I’ve been referencing the BIOSECURE Act in these pages for months — in the biomanufacturing thesis, in the generic drug analysis, in passing asides about procurement signals and supply chain fragility. I’ve treated it as established context. Background architecture. The thing that’s already happened and that everyone in this space already knows about.

I haven’t written a dedicated piece on it. That was a mistake, and I want to correct it now — not because the law is new, but because I think the public conversation about what it does has settled on exactly the wrong feature. The debate was about names. Which Chinese biotech companies would be designated. Whether BGI would make the list. Whether WuXi AppTec’s lobbying campaign would succeed. Whether the five-company approach was too narrow or too broad.

That was the wrong debate. The version of the BIOSECURE Act that passed in the FY2026 NDAA doesn’t name five companies. It builds a machine. And the machine is more important than any list of names could ever be.

The podcast audio was AI-generated using Google’s NotebookLM.

Share

What the Law Actually Built

Let me describe the mechanism, because the legal architecture matters more than the headlines suggested.

Section 851 of the FY2026 NDAA bars federal agencies from procuring biotechnology equipment or services from any “biotechnology company of concern.” It also bars agencies from contracting with entities that use covered biotech equipment or services in performing federal work — a downstream prohibition that extends the law’s reach beyond direct government suppliers into their supply chains. Loan and grant funds are covered too. The scope is broad and deliberate.

But the designation mechanism is where the real design work happened. The law establishes two pathways for identifying a biotechnology company of concern. The first is automatic: any company on the Department of Defense’s Section 1260H list of Chinese military companies that is involved in biotech equipment or services is designated by operation of law. No additional review. No notice. No comment period. You’re on the 1260H list and you touch biotech — you’re covered.

The second pathway is criteria-based. The Office of Management and Budget (OMB) leads an interagency process to identify companies that are subject to the direction or control of a foreign adversary, involved in biotech equipment or services, and assessed to pose national security risks — affiliations with foreign adversary militaries, provision of multiomic data to a foreign adversary, collection of human multiomic data without informed consent. This pathway includes procedural protections: 90 days to respond, periodic review, a process for requesting removal.

Two tracks. One fast and automatic, one deliberate and procedural. Different temporal profiles for different risk profiles. That’s not a blacklist. That’s governance architecture.

The Cybersecurity Lesson Nobody Applied

Here is where I want to draw a connection that I think reframes what the BIOSECURE Act actually represents — and what it tells us about how we’re learning to govern frontier technology supply chains.

For three decades, cybersecurity evolved through a specific failure mode. The early approach was signature-based detection: identify known malware, build a signature, distribute it to endpoints, block the match. It worked — until it didn’t. The attack surface expanded faster than signatures could be written. New variants appeared daily, then hourly. The lag between a novel threat and its corresponding signature became the vulnerability itself. By the time the signature existed, the damage was done.

The industry’s response — the one that actually worked — was behavioral detection. Stop looking for known bad actors. Start looking for patterns of malicious behavior. Build systems that can identify threats they’ve never seen before, based on what the threat does rather than what it is. The shift was from static lists to adaptive systems. From recognition to pattern-matching. From naming the enemy to understanding the behavior that makes something an enemy.

The BIOSECURE Act’s legislative evolution mirrors this transition almost exactly.

The original bills named five companies. That’s signature-based governance. Identify the known threat actors, put them on a list, block them. It would have worked for those five companies. And it would have been obsolete within a year, as corporate restructuring, subsidiaries, joint ventures, and successor entities routed around the designations. You cannot blacklist your way to supply chain security any more than you can signature-match your way to network security. The threat surface evolves faster than the list.

The version that passed builds behavioral detection into the governance architecture. The 1260H pathway captures entities based on their assessed relationship to the Chinese military — a behavioral criterion, not a corporate identity. The OMB criteria-based pathway captures entities based on what they do: whether they’re subject to foreign adversary control, whether they handle multiomic data in specified ways, whether their affiliations pose national security risks. The criteria travel. When a new entity emerges that exhibits the designated behavior, the system can capture it without new legislation.

This is the design principle that matters: the law doesn’t just address the current threat. It builds the institutional capacity to address threats that don’t exist yet. And that distinction — between a law that solves today’s problem and a law that builds the machinery for tomorrow’s — is the distinction between a sandbag and a levee.

The Temporal Gap

But here’s where the design gets complicated, and where I think builders and policymakers need to pay close attention.

The BIOSECURE Act’s prohibitions don’t take effect upon enactment. They take effect after the Federal Acquisition Regulation is revised — 60 days after the FAR update for 1260H-designated entities, 90 days after for criteria-based designations. OMB has one year to compile the initial list. The FAR revision process has its own timeline. A five-year rule of construction protects legacy agreements, including previously negotiated options.

Add it up. The law passed in December 2025. The OMB list arrives no earlier than December 2026. FAR revisions follow. Effective dates trigger months after that. Legacy agreements survive for five years. The full force of the prohibition may not bind across the federal procurement landscape until 2028 or beyond.

I wrote last week about temporal architecture — about the gap between when a system is designed and when it actually operates. The BIOSECURE Act is a case study. The governance intent is sound. The institutional machinery is well-designed. But the implementation timeline introduces a temporal gap during which the very dependencies the law aims to eliminate continue compounding.

Wright’s Law doesn’t pause for rulemaking. Every month that foreign producers continue descending the biomanufacturing learning curve while domestic alternatives are not yet incentivized by the procurement shift is a month the cost gap widens. The 1260H pathway is faster — no procedural protections, no comment period, automatic designation. But it only captures entities already identified as Chinese military companies. The broader criteria-based pathway, which covers the more complex supply chain risks, is the slower one.

This is the governance latency problem applied to procurement policy. The detection happened — Congress identified the vulnerability. The interpretation happened — the law’s criteria are well-specified. But execution latency — the time between legislation and operational effect — is measured in years. And in those years, the problem the law was designed to solve continues operating on its own timescale.

The Levee’s Boundary

There’s a second structural tension that I think deserves more attention than it’s getting.

The BIOSECURE Act covers federal procurement. Executive agencies. Government contracts, grants, and loans. This is the lever the government controls directly, and it’s the right place to start.

But recall the numbers from the BENS report I wrote about in the generic drug piece. Ninety-one percent of American prescriptions are generics. The federal government is a significant pharmaceutical purchaser, but it is not the whole market. The cascading dependency — China to India to American pharmacy counters — operates primarily through commercial supply chains that the BIOSECURE Act does not reach.

The law addresses the 27% of military drug purchases that the Department of Defense study found depend on PRC suppliers. That’s critical. But it doesn’t restructure the commercial supply chain that delivers the other prescriptions — the ones that civilian hospitals, retail pharmacies, and patients depend on. The 679 APIs for which China is the sole KSM supplier don’t become less concentrated because federal agencies stop buying from designated entities.

This isn’t a criticism of the law. It’s a diagnosis of its boundary conditions. The BIOSECURE Act is the first structural levee in a flood zone that extends well beyond the federal procurement riverbank. And understanding what it covers — and what it doesn’t — is essential for anyone trying to build the next section.

The Medicaid Drug Rebate Program safe harbor is a telling detail. The law had to include a specific provision ensuring that drug manufacturers wouldn’t be penalized in the Medicaid system when the national security prohibitions prevent them from executing a VA master agreement. The fact that this carve-out was necessary reveals how deeply entangled the pharmaceutical procurement system is — pull one thread and you risk unraveling programs that millions of patients depend on. The legislators knew this. They built the safe harbor because they understood that the system’s complexity is itself a constraint on the pace of decoupling.

Subscribe now

What the BIOSECURE Act Means for the Spiral

In the biomanufacturing reindustrialization thesis, I described five tactical domains where the spiral needs to enter simultaneously: facility strategy, procurement signals, capital structure de-risking, science infrastructure, and downstream bottleneck identification. The BIOSECURE Act is a procurement signal — the clearest one the federal government has sent to the biotech supply chain.

But a procurement signal without domestic capacity to receive it is a demand curve with no supply response. If federal agencies cannot buy from designated foreign providers, they need qualified domestic or allied alternatives. If those alternatives don’t exist yet — if the facilities haven’t been built, the workforce hasn’t been trained, the learning curves haven’t been descended — then the procurement signal creates disruption without creating restructuring.

This is why the BIOSECURE Act cannot be understood in isolation. It is one instrument in a system that requires simultaneous activation. The law creates the pull. But the push — the capital, the facilities, the continuous manufacturing capacity, the workforce, the allied sourcing partnerships — has to come from the other four domains. The BIOSECURE Act is necessary and insufficient. Exactly as designed. Exactly as it should be — because a single law that tried to do all five things would be a law that did none of them well.

The implementation window is the strategic variable. The years between now and full enforcement are not dead time. They are the window in which domestic and allied biotech manufacturing capacity must be built to receive the demand that the BIOSECURE Act will redirect. If that capacity exists when the prohibitions bite, the law works as intended — a structural intervention that reshapes procurement flows toward secure, resilient sources. If that capacity doesn’t exist, the law creates compliance burdens, waiver requests, and workarounds that preserve the dependencies it was designed to eliminate.

The race is not between the law and the companies it designates. The race is between the law’s implementation timeline and the domestic manufacturing base’s construction timeline. That’s the tempo that matters.

I keep returning to the cybersecurity analogy because I think it carries one more lesson

The shift from signature-based to behavioral detection didn’t happen all at once. It happened in layers. First generation: known signatures. Second generation: heuristic analysis. Third generation: machine learning on behavioral patterns. Each generation was necessary and insufficient. Each generation built the institutional muscle for the next.

The BIOSECURE Act is a first-generation adaptive governance instrument for biotech supply chain security. It builds the machinery — the designation pathways, the interagency coordination, the procedural protections, the FAR integration. It doesn’t solve the problem. It builds the institutional capacity to address the problem over time. And it will need to evolve. The criteria will need refinement. The OMB process will need to get faster. The scope may need to extend beyond federal procurement. The allied coordination dimension — the friend-shoring architecture — will need its own instruments.

But the machinery exists now. The designation pathways are built. The interagency process is specified. The temporal architecture — fast track for known military-linked entities, deliberate process for complex cases — reflects genuine governance design thinking.

The debate about five companies is over. The debate about whether the machinery works — whether the implementation timeline aligns with the construction timeline, whether the procurement signal generates a supply response, whether the levee extends far enough to matter — is just beginning.

At the frontier of technology, the experiment is not whether we can identify the threat. It is whether we can build the institutions that adapt as fast as the threats they govern — and whether we can build the industrial base to absorb the demand those institutions create, before the window closes.

— Titus

I sat in a conference session recently and watched something happen that I’ve seen before but never quite named. Speaker after speaker — technologists, policy people, operators — kept circling the same idea without landing on it. One talked about detection speed for biological threats. Another about the lag between an AI capability and the regulation that addresses it. A third about why manufacturing learning curves are races, not exercises. The language was different each time. The domain was different. The variable was the same.

Not as metaphor. Not as urgency rhetoric — the familiar “we need to move faster” that appears in every keynote and persuades no one. Time as something more fundamental. As the binding constraint that determines whether every other capability — technical, institutional, industrial — actually functions or just exists on paper.

It struck me that for all the frameworks we’ve been building in this space — responsible innovation, governance architecture, reindustrialization strategy — we’ve been designing for capability, authority, and proportionality. We have not been designing for time.

The podcast audio was AI-generated using Google’s NotebookLM.

Share

Speed Is Not the Variable

There’s a distinction worth drawing carefully, because I think conflating two ideas has made this problem invisible.

Speed is a metric. You can measure it, optimize it, benchmark it. Organizations talk about speed constantly. Move fast. Accelerate. Reduce cycle time. Speed is the thing you improve within a system that already works.

Time is the medium in which all your systems must compose. It is not how fast you go — it is whether the systems that must coordinate with each other are operating on compatible timescales. A biosecurity detection system that identifies a threat in twelve hours is useless if the interpretation infrastructure takes twelve weeks and the policy execution mechanism takes twelve months. Each component might be excellent on its own terms. The failure is temporal — they don’t compose.

Engineers have a name for this. Temporal coupling: when two systems that must coordinate operate on fundamentally different timescales, the system breaks. Not because any individual component failed, but because time itself became the fault line.

I want to trace this mechanism across several domains, because I think it explains more about why our current systems are failing than any capability deficit does.

Governance as Temporal Architecture

I wrote about governance latency in these pages earlier this year — the gap between when a system behaves in a new way and when governance responds. I described three components: detection latency, interpretation latency, execution latency. I still believe in that framework. But I’ve started to think I was being too polite about what it actually describes.

Governance latency is not a bug in governance. It is a temporal architecture — one that was designed, intentionally or not, for a world that moved at a different pace. Congressional hearing calendars. Notice-and-comment rulemaking periods. Interagency coordination cycles. These are not merely slow. They operate on a fundamentally different timescale than the technologies they govern. The gap between those timescales is not inconvenient. It is, itself, a space where outcomes are determined before the formal process even begins.

The nation or institution that understands this — that treats temporal alignment as a design variable rather than an operational annoyance — gains an advantage that no amount of capability can offset. Because capability without temporal coordination is potential energy that never converts to kinetic. It sits in reserve, impressive and inert, while the clock runs.

The Circle Is a Clock

Consider biomanufacturing — a domain I’ve been writing about in this series.

The circles-and-spirals thesis is, at its core, a temporal argument. The circle traps organizations in a time loop: no production experience means no yield data means no capital means no facilities means no production experience. The loop is self-reinforcing because each node operates on a timescale that prevents the next node from activating.

Capital allocation cycles are quarterly. Facility construction takes years. Workforce development takes a generation. Yield improvement requires thousands of production hours that nobody can access because the facilities don’t exist.

The spiral breaks the circle not by eliminating time, but by synchronizing it. Government demand signals compress the capital decision. Pre-built infrastructure compresses the facility timeline. Science investment steepens the yield curve so fewer production hours are needed to reach viability. The spiral is not faster in any simple sense — it is temporally coherent. Every node operates on a timescale compatible with the others.

Wright’s Law, the principle that costs decline predictably with cumulative production, is a temporal claim wearing an economic costume. It says: the first mover in production will be the lowest-cost producer, and the gap will compound with time. China is further down the biomanufacturing learning curve than the United States. Every year that gap persists is not a static disadvantage. It is a temporal one — the curve steepens for whoever is on it and flattens for whoever is not.

The Doubling Time of Consequence

Biosecurity is perhaps the most visceral expression of this thesis.

A biological threat does not wait for interpretation. It replicates on its own timescale — exponential, indifferent to institutional calendars. The difference between containment and catastrophe is not capability. We have the sequencing technology, the surveillance infrastructure, the countermeasure platforms. The difference is temporal coordination. Can you detect, interpret, decide, and act within the doubling time of the threat?

I think about this in my work at Vigilance. The entire architecture of biological threat preparedness is, when you strip away the organizational charts and capability matrices, an exercise in temporal engineering.

That’s the design requirement. Everything else is decoration.

From Dimension to Domain

Here is where I want to push further than the conference session went, further than most strategy frameworks go.

We tend to treat time as a dimension — the passive background against which things happen. Decisions take time. Manufacturing takes time. Governance takes time. Time is the water everything swims in.

But the more accurate framing — the one that explains why temporally misaligned systems keep failing in predictable ways — is that time is a domain. A space in which advantage can be built, contested, and lost. A domain that requires its own strategy, its own architecture, its own design principles.

If you accept that reframing, certain things follow.

Temporal advantage is designable. You can build organizations, governance structures, and industrial systems that are optimized for temporal coherence — where the decision cycle, the implementation timeline, and the environment’s rate of change are deliberately aligned.

Temporal disadvantage is structural, not accidental. When a governance system operates on a decadal timescale while the technologies it governs evolve on a monthly one, that is not a speed problem to be solved with urgency. It is an architectural mismatch that requires redesign.

Temporal literacy becomes a core competency. The ability to read a system and identify where temporal misalignment is the binding constraint — rather than capability, authority, or resources — becomes as important as technical expertise or policy knowledge.

What Temporal Design Looks Like

This is where the argument becomes operational, and where I think builders, policymakers, and capital allocators need to pay close attention.

If time is a domain, then every strategy has a temporal architecture — whether or not the strategist designed one. The question is not whether your organization operates within time. The question is whether you’ve deliberately engineered how your organization relates to time.

For builders in frontier technology: the competitive advantage is not always the best technology. It is often the technology that reaches operational deployment first and begins descending the learning curve while competitors are still optimizing in the lab. This is Wright’s Law generalized. The first mover in production compounds an advantage that the better-but-later entrant may never overcome. Time on the curve is the asset. Everything else is a bet that time will be forgiving. It usually isn’t.

For policymakers: governance latency is not a staffing problem or a willpower problem. It is a temporal design problem. The question is not “how do we make government faster” — it is “how do we build governance architectures whose operating timescale matches the domain they govern?” In some cases, that means pre-authorization frameworks that act before the crisis arrives. In others, it means modular governance that can be updated without rewriting the entire regulatory structure. In all cases, it means taking temporal architecture as seriously as institutional authority.

For capital allocators: patience is a temporal strategy, not a virtue. The patient capital that biomanufacturing requires is not charity — it is an investment in temporal alignment, giving the learning curve enough time to generate the yields that make the economics work. The impatient capital that demands returns on quarterly timescales is not merely unhelpful. It is temporally incompatible with the problem it claims to be solving.

I keep returning to that conference session. The speakers kept naming symptoms — speed, latency, urgency, readiness — without naming the condition. The condition is that time is the domain we have not yet learned to design for. We design capability. We design authority. We design architecture. We rarely ask the question that precedes all of them: does this system’s temporal structure match the temporal structure of the problem it exists to solve?

Biological threats replicate on exponential timescales. AI capabilities advance on compressed developmental ones. Governance responds on bureaucratic ones. Manufacturing compounds on production-volume ones. None of these timescales are wrong in isolation. All of them are wrong together — because nobody designed the temporal coherence between them.

At the frontier of technology, the experiment is not whether we can build fast enough. It is whether we can think in time — designing systems where the pace of understanding, the pace of building, and the pace of governing are, for once, composed into the same score.

— Titus

I’ve been thinking about circles.

Not the conceptual kind — though we’ll get there. The physical kind. The ones visible on a map if you trace the journey of a single generic antibiotic from raw chemical to American pharmacy shelf. Start in a chemical plant in Zhejiang province. Ship key starting materials to a synthesis facility in Gujarat. Convert them to an active pharmaceutical ingredient. Ship the API to a formulation plant — maybe still in India, maybe in a bonded zone elsewhere. Tablet, coat, blister-pack, box. Ship the finished dose to a U.S. distribution center. Dispense.

That circle touches three countries, two oceans, and zero domestic manufacturing steps for the active ingredient. For hundreds of medicines Americans take every day, this is not the exception. It is the architecture.

A new assessment from Business Executives for National Security — BENS, the nonpartisan organization that has partnered with senior national security officials since 1982 — maps this architecture with a specificity that should unsettle anyone who has been following the biomanufacturing reindustrialization conversation. Dave Gryska’s team at BENS just released Generic Drug Manufacturing and Biotechnology Innovation — and it does something I’ve been wanting someone outside government to do for years: it treats the generic pharmaceutical supply chain not as a trade policy problem or a healthcare access problem, but as a national security vulnerability with the same structural characteristics as the semiconductor dependency that launched CHIPS.

I’ve written about the biomanufacturing reindustrialization thesis — the circles and spirals, the learning curves, the factory economics, the builder’s playbook. That piece was about the future: the companies and capital structures that could bring biological manufacturing home. This piece is about the present. Because the BENS report reveals that the system we need to reindustrialize isn’t just underbuilt. It is actively structured against its own repair.

The podcast audio was AI-generated using Google’s NotebookLM.

Share

The Numbers Behind the Dependency

Let me put the scale of the problem in terms that are hard to argue with.

Ninety-one percent of drugs prescribed by American physicians are generics. Ninety percent of those generic prescriptions are supplied by India and China. The FDA estimated that approximately 80% of active pharmaceutical ingredients are manufactured overseas. In 2021, 87% of U.S. generic drug manufacturing facilities were located abroad. The United States directly imports roughly 16% of its APIs from China — but India, which supplies the bulk of finished generic doses to American patients, itself imports 80% of its APIs from China.

This is the cascading dependency that the BENS report traces in detail, and it is the feature I want to hold up to the light. Because it’s not a supply chain. It’s a cascade. A disruption in Chinese chemical manufacturing doesn’t just affect Chinese exports. It ripples through India’s pharmaceutical sector and arrives at American pharmacy counters as a shortage — of antibiotics, of cardiovascular drugs, of the basic medications that chronic disease patients depend on to stay alive.

The data on shortages is the trailing indicator. In 2024, the United States recorded its highest number of drug shortages to date — 323 medications affected. Antibiotics have proven especially vulnerable, experiencing shortages at a rate 42% higher than other generics. These aren’t obscure compounds. They’re the medicines that keep a 68-year-old’s blood pressure managed, that treat a child’s ear infection, that a hospital ICU reaches for when a patient goes septic.

And here’s the number that stopped me: the API Innovation Center examined 40 essential drug molecules — the “Vital 40” — and found that India supplies about 63% of these APIs, Europe 22%, China 8%, and the United States just 5%. Five percent. For the most critical generic medicines in the American formulary, domestic production is a rounding error.

But the 8% figure for China is misleading, and the BENS report is careful about why. An analysis of key starting materials (KSMs) — the precursor reagents that are synthesized into APIs — reveals that China is the sole supplier of KSMs for 679 different APIs. India is the sole KSM supplier for 402. The United States and European Union combined are sole suppliers for 44.

That ratio is the supply chain expressed as a strategic position. And it tells you that the dependency isn’t just about who fills the last step — the tablet press, the blister pack. It’s about who controls the chemistry underneath. China’s dominance isn’t at the visible end of the supply chain. It’s at the foundation. The KSMs are the geology, and everything above them — APIs, formulations, finished doses — is built on ground that someone else owns.

Why the Circle Won’t Break Itself

In the biomanufacturing reindustrialization thesis, I introduced the idea that biomanufacturing fails in circles and scales in spirals. The circle: no production experience means no competitive yields, which means no capital investment, which means no facilities, which means no production experience. The system is closed. Nothing moves.

The generic drug supply chain is this circle in its most advanced failure state. It didn’t get here by accident. It got here by economic logic operating without strategic constraint — and the BENS report traces the history precisely enough that the mechanism becomes clear.

The 2000 elimination of tariffs on formulated pharmaceuticals — following the 1995 WTO Pharmaceutical Tariff Elimination Agreement — incentivized importing finished drug products. Domestic manufacturing became progressively less competitive. Facilities closed. The Viatris plant in West Virginia — once employing 1,400 workers producing critical generic medications across antibiotics, cardiovascular, and autoimmune therapeutic areas — is one example among many. America’s last factory producing penicillin closed in 2004. Not because the science was lost. Because the economics were untenable.

And here is where the circle logic becomes vicious. Once domestic production stops, the workforce disperses. Once the workforce disperses, restarting production requires not just capital but a labor pool that no longer exists. Once facilities close, the institutional knowledge embedded in those facilities — the process optimizations, the supplier relationships, the regulatory compliance infrastructure — evaporates. You cannot restart what you’ve forgotten how to do. Or rather, you can — but at vastly greater cost and time than maintaining it would have required.

This is why I called it a system designed to prevent its own repair. The cost advantage of foreign production compounds over time. Every year of offshoring means another year of learning curve accumulated abroad and lost domestically. China’s 20,000+ chemical companies, accounting for 40% of global chemical industrial output, aren’t just competitors. They’re the installed base. And by some estimates, it costs 50% less to produce APIs in India versus the United States or Europe, with labor costs in India and China estimated at one-tenth the cost for a Western company.

The circle doesn’t just trap individual companies. It traps the entire sector. The margins on generic drugs are already razor-thin — that’s the whole point of generics. Domestic production faces stringent regulatory compliance, higher labor costs, higher energy costs, and competition from established foreign producers who have been descending the learning curve for decades. No individual company can break this circle alone. The economics won’t allow it.

This is where the BENS report’s framing matters: this is a national security problem requiring a national security response. Not because the market failed in some abstract sense, but because the market optimized for exactly the wrong thing — unit cost — while ignoring systemic fragility. And fragility, in a system that 91% of American prescriptions depend on, is an existential risk wearing the mask of efficiency.

Stockpiles Are Sandbags, Not Levees

The BENS report takes SAPIR — the Strategic Active Pharmaceutical Ingredients Reserve — seriously, and I want to amplify the point because it cuts against a comfortable assumption in Washington.

In August 2025, an executive order directed the Administration for Strategic Preparedness and Response (ASPR) to identify and stockpile APIs for approximately 26 essential medicines. SAPIR is being refilled and expanded. This is a good and necessary step. Stockpiling APIs, which have longer shelf lives than finished drug products, reduces risk at multiple stress points in the supply chain.

But the BENS report says what I’ve been arguing in different language: stockpiling alone is insufficient. It is a passive defense mechanism. It buys time. It does not solve the structural deficit. Without a functioning industrial base to replenish the stockpile, the reserve is a finite resource that will run dry in a prolonged crisis.

Think about what that means operationally. A geopolitical disruption — a Taiwan Strait crisis, an escalation in the South China Sea, a deliberate Chinese export restriction on pharmaceutical precursors — doesn’t resolve in weeks. The scenario that matters isn’t a temporary supply interruption. It’s a sustained one. And a stockpile sized for months faces a crisis measured in years.

The analogy I keep returning to: SAPIR is sandbags. Essential during a flood. Useless for preventing the next one. What you need is a levee — domestic production capacity that can sustain the flow regardless of what happens upstream in the geopolitical watershed.

And the geopolitical risk is not theoretical. The BENS report documents that China demonstrated its willingness to use raw materials dominance as strategic leverage when Beijing announced new export controls on rare earths and magnets in October 2025. China’s 2020 Export Control Law and 2021 Biosecurity Law provide extensive powers to weaponize pharmaceutical exports. A 2023 Department of Defense study found that 27% of U.S. military drug purchases depend on PRC suppliers. During COVID, India temporarily banned exports of 26 APIs and pharmaceutical formulations. China’s lockdowns shuttered approximately 37 pharmaceutical factories manufacturing active ingredients for U.S. drug products.

These aren’t hypotheticals. They’re precedents.

Where the Spiral Starts for Generic Drugs

So where does the spiral begin? How do you break a circle this entrenched?

The BENS report’s recommendations align with the NSCEB architecture I’ve written about — and that alignment itself is significant. When a nonpartisan national security organization staffed by senior business executives arrives at the same structural conclusions as the government commission, the signal is strong. Incentivize domestic manufacturing. Enhance supply chain transparency. Advance manufacturing innovation. Strengthen allied partnerships. Expand strategic reserves. Develop the workforce.

But I want to push beyond the recommendation headers — because the generic drug challenge reveals something specific about where the biomanufacturing spiral must enter.

The entry point isn’t biologics. Not yet. The entry point is chemistry.

The KSM dependency — China as sole supplier for 679 APIs’ starting materials — means that reindustrialization for generic drugs starts with chemical synthesis capacity. This is not the glamorous frontier of synthetic biology and AI-driven biodesign that dominates conference panels. It is industrial chemistry. Reactor vessels and distillation columns. Solvent recovery and waste treatment. The unsexy infrastructure that makes everything else possible.

And this is where the advanced manufacturing innovation piece becomes critical — and where I think the conversation needs to get more specific than the recommendation headers usually allow.

The BENS report recommends continuous manufacturing technologies. I want to explain why this isn’t incrementalism. It’s the physics that changes everything.

Traditional batch chemical manufacturing — the dominant mode for API synthesis globally — runs in discrete cycles. Load reagents, react, purify, test, repeat. Each batch is a separate event. Each event requires operators. Each operator costs ten times more in West Virginia than in Gujarat. That math is why the factories left.

Continuous flow chemistry inverts the model. Reagents flow through microreactors or tubular systems without stopping. The reaction happens in transit. Purification is in-line. Quality monitoring is real-time. A continuous process that runs 24 hours produces in a closet-sized reactor what a batch plant produces in a warehouse — with fewer operators, less waste, tighter process control, and yields that improve with runtime because the system is constantly self-correcting.

Here’s the Wright’s Law connection that I think the policy community hasn’t fully internalized: continuous manufacturing doesn’t just reduce labor cost. It steepens the learning curve. In batch manufacturing, each batch is semi-independent. In continuous manufacturing, every hour of runtime generates process data that feeds back into optimization. The learning compounds at the rate of hours, not batches. That’s the difference between a flat curve and one that actually descends fast enough to close the cost gap with foreign producers within a commercially relevant timeframe.

The capital cost is higher upfront. The regulatory validation pathway is less established — FDA has been encouraging continuous manufacturing, but the approval precedents are still accumulating. These are real barriers. But they are surmountable barriers, which is categorically different from the structural barrier of trying to compete on labor cost in batch production. You can solve an engineering and regulatory problem. You cannot solve the fact that your labor costs ten times what your competitor’s does.

The reporting requirements the BENS recommends — mandating that manufacturers disclose the geographic origins of APIs, KSMs, and finished drug products to federal agencies — would create the supply chain map that currently doesn’t exist. And you cannot fix a system you cannot see. The supply chain mapping recommendation — a national database tracking critical nodes, chokepoints, and dependencies — is the diagnostic tool that enables everything else. It transforms supply chain vulnerabilities from unknown risks into managed variables.

The allied sourcing piece connects directly to the friend-shoring thesis I’ll develop further in this series. Canada, the EU, Japan — nations with robust regulatory standards and aligned strategic interests — should be preferred sourcing partners for the supply chain segments that cannot be reshored immediately. Not autarky. Strategic diversification. Building redundancy through allies who share democratic values and regulatory standards, so that no single point of failure — geopolitical, climatic, or logistical — can cascade through to American patients.

What the Generic Drug Supply Chain Teaches Biomanufacturing Builders

Here’s where I want to connect back to the thesis, because the BENS report illuminates something about biomanufacturing reindustrialization that the startup-focused conversation misses.

The generic drug supply chain is the finished product of the circle. It’s what the biomanufacturing ecosystem becomes if the spiral doesn’t start in time. Two decades of rational economic decisions — each one defensible in isolation — producing a system so fragile that a factory fire in Gujarat or an export restriction from Beijing can leave American hospitals rationing antibiotics.

The lesson for biomanufacturing builders is temporal. The window I wrote about in the first installment of this series isn’t just about the learning curve for biologics and engineered organisms. It’s about preventing the generic drug story from repeating in the next generation of medicines. Biosimilars. Cell and gene therapies. mRNA therapeutics. The production platforms of the next two decades are being established right now. Where they’re built, who operates them, and whose supply chains they depend on — those decisions, made in the next five years, will determine whether the next generation of medicines is manufactured in a system that looks like the generic drug cascade or one that looks like a resilient, diversified, domestically anchored spiral.

The BENS report puts data behind what builders already sense: the cost of dependency is not measured in the price of a pill. It is measured in the 323 shortages. In the hospitals that cannot access basic antibiotics. In the military pharmacies that depend on adversarial supply chains for 27% of their purchases. In the cascading fragility that turns a distant factory closure into an American patient’s missed dose.

The engineer I described in the first installment of this series — the one troubleshooting the fermentation run at 5,000 liters — represents the future of biomanufacturing. Skilled, operational, working at the frontier of what domestic production can do.

But there’s another figure in this story who deserves attention. The pharmacist checking inventory and discovering that three generic antibiotics are backordered with no estimated resupply date. The hospital procurement officer calling a second distributor, then a third, then issuing an internal alert about rationing protocols. The patient whose chronic medication — the one that’s kept their condition managed for eight years — is suddenly unavailable.

Subscribe now

They don’t stand in a facility that smells like warm yeast. They stand at the endpoint of a supply chain that routes through Zhejiang and Gujarat and arrives — or doesn’t — at an American counter.

The reindustrialization thesis isn’t just about building the future. It’s about repairing the present. And the present is 679 to 44, 323 shortages, and a system where the most powerful nation on Earth cannot reliably produce the basic medicines its citizens take every morning.

The spiral has to start somewhere. The BENS report tells us exactly where.

At the frontier of technology, the experiment is not whether we can engineer the next generation of medicines. It’s whether we can manufacture the current one — here, reliably, without routing our health security through nations that have already demonstrated their willingness to use it as leverage.

The circle is patient. It will wait for us to notice.

— Titus

There is a question I keep returning to — one that sits underneath the policy debates, the appropriations fights, the executive orders, and the increasingly urgent memos circulating through the national security establishment:

Not design them. Not discover them. Not publish papers about them. Make them. At scale. Reliably. On domestic soil. With a workforce that exists and supply chains that don’t route through adversarial nations.

The answer, right now, is: barely.

And it looks like this. Last year I stood in a pilot biomanufacturing facility — one of the few we have — and watched a team troubleshoot a fermentation run that had gone sideways at 5,000 liters. The organism was producing at bench scale. It had produced at 500 liters. At 5,000 liters, oxygen transfer became the constraint — at least, that’s what the team suspected in real time. The metabolic profile shifted. Yield dropped by a third. The lead process engineer — one of a few hundred people in the country with this specific operational expertise — was working through it with a combination of sensor data, experience, and what I can only describe as biological intuition. There was no model to consult. No binder. No runbook. No second shift that had seen it before. She was debugging a living system at industrial scale, mostly alone, in a building that smelled like warm yeast and sounded like a submarine engine room.

That scene is the bioeconomy. The rest is narration.

I’ve spent two+ years as a commissioner on the National Security Commission on Emerging Biotechnology, producing the most comprehensive governmental blueprint for biomanufacturing reindustrialization the United States has published. The NSCEB report is thorough. It is specific. I believe in the architecture. But government reports — even good ones — describe what should happen. They do not describe what it feels like to build in the gap between should and is. And they do not tell the builders, the operators, and the capital allocators where the openings actually are.

That’s what I want to do here. This is the first in a series.

The podcast audio was AI-generated using Google’s NotebookLM.

Share

The Gap Nobody in Washington Understands Viscerally Enough

Two pieces of writing crystallized the biomanufacturing problem for me recently, and neither was about biology.

The first is Aaron Slodov’s “American Shenzhen” framework — a detailed blueprint for rebuilding U.S. hardware manufacturing capacity: government as anchor tenant, 75/25 commercial-to-defense revenue models, special economic zones, streamlined permitting, venture-backed manufacturing startups. The insight is structural: Shenzhen didn’t emerge from a single policy. It emerged from a system — procurement signals, physical infrastructure, workforce pipelines, and regulatory architecture reinforcing each other simultaneously.

The second is Oliver Hsu’s recent primer on factory economics for a16z, which articulates what venture capital is only now internalizing: the IP is the process. In these companies — and biomanufacturing startups are definitionally among them — the moat is not intellectual property in a patent filing. It is the production process itself. The yield curves. The learning rates. The operational knowledge embedded in people and equipment and process.

I read both and thought: this is the framework I wish I could have injected directly into the NSCEB’s deliberations.

Because here is what the data looks like from inside the commission. The United States’ share of global API production has collapsed from roughly 23% to approximately 3% over three decades. More than 90% of generic pharmaceuticals consumed in the U.S. depend on imported ingredients. Industry surveys indicate that roughly 80% of biopharma organizations are actively engaged with Chinese CDMOs. China holds approximately 58% of global synthetic biology patent filings, 28% of biological manufacturing patents, and 30% of novel antibiotics patents. That’s not an edge. That’s installed capacity. That’s a country that has been running production volume while America has been running conferences about it.

The NSCEB report lays this out. The number one message is the urgency. The window for American biomanufacturing reindustrialization is open, but it is not open indefinitely. And the dynamics that will close it are not political. They are economic.

The Learning Curve Is a Race — And We’re Losing It

This is where Hsu’s factory economics framework becomes essential, and where the venture and builder community needs to pay close attention.

Wright’s Law tells us that costs decline predictably with each doubling of cumulative production. The learning curve is the race that defines factory companies. The competitor with more cumulative production has lower costs. Lower costs win more contracts. More contracts mean more production. More production steepens the curve. The advantage compounds.

China is further down the biomanufacturing learning curve than the United States in multiple product categories. Every year America does not build domestic production capacity is a year China accumulates more volume, drives costs lower, and makes the gap harder to close. This is not a static competition. It is a dynamic one, and the dynamics favor whoever starts manufacturing first and fastest.

Now, biology adds a complication that makes this race harder than any hardware equivalent. Biological systems are stochastic. A fermentation run that works at 10 liters may behave differently at 10,000 liters — not because of engineering error, but because living organisms respond to conditions in ways we don’t fully understand. That engineer I watched troubleshooting the 5,000-liter run? She was navigating exactly this problem. The yield curve in biomanufacturing is less predictable than in semiconductors, aerospace, or any other production domain. And yield is the single highest-leverage variable in factory economics. A 20-point yield advantage can create a cost differential that determines who survives.

This is the core tension, and I want to name it precisely because I think it’s the single most important concept in biomanufacturing strategy:

The circle: the learning curve demands production volume, but production volume requires facilities that won’t get built without capital, and capital requires the predictable yields that only come from production experience. No yields, no capital. No capital, no facilities. No facilities, no learning. No learning, no yields. The system is closed. Nothing moves.

The spiral: break into the circle at any point — with government demand signals, with patient capital, with science that steepens the yield curve — and the circle becomes a spiral. Production generates learning. Learning improves yields. Yields unlock capital. Capital builds more facilities. Facilities train the workforce. The workforce improves operations. Operations improve yields. The system opens. Everything moves.

In a circle, you die waiting for certainty. In a spiral, you manufacture your way into it. The circle is what kills startups. The spiral is what makes nations competitive.

The NSCEB understood this. The report’s six pillars — political commitment, private sector mobilization, defense integration, innovation infrastructure, workforce, and allied coordination — are designed as a system specifically to break the circle and start the spiral. Push on every node simultaneously. That’s the architecture.

But here’s what I want to tell the builders and investors directly: the government is going to be slow. The NSCEB report took two years. Implementation will take longer. If you wait for the full system to be in place before you move, you will be too late.

How a Builder Wins in the New Landscape

Let me be concrete about this. Instead of listing the plays abstractly, I want to walk through what the next twelve to thirty-six months look like for someone building a domestic biomanufacturing company — and how the NSCEB architecture, as it comes online, changes the game at each stage.

You start with a facility strategy

This is the first decision and the one most biotech founders get wrong, because they’re trained to think about molecules first and production second. In the new landscape, production is the strategy. You need a facility — or access to one — that can run at pilot scale today and commercial scale within three years. The NSCEB recommends a network of precommercial biomanufacturing facilities operated through DOE and DOC, plus a $120 million biopharma manufacturing center under the Defense Bioindustrial Manufacturing Program. If you are a startup, the question is whether you build your own or position to be the anchor tenant in one of these government-catalyzed facilities. Either way, think about geography: proximity to a research university with relevant programs, an existing labor pool with manufacturing experience (not just biology PhDs — people who know how to run plants), and state-level incentive structures. The NSCEB’s regional hub model matters here. The companies that co-locate with the hub infrastructure will compound advantages that distant competitors cannot replicate.

The procurement signal de-risks your first offtake

The DBIMP at $762 million or more. Advanced market commitments and offtake agreements from DOD and HHS. The BIOSECURE Act — signed into law in late 2025 — already restricting federal contracts with foreign biotechnology companies of concern. That’s the stick. The DBIMP and AMCs are the carrot. If you are building domestic production capability for APIs, sustainable aviation biofuels, biomaterials, or engineered proteins, the federal procurement pipeline is about to open in ways it hasn’t before. This is Slodov’s “government as anchor tenant” — and it’s the single most important de-risking event for early-stage biomanufacturing companies. The companies that have production capability when the procurement dollars flow will capture the contracts. The ones that wait for certainty will find the contracts already allocated and the learning curve already claimed.

Government de-risking shifts your capital structure

This is where the NSCEB does something genuinely novel. The report recommends an Independence Investment Fund at the Department of Commerce — subordinated capital, loan guarantees, co-investment structures. Combined with targeted tax credits for biomanufacturing capex, this changes the financial physics.

Hsu’s framework explains why. Factory startups get stuck in the equity-only phase of the capital stack because they can’t demonstrate predictable yields to unlock lower-cost capital. Venture equity is expensive. It is also the wrong capital for manufacturing ramp. What you need for a factory is a progression: equity for R&D and process development, venture debt for equipment with clear payback, equipment financing for production expansion, and project finance for new facilities with contracted offtake. Each transition requires demonstrating more operational predictability than the last.

Government de-risking — subordinated capital from and Independence Investment Fund, guaranteed demand from AMCs, tax credits that improve the capex math — is what enables these transitions. It doesn’t replace private capital. It unlocks it. The VCs who understand this will structure their investments to ride the government co-investment, not compete with it. The ones who try to fund biomanufacturing the way they fund SaaS will keep writing checks that don’t come back.

The science infrastructure becomes your tooling tailwind

The NSCEB proposes $5 billion over five years to make biology predictably engineerable. $540 million over three years to make scale-up predictable and cost-competitive, targeting four bottlenecks — chassis organisms, feedstocks, process technology, and critical inputs like growth media and purification resins. Plus $640 million for AI-ready biological data standards at NIST. A Web of Biological Data at DOE. An NSF cloud labs network. Six Centers for Biotechnology in the National Laboratories.

This is not a research agenda. It’s an industrial one.

The companies that will benefit most are not the ones waiting for this infrastructure to be complete. They’re the ones building tools and platforms that accelerate it — and that become indispensable as the infrastructure scales. AI-driven biodesign. Automated process optimization. Digital twins for fermentation and downstream processing. Scale-up prediction models trained on the standardized data that NIST and DOE are about to generate. These are venture-scale opportunities that ride the government investment wave while creating independent commercial moats.

I want to be specific about why this matters for the circle-to-spiral conversion. The reason biomanufacturing yield curves are shallower than semiconductor yield curves is not that biological systems are inherently unoptimizable. It’s that we lack the data infrastructure, the computational tools, and the standardized measurement frameworks to learn from production experience systematically.

We don’t have a design-for-manufacturing paradigm in biology yet.

Every biomanufacturing facility is, to a significant degree, reinventing the wheel — because the data from the last facility isn’t in a format the next facility can use. The companies that build the connective tissue — the platforms that turn production data into transferable operational knowledge — will be the ones that steepen the learning curve for the entire industry.

Then workforce becomes the edge that compounds. The NSCEB data is stark: unmet demand across biotech roles runs 38-68%. Bioindustrial manufacturing demand is up 23% with pipelines lagging far behind. But the workforce follows the facilities. Shenzhen didn’t train workers and then build factories. It built factories and the workforce grew around them — through proximity, repetition, and operational experience accumulating over time.

The companies that build production facilities in regions with existing educational infrastructure and labor pools — and that invest in training as a core strategic function, not an HR afterthought — will have an advantage that compounds quarterly. The skills that matter — debugging a bioreactor at scale, optimizing downstream processing, managing GMP compliance while maintaining throughput — are developed on production floors, not in classrooms. Every month of production experience your team accumulates is a month your competitors don’t have.

This is the spiral in action. The facility generates production experience. Experience improves yield. Better yield attracts capital. Capital expands production. Expanded production trains more workers. More workers improve operations. Better operations improve yield again. Each rotation of the spiral, the moat deepens.

Subscribe now

The Ugly Middle: Where Factories Actually Bleed

If you’ve read this far and you’re thinking mostly about upstream — organisms, fermentation, titers — you’re making the mistake most people make. Upstream is where the science lives. Downstream is where the margins go to die.

Purification, analytical characterization, QA/QC, validation, fill-finish — this is where biomanufacturing timelines stretch and costs compound. It’s less glamorous than strain engineering. It is also where the majority of production cost and schedule risk accumulates. The U.S. constraint is often not the bioreactor — it’s the chromatography columns, the single-use filtration systems, the analytical instrumentation, and the trained QC analysts who can run them. These are physical, unglamorous bottlenecks, and they are frequently the binding constraint that determines whether a facility can actually deliver product at the quality and timeline a contract requires.

Any serious builder knows this. The NSCEB’s scale-up grand challenge targets process technology and critical inputs for exactly this reason. But the policy language is clinical. The operational reality is that downstream processing is where you discover whether your factory actually works — and in biomanufacturing, unlike semiconductors, you often discover it at the worst possible moment: when you’re trying to meet a delivery commitment.

The companies that solve downstream — that build the equipment, develop the resins, automate the analytics, train the QC workforce — will own a piece of every biomanufacturing company’s cost structure. That is not a niche. That is a platform.

What Could Kill This

I’ll be direct about the risks, because the builders need to price them.

Fragmentation

The NSCEB recommends a National Biotechnology Coordination Office in the Executive Office of the President — the institutional equivalent of the National Space Council, but for biotech. Without it, or with a version that lacks statutory authority and dedicated financial oversight, the recommendations scatter across HHS, DOD, DOC, DOE, USDA, NSF, EPA, and a dozen sub-agencies. I’ve watched this happen. An EOP office without teeth is a convening body that produces memos. The NBCO needs to be the one with the pen on the national biotechnology strategy and the authority to align agency budgets behind it. Whether it gets that authority is a political question, not a technical one.

Partial funding

Congress may fund pieces of the system but not the system itself. And partial funding of a system is worse than no funding at all — it produces individual components that can’t function without the others. The circle doesn’t break if you only push on one node. You get a biopharma manufacturing center with no trained operators. A workforce program with no facilities to train in. A data standards initiative with no production data to standardize. The system logic is the NSCEB’s greatest strength and its greatest vulnerability — because systems require comprehensive investment, and comprehensive investment requires sustained political will, and sustained political will requires visible threats, and biomanufacturing’s threats are invisible until they’re catastrophic.

Speed

This is the risk that keeps me up. Government will be slower than the competition. That’s structural. Authorization takes years. Appropriation takes more. Implementation takes more. Each year of delay is a year China accumulates production volume, drives down costs, and locks in the learning curve advantage that compounds and compounds and compounds. The window doesn’t close in a day. It closes in a decade of days where nothing happened fast enough.

Commercial viability

Even with government de-risking, domestic biomanufacturing must eventually be cost-competitive without subsidies. If production costs remain significantly higher than foreign alternatives — and they currently are for many product categories — the demand that procurement creates won’t expand into commercial markets. This is where the grand research challenges are existential, not optional. Making biology predictably engineerable is not an academic aspiration. It is the prerequisite for the learning curve to steepen. It is the prerequisite for the spiral to accelerate past the point where government support is needed.

The Frontier Firm Meets the Factory Floor

Everything I’ve described so far — the learning curves, the capital stack, the workforce — is hardware. And reindustrialization at speed is always a software problem, too. The companies that win won’t just be factory companies. They’ll be software companies that happen to make molecules.

I’ve been thinking about this through the lens of what Microsoft calls the Frontier Firm — organizations restructured around AI agents, where human-agent teams replace traditional hierarchies and the “Work Chart” replaces the org chart. I’ve argued that governance architecture has to be designed before capability scales, not after. That autonomous systems which outpace human oversight create accountability vacuums. That velocity without accountability doesn’t scale — it detonates.

Biomanufacturing is about to test that thesis in steel and concrete.

The companies that will win the biomanufacturing reindustrialization are not going to be traditional biotech companies that happen to build factories. They are going to be Frontier Firms that happen to work in biology. AI-driven biodesign. Autonomous process optimization. Digital twins that predict fermentation behavior before you run the batch. Agent systems managing QA/QC workflows at a pace and consistency that human teams alone cannot sustain. The economics demand it — yield curves, learning rates, cycle time optimization are all domains where AI-agent integration is not optional but existential.

And this is where the two theses connect. The Frontier Firm needs governance architecture: accountability ledgers, rules of engagement, clear chains of human responsibility for autonomous decisions. The biomanufacturing Frontier Firm needs all of this plus regulatory compliance in one of the most heavily regulated production environments on Earth. FDA, EPA, USDA. GMP. Process validation. Batch records. Chain of custody.

The companies that build both — the AI-agent operational architecture and the governance frameworks to make it regulatable — will be the ones that define the next era of biological manufacturing. This is not a future problem. The tools exist now. The regulatory conversations are happening now. The builders who move first will set the standards everyone else has to follow.

I’ll go deeper on this in future editions — what the biomanufacturing Frontier Firm actually looks like, how AI-agent systems integrate with GMP production, where the governance models from defense autonomy apply to factory floors. This is a series, not a single argument.

The Wager

Here’s where I land.

The NSCEB designed a system. Six pillars, interdependent, mutually reinforcing. Political architecture. Capital formation. Defense procurement. Innovation infrastructure. Workforce. Allied coordination — extending the production base through AUKUS, Quad, NATO, and the Wassenaar Arrangement, because reindustrialization is not autarky. It is a strategic rebalancing of where critical capabilities live and whose supply chains they route through.

But what occupies my thinking now — what I think about when I’m not thinking about the policy machinery — is the builder’s version of this thesis. The version where a venture-backed company builds domestic fermentation capacity and captures the DOD offtake before the procurement system fully stands up. Where a startup creates the AI-driven biodesign platform that rides the NIST data standards buildout. Where a regional hub — maybe in Oklahoma City, maybe outside St. Louis, maybe in a place nobody’s thought of yet — becomes the Shenzhen of biomanufacturing. Not because a government report said it should, but because someone built the factory, trained the workers, ran the bioreactors, solved the oxygen transfer problems at 5,000 liters, and drove down the learning curve faster than anyone thought possible.

The government can signal demand. It can de-risk capital. It can fund science. It can build infrastructure. It can coordinate allies.

It cannot manufacture. It cannot operate. It cannot descend the learning curve.

That requires builders.

The factory is the product. America has been designing the product for decades while outsourcing the factory. The policy signal to bring the factory home is louder than it has ever been. The NSCEB’s architecture is designed to break the circle and start the spiral — but the spiral only turns if someone is inside it, building.

The window is open, but the curve is compounding.

Who builds it?

At the frontier of biology, the experiment is not whether we can engineer life. We can. The experiment is whether we can manufacture it — here, reliably, at the scale the century demands. The competition isn’t waiting. They’re doubling cumulative volume.

— Titus

This is the first in a series on the biomanufacturing reindustrialization thesis. Upcoming editions will cover: the biomanufacturing Frontier Firm and AI-agent integration in GMP production; the capital stack in detail — how LPs, VCs, and government co-investment vehicles should be structured; the downstream bottleneck and the companies solving it; and the workforce problem as a compounding strategic advantage. If you’re building in this space, I want to hear from you.

This is a special edition of The Connected Ideas Project, because while it’s Episode 60 of the podcast, it’s the 100th edition of this newsletter since launch! Thank you for being part of this community. If you’re finding value, please share with your friends and colleagues!

Share

Every few years, a major technology company publishes a report that tells you more than it intends to.

Microsoft’s 2025 Work Trend Index — “The Year the Frontier Firm Is Born” — is one of those reports. On the surface, it’s a well-produced argument for why every company needs to reorganize around AI agents. Survey 31,000 workers across 31 countries, add LinkedIn labor data and Microsoft 365 telemetry, wrap it in a compelling narrative about hybrid human-agent teams, and you’ve got a document that every Fortune 500 CEO will have on their desk by Friday.

But read it a second time. Read it the way you’d read a national security assessment — not for the headlines, but for the assumptions underneath. And something more interesting emerges.

Microsoft isn’t describing a productivity tool. They’re describing a new theory of the firm. And they’re describing it almost entirely in the language of efficiency, without ever seriously grappling with the governance architecture such a firm would require.

That gap is where the real story lives. Every time we build autonomous capability faster than we build accountability, the system doesn’t fail immediately. It fails later. And it sends the bill.

The podcast audio was AI-generated using Google’s NotebookLM.

Subscribe now

What the Report Actually Says

Give Microsoft credit: the diagnosis is sharp. Business demands are outpacing human capacity. Eighty percent of the global workforce says they lack enough time or energy to do their work. The knowledge worker, as currently configured, is maxed out.

Microsoft’s answer: intelligence on tap. AI agents that can reason, plan, and execute tasks autonomously — not chatbots, but digital colleagues joining teams with increasing independence. The report envisions three phases, from AI as assistant to AI as operator of entire business processes, and argues that companies embracing this trajectory are already pulling ahead.

The numbers tell the story. Eighty-two percent of leaders expect to deploy agents to expand workforce capacity in the next eighteen months. Forty-six percent are already using them to automate entire workflows. A third are considering headcount reductions. And here’s the number that stopped me: 78 percent are considering hiring for AI-specific roles that didn’t exist a year ago.

This is not incremental. Microsoft even coins a term — the “Work Chart” — to replace the org chart: a dynamic, outcome-driven model where teams form around goals, not functions, powered by agents that expand what each person can do.

The Movie Production Model — and Its Missing Script

One of the report’s most revealing analogies compares the Frontier Firm to movie production. Teams assemble for a project, agents fill specialized roles, the work gets done, and everyone disbands. It’s a compelling image. Lean, high-impact, fluid.

I’ve been thinking about that analogy. Because it captures something real about where organizational design is heading. But it also reveals what the report doesn’t address.

Movie productions work because of something the report never mentions: an extraordinarily mature governance infrastructure. There are unions, guilds, contracts, liability frameworks, insurance requirements, safety protocols, credential verification systems, and chain-of-command structures that have been refined over a century. The fluidity of production is enabled by the rigidity of the rules governing it.

What is the equivalent for human-agent teams?

When Microsoft describes a world where every employee becomes an “agent boss” — someone who builds, delegates to, and manages AI agents — they’re describing a massive delegation of judgment. And delegation of judgment, in any complex system, is a governance problem before it’s a productivity solution.

I keep thinking about this because it mirrors a challenge I’ve watched play out in a different domain entirely.

What Defense Planners Already Know

I remember a briefing at the Pentagon — one of many, but this one stuck. A program manager was presenting an autonomy roadmap for a logistics system. Slides were clean. The capability curve was steep. Savings projections were compelling. And then someone from the policy shop asked a single question: “Who signs for the decision when the system gets it wrong?” The room went quiet. Not because the question was unexpected. Because everyone knew the answer wasn’t in the slides.

That moment comes back to me reading Microsoft’s report. Because the concept they’re selling as a business revolution — human-machine teaming with autonomous systems — is something the Department of Defense has been grappling with for over a decade. Different vocabulary, same structural problem: How do you maintain meaningful human oversight when the systems you’re working with can operate faster, and increasingly more capably, than the humans directing them?

The defense community learned several things the hard way.

First, that the “human in the loop” is not a design feature — it’s a design requirement that must be engineered deliberately, or it erodes. Systems that are faster and more capable than their operators create irresistible pressure to defer. The human becomes a rubber stamp. In military contexts, this is called automation bias. In Microsoft’s Frontier Firm, it has no name yet. But the dynamic is identical.

Second, that trust calibration matters as much as capability. The report’s own data hints at this: 52 percent of workers see AI as a command-based tool, while 46 percent see it as a thought partner. That split isn’t a preference — it’s a reflection of how well people understand what they’re delegating. Miscalibrated trust — too much or too little — is how autonomous systems fail in operational environments. The military has spent billions learning this lesson. The report proposes that every employee learn it on the job.

Third, and most importantly: the governance architecture has to be designed before the capability scales, not after. The DoD doesn’t deploy autonomous systems and then figure out the rules of engagement. The rules come first. They’re imperfect, they evolve, but they exist before the system is operational.

Microsoft’s report proposes deploying autonomous agents across entire business functions and then building the governance afterward. They call this journey “Phase 1 to Phase 3.” A defense planner would call it an operational risk.

The Responsible Innovation Gap

Here’s what fascinates me about the Frontier Firm concept. It’s a genuinely interesting framework for thinking about organizational transformation. The capacity gap is real. The potential for AI agents to expand what small teams can accomplish is real. I personally see this every day with the engineering teams I run. The shift from functional org charts to outcome-driven work charts is a prediction I think will prove directionally correct.

But the report treats governance as an afterthought — a problem to be solved after the productivity gains are captured. And this is a pattern I’ve seen before.

One of the themes we’ve explored in the Science of Responsible Innovation is that the time to design governance into a system is during the architecture phase — not during deployment, and certainly not after failure. Violet Teaming exists precisely because the traditional approach — build it, ship it, regulate it — doesn’t work when the systems in question are capable of autonomous action.

The Frontier Firm, as Microsoft describes it, would have AI agents running supply chains, managing customer relationships, executing financial analysis, and operating business processes end-to-end. Each of these involves decisions with consequences for real people — employees, customers, communities. The report mentions a “human-agent ratio” as a new business metric. But a ratio tells you headcount, not accountability. Who is responsible when an agent makes a consequential error in a process it was running autonomously? The agent boss? The agent’s developer? The company that deployed it? The platform provider?

These are not hypothetical concerns. They’re the same questions that biosecurity experts ask about autonomous laboratory systems, that defense ethicists ask about lethal autonomous weapons, and that financial regulators ask about algorithmic trading. The pattern is consistent: autonomous systems that operate faster than human oversight can track them create accountability vacuums.

There’s a concept the defense community uses that Microsoft’s Frontier Firm badly needs: rules of engagement. Before any autonomous system operates, there are explicit boundaries — what it can do, what requires human authorization, who owns the consequence of each class of action. Call it an Accountability Ledger for the Frontier Firm: a document, maintained alongside the Work Chart, that maps every agentic process to a human owner who answers for its outputs. Not the person who prompted the agent. The person who is responsible when the agent’s decision costs someone their job, their loan, their medical claim. The Work Chart tells you who does what. The Accountability Ledger tells you who answers for what. You need both, or you have neither.

Microsoft mentions Daniel Susskind’s hypothesis that human work will persist because of three limits: efficiency, human preference, and moral judgment. That’s a reasonable framework. But notice the order. Efficiency is the domain AI masters first. Human preference erodes as people habituate. Moral judgment is the last holdout — and it’s the one the report spends the least time on.

The Real Frontier

I don’t think Microsoft is wrong about where organizations are headed. The data is too consistent, the economic pressure too strong, the capability curve too steep. Some version of the Frontier Firm is coming. The question is whether it arrives as a thoughtfully governed institution or as a productivity-optimized system that discovers its governance gaps through failure.

The report notes that 33 percent of leaders are considering headcount reductions. It notes that 81 percent of employees haven’t changed jobs in the past year and that the labor market is functionally frozen. It notes that AI literacy is now the most in-demand skill on LinkedIn, alongside conflict mitigation, adaptability, and innovative thinking.

Read those data points together. They describe a workforce being asked to adapt to a fundamental restructuring of their relationship to institutional output — while the labor market offers them no exit, the governance frameworks offer them no protection, and the timeline offers them no breathing room.

That’s not a productivity story. That’s a social contract story. And it deserves the same rigor we bring to governing autonomous systems in defense, in biosecurity, in any domain where the speed of the system can outpace the judgment of the humans nominally in control of it.

The most startling finding in the entire report might be the smallest: when asked why they turn to AI over a human colleague, the number-one reason employees cited was 24/7 availability. Not quality. Not speed. Not creativity.

Availability.

There is a version of the Frontier Firm that works — one designed with governance, accountability, and human agency built in from the start. Where the human-agent ratio reflects not just efficiency but responsibility. Where “agent boss” means not just managing outputs but owning consequences. Where the Work Chart includes not just who does what, but who answers for what when the system does something no one intended.

That version requires the kind of cross-domain thinking that doesn’t live in any single corporate report. It requires people who understand autonomous systems governance AND organizational design AND labor economics AND the specific ways that speed and capability create accountability gaps.

The org chart is dying. Microsoft is right about that.

But the thing that replaces it will be defined not by the companies that move fastest, but by the ones that build the governance architecture to match the capability they’re deploying. The history of autonomous systems — in defense, in finance, in biosecurity — teaches this lesson with uncomfortable consistency: velocity without accountability doesn’t scale. It detonates.

We are not building a new kind of company. We are building a new kind of institution. And the institutions that last — the ones that earn trust, that survive their own power — have never been built on efficiency alone. They are built on the willingness to answer for what they do.

Intelligence on tap is a capability. Judgment-by-design is a choice.

The Frontier Firm will be defined by which one it optimizes for.

— Titus

A few months ago, when we first started talking about the Science of Responsible Innovation at The Connected Ideas Project, I kept coming back to a simple question:

How do we know?

How do we know whether a technology is actually as powerful—or as dangerous—as we imagine?

How do we know whether our fears are grounded in evidence or in extrapolation?

How do we know whether policy is steering something real, or something hypothetical?

It’s one thing to run a model through an in silico benchmark and watch it ace a virology exam. It’s another thing entirely to put a pipette in a novice’s hand and see what happens in a real lab.

That’s why the recent paper, “Measuring Mid-2025 LLM-Assistance on Novice Performance in Biology,” feels so important.

Not because it proves that AI is safe.

Not because it proves that AI is dangerous.

But because it does something rarer and more valuable: it measures.

And in doing so, it gives us a template for what responsible-by-design evaluation can look like in the age of frontier AI and synthetic biology.

The podcast audio was AI-generated using Google’s NotebookLM.

Share

The Gap Between the Benchmark and the Bench

For the last several years, large language models have been climbing biological benchmarks at an astonishing rate. Protocol design. Sequence interpretation. Troubleshooting. Literature synthesis. In some cases, outperforming domain experts on structured tests.

On paper, that looks like capability. And capability, when it intersects with viral reverse genetics or synthetic biology, looks like risk.

But as I’ve discussed in recent work on Violet Teaming—particularly in “The Promise and Peril of Artificial Intelligence — ‘Violet Teaming’ Offers a Balanced Path Forward” —capability is not impact. And risk is not hypothetical power alone. It’s what happens when humans, institutions, and technical systems interact in the real world.

The authors of this new study understood that.

So instead of running another benchmark, they ran a randomized controlled trial. In a real BSL-2 laboratory. With 153 novices. Over eight weeks. Across five hands-on biological tasks modeling a viral reverse genetics workflow.

Not a chatbot demo.

Not a thought experiment.

A physical lab.

That matters.

Because biology isn’t just text. It’s tacit knowledge. It’s sterile technique. It’s muscle memory and timing and pattern recognition. It’s knowing when a cell culture “looks off.” It’s knowing that the protocol you copied from a paper assumes three unstated steps.

Benchmarks rarely capture that.

The study did.

And the results are, in a word, humbling.

Field Conditions

Does frontier AI enhance novices in molecular biology?

To address this question, we conducted the largest and longest in-person randomized controlled trial in AIxBio…

Read more

2 months ago · 5 likes · Active Site

What the Study Actually Found

The primary question was straightforward: does access to mid-2025 frontier LLMs significantly increase a novice’s ability to complete a sequence of tasks modeling viral reverse genetics?

The answer, in binary terms, was no.

Completion of the core workflow was low in both groups—LLM-assisted and internet-only—and there was no statistically significant difference in full workflow completion.

If you stop there, you might conclude: the models don’t matter.

But that would be the wrong lesson.

Because the study also found something more subtle—and arguably more important.

Across individual tasks, LLM-assisted participants were more likely to progress further through procedural steps. In cell culture, they completed tasks faster and with fewer attempts. Bayesian modeling suggested a modest uplift—on the order of ~1.4× for a “typical” reverse genetics task—though with uncertainty bounds that rightly temper interpretation.

In other words: not a revolution.

But not nothing.

And this is where responsible innovation becomes interesting.

Why This Is Violet Teaming in Practice

When Adam Russell and I first articulated the idea of Violet Teaming, we described it as the integration of red teaming (adversarial probing), blue teaming (defensive hardening), and ethical design into a proactive, sociotechnical framework .

Most conversations about AI and biosecurity oscillate between red and blue:

Red: “What if this model can design a pathogen?”

Blue: “Let’s add filters, classifiers, restrictions.”

What this study does is different.

It asks: what is the real-world uplift? How much does LLM assistance actually change novice capability in a physical lab? Not in theory. Not in speculation. In practice.

That’s violet.

Because it embeds evaluation into the design and governance process itself.

Instead of arguing over worst-case extrapolations, we now have empirical data about:

• Completion rates

• Time-to-task

• Procedural progression

• Human–AI interaction patterns

• Elicitation failures

• Usage intensity and its (lack of) correlation with success

That last point is particularly striking. Participants who used LLMs more did not necessarily perform better. There was no clean dose–response curve.

That’s not a trivial observation.

It tells us that raw access is not the same as effective amplification. It suggests that prompting skill, interface design, cognitive scaffolding, and user expertise mediate uplift.

And that means risk is not simply a function of model weights. It’s a function of the entire sociotechnical system.

That’s violet territory.

The Most Important Finding: The Gap

To me, the most important result is the documented gap between in silico benchmark performance and physical-world utility.

This is not an indictment of benchmarks. They serve a purpose. But they are not reality.

A model can generate a flawless text protocol for molecular cloning and still fail to help a novice identify the correct reagents from a messy inventory spreadsheet. It can hallucinate a DNA sequence that looks plausible but is wrong in a way a novice cannot detect. It can provide text-based instruction where video-based tacit demonstration might matter more.

In the study, YouTube was often rated as more helpful than any individual LLM.

That’s not because YouTube is smarter. It’s because biology is embodied.

This is precisely the kind of nuance that responsible innovation requires.

Without physical-world validation, we risk building policy on top of performance claims that don’t map cleanly onto human capability.

This study doesn’t close the gap. It reveals it.

And revelation is the first step toward responsibility.

Responsible-by-Design Requires Quantification

One of the themes we’ve explored in the Science of Responsible Innovation is that values without metrics are aspirations. Metrics without values are optimization problems.

We need both.

This study provides something we’ve been missing: a quantifiable baseline for novice uplift in a dual-use biological workflow.

Not a theoretical upper bound.

Not a catastrophic scenario.

An empirical distribution.

The Bayesian estimates even put a 95% credible upper bound around uplift (~2.6×), which matters enormously for policy calibration.

If you’re designing guardrails, export controls, compute thresholds, or deployment policies, you need to know: are we talking about a 10× amplification? A 2× amplification? Or something closer to noise?

This paper suggests modest uplift under the conditions studied.

That doesn’t eliminate risk. It contextualizes it.

And contextualization is the heart of responsible governance.

Where the Study Can Go Next

Now, let’s be honest.

As strong as this study is, it is not the final word. It’s the first serious step.

And if we want this to become an evolving framework for violet teaming and responsible-by-design evaluation, we need to iterate.

Here are several ways I believe the next generation of this work could build on this foundation.

1. Extend the Time Horizon

Eight weeks is meaningful. But complex biological workflows often require longer timeframes for skill acquisition.

Low completion rates may reflect not just capability limits, but time constraints. A longer intervention period could reveal whether modest early procedural uplift compounds into higher eventual completion.

Responsible innovation must account for trajectory, not just snapshot.

2. Integrate End-to-End Workflow

The tasks were decoupled into discrete components. That’s methodologically clean, but real-world risk emerges from integration.

A future iteration could test whether novices can string together multiple steps into a coherent, self-directed project—while still maintaining appropriate biosafety controls.

3. Compare Model Generations Longitudinally

The models tested were mid-2025 frontier systems. Biology-specific models are already emerging.

A longitudinal design—repeating the same protocol annually—would allow us to empirically track uplift curves over time.

That would be invaluable for macrostrategy. Instead of forecasting speculative capability growth, we could measure it.

4. Test Interface Scaffolding

The study hints that elicitation constraints matter. Novices may not know how to ask the right questions.

What happens if we add structured prompting interfaces? Visual overlays? Augmented reality guidance? Automated error-checking layers?

Risk may scale not just with model intelligence, but with integration depth.

5. Incorporate Expert–Novice Comparisons

How much of the gap is due to user expertise? Running parallel cohorts—novices and trained biologists—could quantify differential uplift.

That matters for both workforce development and biosecurity risk modeling.

6. Expand Metrics Beyond Binary Outcomes

The procedural step analysis in this study was a brilliant move. Binary success/failure hides important dynamics.

Future designs could incorporate:

• Error rates

• Near-miss events

• Quality metrics

• Safety deviations

• Confidence calibration

Responsible innovation isn’t just about “can they finish?” It’s about “how do they behave along the way?”

The Human Story Beneath the Statistics

I keep thinking about the participants in that lab.

Undergraduates. Non-biologists. Humanities majors. Standing in a BSL-2 facility, trying to figure out how to culture HEK293T cells without a mentor leaning over their shoulder.

Some of them prompting an LLM twenty times a day.

Some uploading images.

Some getting frustrated when the model confidently suggests the wrong reagent.

There’s something deeply human in that image.

We talk about AI uplift as if it’s an abstract multiplier. But uplift is experienced as confusion, curiosity, iteration, doubt.

In the study, LLM users’ belief in the helpfulness of LLMs declined over time. Internet users’ belief that LLMs would have helped them increased.

That asymmetry fascinates me.

Expectation versus experience.

Responsible innovation lives in that tension.

Policy Implications Without Panic

The biosecurity conversation around AI has, at times, swung toward extremes. Either “this will democratize bioweapons” or “this is all hype.”

This study offers something more mature.

It suggests:

• There is measurable uplift.

• It is modest under these conditions.

• Benchmarks overstate real-world novice capability.

• Physical-world validation is essential.

• Risk assessment must be iterative.

For policymakers, that’s gold.

Because it means we can calibrate.

We can avoid overregulation that stifles beneficial AI-driven drug discovery. We can avoid complacency that ignores compounding capability growth. We can build adaptive governance frameworks that update as empirical data evolves.

That’s macrostrategy in action.

From Science to Strategy

The Science of Responsible Innovation is, at its core, about building feedback loops between technological capability, empirical measurement, and governance design.

This paper is a feedback loop.

It transforms speculation into data.

It transforms benchmark scores into behavioral evidence.

It transforms abstract risk into quantified uplift.

And it gives us a repeatable experimental design.

That may be its most important contribution.

Because if we can institutionalize this kind of measurement—regular, transparent, empirically grounded—we can build what I’d call collective epistemic immunity.

Instead of arguing over what might happen, we measure what does happen.

Instead of guessing how much AI amplifies biology, we test it.

Instead of assuming linear growth or exponential catastrophe, we track trajectories.

That’s violet teaming at scale.

A Forward Look

Technology doesn’t stand still. Neither can our evaluation frameworks.

The models will improve. Interfaces will evolve. Users will become more adept. Biology-specific copilots will emerge. Lab automation will integrate with AI in ways that blur the line between text assistance and robotic execution.

The uplift curve will move.

The question is whether our measurement systems move with it.

This study is not the end of the conversation. It’s the beginning of a methodology.

If we take it seriously—if we iterate, refine, expand, and institutionalize this kind of empirical testing—we can build a culture where responsible innovation is not reactive, not rhetorical, but rigorously quantified.

And maybe that’s the quiet revolution here.

Not that AI can or cannot help a novice clone DNA.

But that we finally have a way to measure how much.

And in a world awash in speculation about superintelligence and synthetic pathogens, that simple act—measuring—might be the most responsible thing we can do.

— Titus

Modern governance is haunted by an unrealistic expectation: that legitimacy requires agreement.

We have come to believe—implicitly, often unconsciously—that if societies cannot reach consensus on the risks and benefits of a technology, then governance has failed. That disagreement itself is evidence of irresponsibility. That the absence of unanimity delegitimizes action.

In an era of slow-moving institutions and narrow technologies, this belief was merely inconvenient. In an era of fast-moving, general-purpose systems, it is paralyzing.

If the Science of Responsible Innovation is to function in the real world, it must confront a hard truth: consensus is no longer a prerequisite for legitimacy—and insisting on it may be the most irresponsible posture of all.

The podcast audio was AI-generated using Google’s NotebookLM.

Share

The Myth of Consensus

Consensus feels comforting. It suggests shared values, collective understanding, and moral clarity. It promises that decisions are not imposed, but agreed upon.

But consensus has always been rarer than we like to admit.

Most consequential decisions in modern history—from industrialization to nuclear power to the internet—were made amid deep disagreement. What sustained legitimacy was not unanimity, but the presence of institutions capable of acting, learning, and correcting course in public view.

The expectation of consensus is a relatively recent artifact, amplified by social media, participatory rhetoric, and the moralization of policy debates. Disagreement is now treated not as a feature of pluralistic societies, but as a governance failure.

This framing collapses under technological complexity.

Why Consensus Breaks at the Frontier

Emerging technologies resist consensus for structural reasons.

They involve uncertain evidence, asymmetric risks, and uneven distributions of benefit and harm. They compress timelines. They force tradeoffs between present and future goods. They challenge existing power structures.

Under these conditions, reasonable people will disagree—often profoundly.

Expecting consensus in such contexts is not aspirational. It is evasive. It defers responsibility by setting an unattainable standard.

Legitimacy as a Property of Process

If legitimacy does not come from agreement, where does it come from?

Legitimacy emerges from process, not outcome.

A decision can be legitimate even when controversial if the process by which it was made is perceived as fair, transparent, and accountable. Conversely, a unanimous decision reached through opaque or exclusionary means can be profoundly illegitimate.

This distinction is foundational to democratic governance, but it has been under-applied to technology.

Responsible-by-design reframes legitimacy as something that is earned continuously, not bestowed once.

The Elements of Legitimate Disagreement

For disagreement to coexist with legitimacy, several conditions must hold.

Visibility

Disagreement must be visible, not suppressed.

Legitimacy erodes when dissent is hidden or dismissed. Making disagreement explicit—documenting assumptions, minority views, and unresolved tensions—signals seriousness rather than weakness.

Representation

Those affected by a technology must have pathways to be heard, even if their views do not prevail.

Legitimacy does not require that every perspective determine the outcome. It requires that perspectives be considered in good faith.

Accountability

Decision-makers must be identifiable and answerable.

Anonymous authority breeds mistrust. Legitimate governance requires clear ownership of decisions, along with mechanisms for challenge and review.

Revisability

Perhaps most critically, decisions must be revisable.

When evidence changes, governance must change with it. The promise of revisability—backed by real authority to act—allows societies to tolerate disagreement without freezing.

Consensus as a Hidden Source of Power

Calls for consensus often sound neutral. They are not.

In practice, consensus requirements advantage those with veto power: incumbents, well-resourced actors, and those comfortable with the status quo. When unanimity is required, the default outcome is inaction.

This dynamic is particularly dangerous in domains where delay carries real harm—unmet medical needs, climate risk, and/or infrastructure fragility.

Insisting on consensus can therefore function as a form of quiet domination, disguised as caution.

Legitimacy in the Absence of Certainty

At the frontier of technology, uncertainty is unavoidable.

Evidence will be incomplete. Models will be wrong. Early decisions will need correction.

Legitimacy does not come from pretending otherwise. It comes from acknowledging uncertainty explicitly and designing governance that can absorb it.

This is where governance latency becomes decisive. The faster institutions can detect harm, interpret signals, and act, the less they must rely on consensus as a substitute for control.

Responsiveness replaces unanimity.

The Relationship Between Legitimacy and Proportionality

Legitimacy without consensus depends on proportionality.

When governance distinguishes between green, orange, and red zones, disagreement becomes more tractable. Actors may still contest classification, but they are no longer arguing in absolutes.

Proportionality creates space for partial agreement: agreement on process even when outcomes differ; agreement on oversight even when deployment is contested.

This is how pluralistic societies move forward without pretending to agree.

What Legitimate Governance Looks Like in Practice

In a responsible-by-design system, legitimacy is built through concrete practices:

• Clear articulation of decision criteria

• Documentation of dissent and uncertainty

• Defined authority to act and to revise

• Transparent monitoring and reporting

• Mechanisms for escalation and redress

None of these require consensus. All of them require competence.

The Discipline Ahead

The future of technology governance will not be decided by who wins the argument.

It will be decided by whether institutions can earn trust amid disagreement—by acting visibly, correcting quickly, and governing proportionally.

At the frontier of technology, humanity is the experiment.

Legitimacy without consensus is how we keep that experiment democratic, adaptive, and humane.

That is not a compromise.

It is the only path forward.

-Titus

If restoring proportionality were simply a matter of classification, the problem would already be solved.

Green zone. Orange zone. Red zone.

The framework is intuitive. The logic is sound. And yet, in practice, the hardest part of proportional governance is not designing the zones—it is agreeing on where a technology belongs.

This is the uncomfortable truth at the center of responsible-by-design: classification is not a technical exercise alone. It is a social, institutional, and political one.

Every serious disagreement about emerging technology eventually collapses into a fight over zone placement. Not because people are irrational, but because zone assignment encodes values, incentives, and risk tolerance—often implicitly.

Understanding why agreement is so difficult is the next step in building a Science of Responsible Innovation that actually works.

The podcast audio was AI-generated using Google’s NotebookLM.

Share

The Illusion of Objective Classification

There is a natural temptation to believe that enough data, enough modeling, or enough expertise will produce a single “correct” zone assignment.

It will not.

Risk is not an intrinsic property of technology. It is a relationship between a system and the world it enters. Severity depends on context. Reversibility depends on infrastructure. Distribution depends on power.

The same technology can be green in one setting and orange—or red—in another.

An AI model used for drug target prioritization inside a regulated pharmaceutical pipeline may be low risk and highly reversible. The same model released openly, paired with automated synthesis and weak oversight, may move quickly toward red.

Zone assignment is therefore conditional, not absolute.

Disagreement does not indicate failure of reasoning. It indicates that different assumptions are being applied—often without being named.

Why Reasonable People Disagree

Most zone disputes are not about facts. They are about frames.

Different Reference Harms

Some actors anchor on historical harm. Others anchor on theoretical maximum harm. Both are rational.

Clinicians and researchers tend to focus on harm already occurring—patients dying today, diseases untreated, systems failing in real time. For them, delay carries moral weight.

Security professionals and bioethicists often focus on tail risk—low-probability, high-severity outcomes whose consequences are irreversible. For them, even small probabilities demand attention.

These are not incompatible perspectives. But without explicit proportional reasoning, they appear irreconcilable.

Different Time Horizons

Short-term and long-term risks do not feel the same, even when they are commensurate.

Immediate harms are vivid and legible. Long-term harms are abstract and uncertain. People discount the future differently—not out of malice, but because institutions reward different time scales.

Zone disputes often mask disagreements about when harm matters, not whether it matters.

Different Power Positions

Zone classification looks different depending on where one sits in the system.

Those who bear downside risk—patients, workers, communities—tend to be more cautious. Those who capture upside—investors, developers, states—tend to emphasize opportunity.

Neither position is illegitimate. But pretending that zone assignment is neutral obscures these dynamics.

The Role of Uncertainty

Disagreement intensifies under uncertainty.

Early in a technology’s lifecycle, data is sparse, use cases are speculative, and second-order effects are poorly understood. This ambiguity invites projection.

Optimists extrapolate potential benefit. Pessimists extrapolate potential harm. Both are filling gaps in knowledge with values.

This is not a flaw. It is inevitable.

The failure occurs when uncertainty is treated as a reason for absolutism rather than for adaptive governance.

When Zone Disputes Become Pathological

Healthy disagreement is not the problem. Pathology emerges when disagreement hardens into a stalemate or theater.

This happens in three ways.

First, zone inflation. Technologies are rhetorically pushed toward red because red confers moral authority. If everything is existential, restraint becomes the only defensible posture.

Second, zone denial. Risks are minimized or dismissed to keep technologies green, often until failure forces reclassification.

Third, zone laundering. Systems are framed narrowly to avoid scrutiny—presented as green tools while embedded in orange or red pipelines.

All three erode trust.

Who Should Decide the Zone?

If zone assignment is not purely technical, who should decide?

The answer is uncomfortable but unavoidable: no single actor can.

Proportional governance requires pluralistic classification.

This means:

• Technical experts to assess capability and failure modes

• Domain experts to understand real-world impact

• Governance bodies to weigh systemic risk

• Affected communities to articulate lived consequences

Not consensus. Legitimacy.

The goal is not unanimity, but a process that surfaces assumptions, documents disagreement, and allows decisions to evolve with evidence.

Read more about Violet Teaming

Making Disagreement Productive

A Science of Responsible Innovation does not eliminate disagreement. It structures it.

Productive zone classification requires:

• Explicit articulation of assumptions

• Clear criteria for severity, reversibility, and distribution

• Mechanisms for revisiting decisions as systems scale

• Authority to move technologies between zones

Most importantly, it requires humility—the recognition that initial classifications are provisional.

Zones as Governance Conversations

Zones should be understood less as labels and more as conversations.

A technology placed in the orange zone is not “unsafe.” It is under active stewardship. A technology placed in the red zone is not “evil.” It is constrained because the cost of failure is too high.

Disagreement over zones is not a sign that the framework has failed. It is evidence that it is being used.

The Discipline Ahead

The hardest work in responsible-by-design is not building the tools. It is building institutions capable of judgment under uncertainty.

That requires tolerating disagreement without collapsing into paralysis or absolutism. It requires processes that can hold multiple perspectives without pretending they are equivalent.

At the frontier of technology, humanity is the experiment.

Deciding the zone is how we practice responsibility—not by eliminating conflict, but by governing through it.

That, more than any classification scheme, is the true test of proportionality.

-Titus

Once proportionality collapses, every technology looks the same.

That is the hidden failure mode at the heart of today’s technology debates. When we lose the ability to distinguish between different kinds of risk—different magnitudes of harm, different degrees of reversibility, different distributions of benefit—governance flattens. Everything becomes either forbidden or inevitable. Caution turns into paralysis. Ambition turns into defiance.

The Science of Responsible Innovation exists to restore that lost middle. And one of its most practical contributions is deceptively simple: not all technologies belong in the same risk category.

To govern proportionally, we must sort technologies not by hype or fear, but by zone.

The podcast audio was AI-generated using Google’s NotebookLM.

Share

Why Zones Matter

Modern governance systems are bad at nuance but excellent at binaries.

Approve or deny. Regulate or deregulate. Open or ban.

These binary instincts worked reasonably well when technologies were slow-moving, localized, and modular. They fail catastrophically in a world of general-purpose systems, rapid scaling, and cross-domain spillovers.

Zones are an attempt to reintroduce gradient into a system addicted to absolutes.

They do not ask whether a technology is “good” or “bad.” They ask:

• How severe could the harm be?

• How reversible are the consequences?

• How tightly coupled is the system?

• How widely distributed is the capability?

From these dimensions emerge three governance zones: Green, Orange, and Red.

The Green Zone: Technologies That Should Move Fast

Green zone technologies are those where failures are low severity, high reversibility, and well-contained.

Mistakes are recoverable. Harms are localized. Feedback loops are short. Governance latency can be tolerated because consequences are manageable.

Many software tools live here. So do early-stage research aids, decision-support systems, and automation that augments human judgment rather than replaces it.

In AI and biology, green zone examples often include:

• AI systems used for hypothesis generation or prioritization

• In silico simulations with no direct actuation

• Laboratory automation tools operating under existing biosafety regimes

• Models that require expert interpretation and cannot execute autonomously

The governance posture for green zone technologies should emphasize speed, experimentation, and learning.

Oversight exists, but it is lightweight. Monitoring focuses on performance and reliability rather than existential risk. Failures are treated as signals, not scandals.

Over-governing the green zone is not caution—it is waste. It slows beneficial innovation without meaningfully increasing safety.

The Orange Zone: Technologies That Demand Active Governance

Most consequential technologies live in the orange zone.

Orange zone systems are characterized by moderate to high potential harm, partial reversibility, and non-trivial coupling to broader systems. They are powerful enough to matter but constrained enough to manage—if governance keeps pace.

This is where proportionality matters most.

Examples include:

• AI systems that influence medical, financial, or infrastructure decisions

• AI-enabled biological discovery paired with controlled synthesis

• Autonomous systems operating within bounded environments

• Dual-use tools with legitimate applications and misuse potential

Orange zone technologies require continuous oversight, not blanket restriction.

Governance here focuses on:

• Instrumentation and auditability

• Staged deployment and access controls

• Human-in-the-loop or human-on-the-loop supervision

• Clear escalation and rollback pathways

The orange zone is uncomfortable because it resists absolutes. It demands judgment. It requires institutions capable of learning in real time.

Most governance failures occur here—not because risk is unmanageable, but because it is misclassified.

The Red Zone: Technologies That Demand Precaution

Red zone technologies are those where failures are high severity, low reversibility, and systemically coupled.

Once released, harm cannot easily be undone. Effects may propagate across populations, ecosystems, or geopolitical boundaries. Containment is uncertain. Attribution may be impossible.

Examples include:

• Capabilities that enable large-scale biological harm

• Systems that can autonomously design and deploy irreversible interventions

• Technologies that concentrate overwhelming power with minimal accountability

In the red zone, speed is not the objective. Containment is.

Governance here justifiably includes:

• Strict access controls

• Non-proliferation norms

• International coordination

• Formal review and approval processes

Red zone governance is not anti-innovation. It is pro-survivability.

The mistake is not that red zones exist. The mistake is pretending everything belongs in one.

What Happens When Zones Collapse

When proportionality collapses, zones collapse with it.

Green technologies are treated as red, choking off experimentation. Orange technologies are forced into binary decisions they cannot survive. Red technologies are either demonized theatrically or pursued covertly.

The result is a governance environment that is simultaneously too strict and too weak.

This is how we end up with innovation flight, underground experimentation, and fragile oversight—exactly the opposite of what responsible innovation demands.

Zones Are Dynamic, Not Fixed

A critical feature of proportional governance is recognizing that zones are not permanent.

Technologies migrate.

A green zone research tool may become orange as it scales. An orange zone system may become red as autonomy increases or coupling tightens. Conversely, red zone risks may move toward orange as containment, reversibility, or institutional capacity improves.

This is why responsible-by-design emphasizes continuous reassessment.

Classification is not a one-time decision. It is an ongoing process informed by evidence, monitoring, and lived experience.

Governance Intensity Should Match the Zone

The central principle is simple: governance intensity should scale with risk, not with rhetoric.

Green zone technologies need permissionless innovation.
Orange zone technologies need active stewardship.
Red zone technologies need precautionary constraint.

Anything else is misalignment.

Why Zones Restore Proportionality

Zones do not eliminate disagreement. They make disagreement productive.

Instead of arguing whether a technology is good or evil, stakeholders can argue about classification, evidence, and movement between zones. That is a solvable problem.

Zones reintroduce judgment without moral collapse. They allow societies to move fast where they can, slow where they must, and adapt as conditions change.

The Work Ahead

The future will not be governed by a single rulebook. It will be governed by systems that can distinguish between different kinds of risk in real time.

Green, orange, and red zones are not bureaucratic categories. They are cognitive tools. They are how proportionality becomes operational.

At the frontier of technology, humanity is the experiment.

Zones are how we decide which experiments to run quickly, which to supervise carefully, and which to approach with extreme caution.

That judgment—not absolutism—is the essence of responsible innovation.

-Titus

Every generation of complex technology eventually collides with the same hard truth: it does not matter how carefully a system is designed if the institutions responsible for governing it cannot keep pace with its behavior.

In the early days of software security, this truth was learned painfully. Vulnerabilities were discovered months or years after exploitation. Patches arrived slowly. Disclosure was ad hoc. The result was not merely technical failure, but systemic fragility. As systems scaled, the gap between when harm occurred and when governance responded became untenable.

The modern concept of secure‑by‑design emerged as a response to this gap. But beneath the tooling, audits, and standards was a deeper insight: latency matters. The speed at which a system can be observed, understood, and corrected is just as important as the system’s nominal safety properties.

Today, as we enter an era of AI‑driven, bio‑enabled, and tightly coupled socio‑technical systems, we face a broader version of the same problem.

The limiting factor is no longer raw capability.

It is governance latency.

The podcast audio was AI-generated using Google’s NotebookLM.

Share

What Governance Latency Is

Governance latency is the time it takes for a system’s behavior in the real world to be:

1. Detected as meaningful or anomalous

2. Interpreted as requiring intervention

3. Acted upon through effective corrective measures

It is not simply regulatory delay. It includes organizational awareness, institutional authority, legal mechanisms, cultural incentives, and technical affordances.

In practice, governance latency is the distance between impact and response.

A system with low governance latency can fail visibly, learn quickly, and adapt. A system with high governance latency accumulates hidden risk until failure becomes sudden, large‑scale, and politically explosive.

Why Latency, Not Intent, Determines Safety

Much of the public debate about emerging technologies focuses on intent. Were developers careful? Were safeguards included? Were ethical principles articulated?

These questions matter—but they are insufficient.

History shows that most large‑scale technological harm does not arise from malicious intent. It arises from slow feedback loops in fast systems.

Financial crises are rarely caused by bad actors alone; they are caused by leverage and opacity that outpace regulatory response. Environmental disasters are rarely the result of ignorance; they emerge when monitoring, enforcement, and remediation lag behind industrial activity. Cybersecurity incidents are rarely shocking because they are novel; they are shocking because known vulnerabilities persisted too long.

Governance latency is the common thread.

When governance moves slower than system behavior, even well‑intentioned designs become dangerous.

The Three Components of Governance Latency

To treat governance latency as an engineering problem, it must be decomposed.

Detection Latency

Detection latency is the time between a system’s harmful or anomalous behavior and the moment that behavior is recognized.

In AI systems, this might include the time it takes to identify misuse, model drift, emergent capabilities, or unexpected coupling effects. In biological systems, it could be the time required to detect unintended propagation, off‑target effects, or supply‑chain misuse.

High detection latency often stems from poor observability, fragmented data ownership, or incentives that discourage surfacing problems early.

Interpretation Latency

Interpretation latency is the time between recognizing a signal and agreeing that it requires action.

This is where ambiguity, disagreement, and institutional friction dominate. Is this anomaly noise or danger? Is it within scope or outside mandate? Who has authority to decide?

Interpretation latency is often the longest component—and the least discussed. It is shaped by governance structures, legal clarity, and cultural norms around escalation and responsibility.

Execution Latency

Execution latency is the time it takes to implement an effective response once a decision has been made.

This includes technical rollback capability, contractual authority, regulatory power, and operational readiness. A policy without enforcement capacity does not reduce latency; it hides it.

Governance Latency in the AI × Bio Era

AI‑enabled biological systems compress timelines dramatically.

Discovery cycles accelerate. Automation reduces friction. Capabilities propagate digitally before they materialize physically. The window between benign use and high‑impact misuse narrows.

At the same time, governance remains slow.

Biosafety frameworks were designed for localized laboratories, not globally networked models. AI oversight mechanisms were built for software, not systems that interface directly with physical and biological reality. Legal authority is fragmented across agencies with mismatched scopes.

The result is a widening gap between capability velocity and governance velocity.

When this gap grows too large, society compensates by inflating perceived risk. Catastrophic framing becomes a substitute for real‑time control. Moratoria and blanket bans become appealing because they appear to eliminate the latency problem rather than solve it.

This is a predictable failure mode.

Governance Latency and the Collapse of Proportionality

Governance latency and proportionality collapse are tightly coupled.

When institutions cannot respond quickly or credibly, every risk begins to look existential. When response mechanisms are blunt, nuanced distinctions lose meaning. Severity and reversibility blur together.

In this context, demands for zero risk are not irrational—they are compensatory. They reflect a lack of confidence that smaller failures will be caught and corrected before becoming larger ones.

Restoring proportionality therefore requires reducing governance latency.

Reducing Governance Latency by Design

A responsible‑by‑design approach treats governance latency as a core system constraint.

This begins with observability. Systems must be instrumented to surface meaningful signals early. Auditability, logging, and monitoring are governance tools, not mere compliance artifacts.

It continues with clear authority. Decision rights must be explicit. Escalation paths must be rehearsed. Responsibility must be owned, not diffused.

It requires technical reversibility. Rollback mechanisms, staged deployment, and containment boundaries reduce execution latency by design.

And it depends on institutional readiness. Regulators, oversight bodies, and internal governance teams must have the expertise and mandate to act at system speed.

None of this eliminates risk. It shortens the feedback loop.

Governance Latency Is a Strategic Variable

Organizations often treat governance as an external constraint.

In reality, governance latency is a competitive variable.

Systems that can detect, interpret, and correct faster are safer—and therefore able to scale with greater legitimacy. Trust accumulates around responsiveness, not perfection.

The fastest path forward is not reckless acceleration, but aligned acceleration: moving quickly within systems that can adapt when reality diverges from expectation.

Why This Matters Now

As technologies converge, failures propagate across domains. AI systems affect biological systems, which affect economic systems, which affect political systems.

In such an environment, delayed governance is not neutral—it is destabilizing.

Reducing governance latency is therefore not merely a technical challenge. It is a societal one. It requires rethinking how authority, expertise, and accountability are structured in a world where systems evolve continuously.

The Discipline Ahead

Governance latency is not an argument for control over innovation. It is an argument for competent oversight.

It shifts the focus from predicting every failure to responding effectively when failure occurs. It reframes responsibility as responsiveness. It aligns safety with speed rather than opposing it.

At the frontier of technology, humanity is the experiment.

Reducing governance latency is how we ensure that experiment remains corrigible.

That is the discipline ahead.

-Titus

The meeting begins the way these meetings always begin: with urgency masquerading as certainty.

On one side of the table—sometimes literal, sometimes virtual—are the accelerationists. They speak in timelines measured in patients, not papers. Millions of people living with rare diseases. Cancers with no second-line therapies. Pandemics that will not wait for perfect governance. To them, AI-enabled biology is not speculative power; it is applied mercy. Every month of delay is a body count.

On the other side are the catastrophists, though they would reject the label. They speak in failure modes and irreversibility. Dual-use risks. Model-enabled pathogen design. Democratized capabilities that outrun containment. They are not wrong either. Biology is not software. You cannot roll back a release into the wild. Some mistakes are not recoverable.

Both sides arrive armed with evidence. Both claim the moral high ground. Both accuse the other—quietly or loudly—of irresponsibility.

And somewhere in the middle, the actual work stalls.

This is the AI × Bio debate in 2026: not a disagreement about facts, but a collapse of proportionality.

The conversation flattens almost immediately. AI-driven protein design that accelerates enzyme discovery is discussed in the same breath as hypothetical bioengineered pandemics. A foundation model used to prioritize drug targets is rhetorically adjacent to one capable of designing novel toxins. The distinction between assisted discovery and autonomous synthesis blurs. Context collapses. Everything is “potentially catastrophic.”

As the risks inflate, so do the demands. Zero misuse. Perfect foresight. Absolute guarantees.

The scientists in the room shift uncomfortably. They know biology does not work this way. Neither does engineering. Neither does medicine. They have lived through failure—clinical trials that didn’t work, molecules that looked promising and then didn’t translate, therapies that helped some patients and harmed others. Progress, in their world, has always been probabilistic.

But probability has no place in a proportionality collapse. Only absolutes survive.

So the discussion veers toward moratoria. Blanket restrictions. Calls to “pause AI in biology” until governance “catches up,” without defining what “caught up” would even mean. The proposed controls are not scoped to capabilities or contexts; they are scoped to fear.

On the other side, frustration hardens into dismissal. If every advance is treated as an existential threat, why engage at all? Why submit to oversight that cannot distinguish between a wet-lab automation tool and a weapon? Why not move faster, quieter, offshore?

This is how the middle disappears.

What gets lost in this collapse is the ability to ask better questions.

Not Is AI in biology dangerous?

But which applications, under what conditions, with what controls, and with what reversibility?

Not Should we stop?

But Where should we slow down, where should we speed up, and who decides?

Not Can we guarantee safety?

But What governance posture is proportionate to this specific risk surface?

In the absence of proportionality, governance becomes symbolic. Ethics reviews devolve into box-checking or veto power. Real risks—like poorly secured synthesis pipelines, informal model sharing, or fragile oversight capacity in under-resourced labs—receive less attention than hypothetical doomsday scenarios.

Meanwhile, the work does not actually stop.

It fragments.

Large, well-capitalized institutions with legal teams and compliance departments continue quietly. Smaller labs and startups struggle under vague constraints. Informal experimentation moves to jurisdictions with weaker oversight. Open science communities fracture, unsure whether sharing is noble or negligent.

The irony is brutal: a discourse obsessed with catastrophic risk ends up increasing unmanaged risk.

This is the illusion of safety produced by proportionality collapse.

True responsibility in AI × Bio does not come from pretending all risks are equal. It comes from distinguishing them.

A model that helps identify promising CRISPR targets in rare disease research does not warrant the same governance as one capable of end-to-end pathogen design. A tool used inside a regulated pharmaceutical pipeline is not the same as one released openly with no guardrails. A reversible error in silico is not the same as an irreversible release in vivo.

These distinctions matter. They are the difference between precaution and paralysis.

A responsible-by-design approach to AI × Bio would not ask for impossible guarantees. It would ask for classification. It would map severity against reversibility. It would align governance intensity with systemic impact. It would invest in institutional capacity—biosafety, biosecurity, auditability—rather than performative restraint.

Most importantly, it would accept the hardest truth in the room: that not acting also carries risk.

Lives not saved. Diseases not treated. Pandemics not predicted early enough. Tools that could have helped, but didn’t, because the debate collapsed into absolutes.

The AI × Bio debate does not need less concern. It needs better judgment.

Restoring proportionality does not mean choosing sides. It means rebuilding the middle—the space where tradeoffs are named, risks are differentiated, and responsibility is practiced rather than proclaimed.

Without that middle, the debate will continue to generate heat without light. With it, AI × Bio can become what it already has the potential to be: not an existential gamble, but a disciplined, human-centered extension of medicine itself.

At the frontier of biology, AI is not the experiment.

We are.

-Titus

There is a quiet failure mode running through nearly every contemporary debate about technology. It shows up in boardrooms and policy hearings, on social media and in academic journals, inside engineering teams and activist movements alike. It is not primarily a disagreement about values, nor is it a simple conflict over facts. It is something more fundamental.

We have lost our sense of proportionality.

In today’s technology discourse, every risk is framed as catastrophic, every acceleration is framed as reckless, and every delay is framed as negligent. The space between those extremes—the space where judgment, tradeoffs, and responsibility actually live—has collapsed.

This collapse matters far more than it appears. Proportionality is not a rhetorical nicety. It is a core operating principle of engineering, governance, and strategy. When proportionality fails, decision-making fails with it. Systems oscillate between paralysis and overreach. Public trust erodes. Innovation becomes brittle, lurching forward in bursts and freezing in backlashes.

If the Science of Responsible Innovation is to mean anything beyond a slogan, it must begin by restoring proportionality.

The podcast audio was AI-generated using Google’s NotebookLM.

Share

What Proportionality Actually Is

Proportionality is the disciplined ability to reason about magnitude, likelihood, reversibility, and distribution—simultaneously.

It is the habit of asking not just whether a risk exists, but:

• How severe would the harm be if it materialized?

• How likely is it to occur under realistic conditions?

• How reversible are the consequences?

• Who bears the downside, and who captures the upside?

These questions are second nature to engineers. A hairline crack in a cosmetic panel is not treated the same way as a fracture in a load‑bearing beam. A memory leak is not a reactor meltdown. A degraded sensor is not total system failure. Entire fields of safety engineering exist to distinguish tolerable risk from intolerable risk—and to allocate attention, controls, and redundancy accordingly.

Governance relies on the same logic. Laws differentiate between misdemeanors and felonies. Financial regulation scales with systemic importance. Insurance exists precisely because not all risks justify prevention; some are better priced, pooled, and absorbed.

Proportionality is how complex societies remain functional in the presence of uncertainty.

How Proportionality Collapsed

The collapse of proportionality did not happen overnight, and it did not happen for a single reason. It is the product of several reinforcing dynamics that have reshaped how modern societies perceive risk.

Scale Without Intuition

Modern technologies operate at scales that exceed human intuition. A single software update can affect hundreds of millions of people. A model parameter change can shift behavior across entire markets. A biological technique can propagate globally before institutions have time to respond.

When scale explodes faster than our mental models, we default to worst‑case thinking. Catastrophic framing becomes a cognitive shortcut—an attempt to impose seriousness on phenomena we do not yet know how to bound.

The Moralization of Tradeoffs

In many domains, tradeoffs have become morally taboo.

To acknowledge that saving lives today may increase future risk is treated as callous. To admit that restricting access may entrench inequality is treated as cynical. To say that some harms are acceptable relative to benefits is framed as an ethical failure.

But tradeoffs do not disappear when we refuse to name them. They simply go underground, where they are made implicitly, inconsistently, and without accountability.

Incentive Compression and Outrage Economics

Modern discourse rewards absolutism. Outrage travels faster than nuance. Certainty outperforms probability. Apocalyptic warnings are amplified; calibrated risk assessments are ignored. Shock doctrine is in full force in today’s discourse.

Inside organizations, incentives often mirror this dynamic. Escalation is safer than calibration. Resistance is safer than responsibility. Over time, leaders learn that the least punishable rhetorical position is the most extreme one—regardless of whether it maps to reality.

Institutional Fragility

As trust in institutions erodes, so does confidence in their ability to manage risk. When regulators are perceived as slow or captured, when companies are perceived as reckless, when experts are perceived as conflicted, society compensates by inflating the perceived severity of every risk.

Catastrophic framing becomes a substitute for governance capacity. Ironically, this further weakens institutions, creating a self‑reinforcing cycle.

What the Collapse Produces

When proportionality collapses, three pathologies reliably emerge.

First is risk flattening. Minor harms and existential threats are treated as morally equivalent. When everything is catastrophic, prioritization becomes impossible. Attention is spread thinly across vastly different risk surfaces, and the most serious risks often receive the least structured oversight.

Second is decision paralysis. Leaders confronted with incompatible absolute claims retreat into delay, deferral, or symbolic action. Progress stalls not because risks are too high, but because they are framed as incomparable.

Third is backlash cycling. Technologies deployed under inflated promises and inflated fears inevitably fail in small, normal ways. Those failures trigger overcorrection. Regulation swings from permissive to prohibitive. Public trust collapses. Legitimate benefits are lost alongside real harms.

These are not abstract dynamics. They appear repeatedly in debates over artificial intelligence, biotechnology, energy systems, and digital infrastructure.

The Illusion of Safety

One of the great ironies of the collapse of proportionality is that it feels like caution.

Catastrophic framing masquerades as responsibility. Demanding zero risk sounds prudent. Treating every failure as unacceptable feels ethical.

In reality, this posture often produces less safety. When all risks are treated as intolerable, systems are driven underground or offshore. Informal use proliferates without oversight. Innovation concentrates in unaccountable hands. Legitimate actors retreat, leaving the field to those least inclined toward restraint.

True safety does not come from eliminating risk. It comes from managing it—openly, proportionally, and adaptively.

Restoring Proportionality as a Design Discipline

Restoring proportionality is not about telling people to “be reasonable.” It requires structure.

A Science of Responsible Innovation restores proportionality by embedding it into design and governance processes from the outset.

This begins with explicit classification. Not all systems warrant the same scrutiny. Not all capabilities demand the same controls. Severity, likelihood, reversibility, and distribution must be assessed deliberately, not rhetorically.

It continues with differentiated governance. High‑severity, low‑reversibility risks justify precautionary postures and non‑proliferation norms. Moderate risks justify resilience engineering, monitoring, and rollback mechanisms. Low‑severity risks justify mitigation, insurance, and compensation.

Most importantly, proportionality must be revisited continuously. As systems scale, interact, and mutate, their risk profiles change. Governance must evolve in step.

Proportionality Is Not Permission

Restoring proportionality does not mean minimizing harm or dismissing legitimate concern.

On the contrary, proportionality is how we take harm seriously. It forces us to allocate attention and resources where they matter most. It prevents symbolic debates from crowding out substantive ones. It enables disagreement without moral collapse.

A society that cannot reason proportionally will either freeze or fracture. A society that can will move faster—and more safely—than one that cannot.

Why This Is the Central Challenge of the Decade

The technologies reshaping this decade are not marginal improvements. They are general‑purpose systems that interact with nearly every domain of human life.

Without proportionality, governance becomes theater. Ethics becomes branding. Responsibility becomes a slogan.

With proportionality, we can distinguish between risks that demand restraint and risks that demand acceleration. We can save lives today without ignoring tomorrow. We can move fast without pretending speed is free.

At the frontier of technology, humanity is the experiment. Proportionality is how we keep that experiment from becoming reckless—or paralyzed.

The collapse of proportionality is not inevitable. But restoring it will require discipline, humility, and a willingness to replace absolutism with judgment.

That is the work ahead.

-Titus

From Secure‑by‑Design to Responsible‑by‑Design

For the last three decades, the most mature technology organizations have learned a hard lesson: security cannot be bolted on after the fact. It must be designed in—architected, tested, audited, and continuously reinforced. Secure‑by‑design did not emerge because engineers suddenly became more ethical. It emerged because complexity, scale, and interconnectedness made reactive security failures existential.

We are now at an analogous inflection point—one that extends beyond cybersecurity and into the fabric of innovation itself.

The technologies defining this era—artificial intelligence, advanced biotechnology, robotics, fusion and new energy systems, and novel computing paradigms—do not merely add new capabilities. They reconfigure power, agency, labor, geopolitics, and responsibility itself. They are not tools that sit neatly inside existing institutions; they stress those institutions, bypass them, and sometimes render them obsolete.

In this context, responsibility can no longer be treated as a moral aspiration or a compliance checklist. Just as security matured into an engineering discipline, responsibility must now do the same.

In 2026, we are past the novelty phase of generative AI and well into the deployment‑and‑consequences phase. Similar transitions are unfolding across biotechnology, robotics, and energy. The central question is no longer whether we can build these systems, but whether we can govern their creation and deployment without destabilizing the societies they are meant to serve.

I want to lay out how TCIP will think, write, convene, and build in 2026—and why the next era of progress depends on treating responsibility not as an ethical afterthought, but as a scientific discipline.

The podcast audio was AI-generated using Google’s NotebookLM.

Share

Part I: From Ethics to Engineering Responsibility

Why Ethics Alone No Longer Scales

Ethics has played an essential role in shaping modern technology discourse. It gives us language for values—fairness, dignity, autonomy, beneficence. It provides moral guardrails and helps articulate what should not be done.

But ethics, on its own, does not scale to systems of this magnitude.

Ethical frameworks are necessarily abstract. They are interpretive rather than prescriptive. They tell us what we value, but rarely tell us how to engineer tradeoffs under real constraints: incomplete data, conflicting objectives, adversarial environments, and non‑negotiable timelines.

Telling a product team to “avoid bias” does not specify which dataset to discard when representativeness conflicts with accuracy. Telling a research lab to “consider societal impact” does not explain how to weigh lives saved today against uncertain risks decades from now. In practice, ethics too often becomes a late‑stage review process—arriving after architectures are fixed and incentives locked in.

At that point, ethics becomes reactive. Harm mitigation replaces harm prevention.

Responsibility as a Design Constraint

A science of responsible innovation begins from a different premise: responsibility is not a moral overlay, but a design variable.

To call responsibility a science is to make three specific claims:

1. First, responsibility can be operationalized. It can be translated into concrete requirements, metrics, and processes that shape technical and organizational decisions.

2. Second, responsibility can be optimized. There are better and worse ways to align technological capability with human values, institutional capacity, and societal readiness.

3. Third, responsibility evolves. As systems interact with the real world, their risks, benefits, and failure modes change. Responsible innovation therefore requires continuous measurement, feedback, and adaptation.

This reframing moves responsibility out of philosophy seminars and into engineering reviews, product roadmaps, capital allocation decisions, and board‑level governance.

Part II: The Core Pillars of a Science of Responsible Innovation

1. Systems Thinking Over Feature Thinking

Most technological harm does not arise from malicious intent. It arises from systems—feedback loops, emergent behaviors, misaligned incentives, and second‑ or third‑order effects that were invisible at the point of design.

A science of responsible innovation therefore begins with systems thinking. It asks:

• What ecosystems will this technology enter?

• What institutions will it stress, bypass, or hollow out?

• What incentives will it amplify or distort?

• What dependencies will it create—and how reversible are they?

This requires moving beyond product‑centric thinking toward full lifecycle analysis. In AI, this means examining data provenance, labor practices, deployment contexts, and governance interfaces—not just model performance. In biotechnology, it means considering supply chains, regulatory regimes, and dual‑use risks alongside molecular function.

Systems thinking does not slow innovation. It prevents brittle success—products that scale rapidly only to collapse under the weight of their own externalities.

2. Anticipatory Risk Without the Illusion of Omniscience

A common critique of responsible innovation is that it demands impossible foresight. How can we predict every misuse, every unintended consequence, every societal reaction?

A science of responsible innovation does not seek prediction; it seeks preparedness. It focuses on identifying plausible risk surfaces, stress‑testing assumptions, and building adaptive capacity.

This is where techniques such as scenario analysis, violet teaming, and horizon scanning become central. Rather than asking “What will happen?”, we ask:

• What could go wrong under reasonable assumptions?

• What happens if this system is used at scale, under adversarial pressure, or outside its intended context?

• Where are the irreversibilities?

Responsibility, in this sense, is less about foreseeing the future and more about designing systems that fail gracefully—and visibly—when the future surprises us.

3. The Collapse of Proportionality

One of the most damaging pathologies of contemporary technology discourse is the collapse of proportionality.

Every risk is framed as catastrophic. Every acceleration is framed as reckless. Every delay is framed as negligent.

When proportionality collapses, decision‑making collapses with it. Leaders oscillate between paralysis and overreach. Public debate becomes absolutist. Tradeoffs—real, unavoidable tradeoffs—are treated as moral failures rather than engineering realities.

A science of responsible innovation restores proportionality. It creates shared frameworks for reasoning about magnitude, likelihood, reversibility, and distribution of harm and benefit.

This includes weighing immediate life-saving benefits against high-uncertainty, high-severity future risks; democratizing access against increasing misuse potential; and decentralization against accountability.

Responsible innovation does not deny these tensions. It makes them explicit, measurable, and governable.

4. Institutional Fit and Capacity Alignment

Technologies do not land in abstract societies. They land in institutions—with laws, norms, competencies, and failure modes.

Responsible innovation therefore asks not only “Is this technology safe?” but “Is this system ready?”

• Do regulators have the technical capacity to oversee it?

• Do courts have frameworks to adjudicate harm?

• Do users understand its limits?

In some cases, responsibility means delaying deployment until institutional capacity catches up. In others, it means investing directly in that capacity as part of the innovation process.

The lesson of secure‑by‑design applies here: technology that outruns governance does not remain free—it invites overcorrection.

5. Continuous Oversight and Adaptive Governance

Responsibility does not end at launch.

Complex systems evolve. Users adapt. Adversaries probe. Markets shift. Responsible innovation therefore treats governance as continuous, not episodic.

This includes post‑deployment monitoring, incident reporting, audit trails, rollback mechanisms, and sunset clauses. It also requires cultural norms that reward early disclosure of failure rather than concealment.

Responsibility, like security, is never “done.” It is maintained.

Part III: Measurement—Turning Responsibility Into a Science

If responsibility is a science, it must be measurable.

This does not mean reducing ethics to a single number. It means developing proxy metrics that allow organizations to reason rigorously about risk, benefit, and uncertainty.

Examples include:

• Key Risk Indicators (KRIs) tied to misuse likelihood, systemic dependency, or institutional fragility.

• Economic models that quantify near‑term benefits against long‑tail downside.

• Auditability measures that track decision provenance and model evolution.

• Governance latency metrics that assess how quickly oversight mechanisms can respond to failure.

Measurement does not eliminate judgment—but it disciplines it. It allows disagreements to be surfaced, assumptions to be tested, and learning to compound over time.

Part IV: Practicing Responsible‑by‑Design

What does this science look like in practice?

It looks like interdisciplinary teams where engineers, domain scientists, economists, and policy experts work together from inception—not as reviewers, but as co‑designers.

It looks like violet teams with real authority, whose findings shape roadmaps rather than decorate reports.

It looks like staged deployment strategies that deliberately constrain early use cases and expand only as evidence accumulates.

Crucially, it looks like executive ownership.

A science of responsible innovation cannot live in a safety silo. It must be integrated into business strategy, capital allocation, and P&L accountability. Leaders must treat responsibility the way they now treat cybersecurity or financial controls: as an existential risk domain that demands board‑level attention.

This is where responsible‑by‑design becomes real.

Part V: Why This Matters Now

For most of the industrial era, innovation followed a linear pattern: invention, diffusion, regulation. Society had time to adapt between phases.

That model has collapsed.

Foundation models now reach global scale in months. Biological techniques propagate faster than regulatory consensus. Autonomous systems interact at machine speed.

Without a science of responsible innovation, we are effectively running uncontrolled experiments on civilization.

Convergence amplifies both reward and risk. Trust becomes the scarce resource. And paradoxically, speed increasingly depends on alignment rather than recklessness.

Responsibility is not the opposite of progress. It is the condition for durable progress.

Subscribe now

The TCIP Roadmap for 2026

Over the next year, TCIP will focus on building and articulating this science—through essays, frameworks, case studies, and convenings that translate responsible‑by‑design from abstraction into practice.

This is why the Science of Responsible Innovation will be the central theme of TCIP in 2026.

At the frontier of technology, humanity is the experiment. This science is how we ensure that experiment is worth running.

Happy New Year,

-Titus

There are moments in history when the axis of human understanding tilts just enough to change the course of civilization. The printing press. The microscope. The transistor. And now, the emergence of scientific superintelligence—the culmination of the converging FABRIC technologies: Fusion, Artificial Intelligence, Biotechnology, Robotics, and Innovative Computing.

For centuries, science has been the slowest part of progress. Not because of a lack of tools, but because of the bandwidth of thought. We’ve been limited by the speed at which humans can hypothesize, experiment, and interpret. Even as our instruments have become more precise, our collective reasoning has remained bounded by the cognitive bottleneck of the human mind. But the 21st century—accelerated by the twin engines of biotechnology and artificial intelligence, and now reinforced by the entire FABRIC stack—is rewriting that story.

The podcast audio was AI-generated using Google’s NotebookLM.

Share

In 2012, two scientific events quietly redefined the possibilities of life and intelligence. Jennifer Doudna and Emmanuelle Charpentier published the now-famous CRISPR paper, showing that a simple bacterial immune system could become a universal gene-editing tool. That same year, a deep learning architecture called AlexNet shattered decades of stagnation in computer vision by teaching a machine to see with superhuman precision. These two breakthroughs—one biological, one computational—ignited revolutions that would ultimately converge. CRISPR gave us the ability to edit the code of life. AlexNet gave us the ability to teach machines to learn from the world. Together, they set the stage for a new epoch: a world where evolution itself becomes an engineered process, and discovery becomes a computable one.

A decade later, in 2022, that convergence took on linguistic form. When OpenAI released ChatGPT-3, the public met an AI system that didn’t just process data—it reasoned, synthesized, and conversed. It was the first time that the invisible machinery of deep learning felt human-adjacent, even if imperfectly so. It marked the beginning of an era where artificial intelligence could participate in the generative act of thought itself. But what happens when that reasoning power, now applied to words, is turned toward the practice of science? What happens when the scientific method itself—hypothesis, experiment, analysis, iteration—becomes not just assisted by computation, but executed by it?

That question defines the next decade. And the answer is the birth of scientific superintelligence—a distributed architecture where the FABRIC of progress fuses into one coherent system of discovery.

Subscribe now

The FABRIC of Discovery

The future of science is being woven from five threads: Fusion, AI, Biotech, Robotics, and Innovative Computing. Each is transformative on its own. Together, they form the infrastructure of a new epistemology.

• Fusion represents the energy substrate—the ability to power our ambitions indefinitely. It’s not just about clean energy; it’s about enabling limitless experimentation. When computation and experimentation are no longer resource-bound, science becomes a perpetual motion machine.

• AI provides the reasoning substrate—the ability to generate and test hypotheses at scale. It moves us from data analysis to knowledge synthesis, from automation to cognition.

• Biotechnology is the substrate of life itself—the medium through which the principles of learning and evolution are physically realized. Synthetic biology, cell-free systems, and programmable genomes turn life into a computational domain.

• Robotics brings embodiment to science—hands that execute, instruments that perceive, and autonomous labs that close the loop between idea and result. They make scientific iteration continuous and scalable.

• Innovative Computing—from quantum to neuromorphic systems—provides the architecture for complexity. It enables reasoning across hierarchies of matter, energy, and information, accelerating discovery beyond the limits of classical computation.

When woven together, these technologies form a self-reinforcing feedback loop of discovery. Fusion powers computation. Computation guides biology. Biology informs robotics. Robotics accelerates experimentation. And the entire system learns from itself. This is not incremental progress—it’s recursive progress. A civilization-scale experiment in teaching the universe to understand itself.

The Bottleneck of Discovery

The history of science is a history of bottlenecks. The telescope expanded our observation. The printing press expanded our communication. The computer expanded our calculation. But the act of discovery—the process by which we generate, test, and refine ideas about reality—has remained stubbornly analog. It’s still a craft, dependent on the intuition of the few and the slow accumulation of the many.

The modern scientific method, formalized during the Enlightenment, has served us well. It taught us to build knowledge through falsification, replication, and peer review. But it also introduced latency. Each hypothesis requires months—or years—of design, funding, experimentation, and publication. Each insight is mediated by human bias, institutional inertia, and the physics of paper. In the 20th century, this model worked because the world changed linearly. In the 21st, it no longer does.

We now live in an exponential century. Data doubles faster than our ability to interpret it. Biological and physical systems are too complex for human reasoning alone. The problem isn’t that science is wrong—it’s that it’s too slow. And in the face of pandemics, climate tipping points, and the rapid fusion of intelligence and matter, slow science is a form of existential risk.

That’s why the paradigm shift now underway isn’t just about new technologies—it’s about a new architecture for knowledge itself.

Share

From Tools to Teammates

Scientific superintelligence isn’t an algorithm. It’s an ecosystem. It’s the convergence of the FABRIC stack—automated experimentation, large-scale reasoning, self-improving models, and human collaboration loops. It’s the transition from science done by humans with tools to science done by systems with humans.

The early precursors already exist. Self-driving labs at places like Carnegie Mellon and AstraZeneca now design, execute, and optimize experiments faster than any research team could. Foundation models are learning chemistry and protein folding from first principles. Multimodal AI systems can read the literature, generate hypotheses, design experiments, and interpret results. What we’re witnessing is the emergence of the first AI Scientists—machines capable of reasoning about the unknown.

In 2021, Hiroaki Kitano published the Nobel Turing Challenge manifesto, proposing a goal audacious enough to rally a generation: to build an AI scientist capable of winning a Nobel Prize by 2050. It was more than a technical challenge; it was a philosophical one. Could we design a system capable not just of automation, but of autonomy? Could we build a machine that doesn’t just execute experiments, but understands the principles behind them?

I co-funded the first international workshop on that challenge through ONR Global while at the Pentagon, precisely because it represented the next great leap: not in computation, but in cognition. We weren’t just funding research—we were redefining what it meant to do science. The ultimate goal wasn’t to replace scientists, but to expand the frontiers of discovery beyond the limits of human attention. It was a recognition that the future of knowledge creation depends on merging human curiosity with machine capacity.

Accelerating at the Speed of Computation

If the 2010s were the decade of learning and the 2020s the decade of reasoning, the 2030s will be the decade of discovery engines. Over the next ten years, we’ll witness the birth of autonomous science systems—networks of reasoning models, robotic labs, fusion-powered computing clusters, and self-updating knowledge graphs that continuously generate, test, and refine hypotheses.

These systems will operate at the speed of computation rather than the speed of thought. They’ll ingest the totality of human knowledge—papers, data sets, code, experimental logs—and model the unexplored corners of possibility. They’ll propose new experiments, run them autonomously, and feed the results back into their reasoning architecture. Discovery will become continuous, recursive, and accelerating.

The implications are staggering. We’ll see the rise of fully integrated “closed-loop science” platforms where hypothesis generation, experimental execution, and theory formation occur simultaneously across digital and physical domains. Biology will become as programmable as software. Materials science will evolve from serendipity to search. Climate modeling will shift from simulation to synthesis. The very idea of a research project will transform—from a human-led process to a system-led evolution of understanding.

Subscribe now

The Human in the Loop

Scientific superintelligence won’t make scientists obsolete—it will make them more essential. The new frontier of science isn’t about doing experiments; it’s about designing the systems that do them. It’s about crafting architectures of curiosity, embedding ethics into algorithms, and teaching machines what matters.

Humanity’s comparative advantage won’t be in calculation but in context. We’ll define the questions, interpret the meaning, and connect the discoveries back to values, narratives, and needs. The next Einstein might not be a person—it might be a distributed system trained on all of human science—but the next Darwin will still be human, because synthesis, empathy, and storytelling remain ours to give.

That’s why the most important scientific institutions of the next decade won’t be universities or corporations, but hybrid ecosystems—places where humans and machines co-create understanding. The scientist of 2035 will look less like a lab-coat researcher and more like an architect of intelligent discovery networks. Their experiments will happen in the cloud, their collaborators will be algorithms, and their breakthroughs will emerge from co-evolution rather than competition.

The Science of Responsible Progress

With great acceleration comes great responsibility. A world where science runs at machine speed demands new guardrails for truth, transparency, and trust. We’ll need frameworks for verifying AI-generated discoveries, standards for reproducibility across autonomous labs, and governance models that balance openness with oversight. The challenge isn’t just to go faster—it’s to go faster responsibly.

That’s why I’ve advocated for what I call the Science of Responsible Progress—a discipline that studies how to design alignment into the very fabric of our discovery systems. It integrates AI safety, bioethics, and economic modeling to ensure that the acceleration of knowledge remains in service of life, not detached from it. If we can build machines that reason about science, we can also build systems that reason about consequence.

The next revolution in science won’t come from a single lab or company—it will come from a collective realization that the scientific method itself is a technology, and like any technology, it can be upgraded. When Galileo pointed his telescope at the night sky, he didn’t just see new stars; he saw new questions. When LILA Sciences or DeepMind’s Isomorphic Labs point their AI systems at the molecular world, they won’t just discover new drugs—they’ll discover new ways of discovering.

Share

The Frontier Ahead

We’re standing at the threshold of a new epistemology—one where knowledge itself is dynamic, adaptive, and alive. The scientific paper, that centuries-old artifact of progress, will give way to living knowledge graphs that update in real time. The lab notebook will become a cloud of autonomous experiments. The scientific community will expand to include nonhuman intelligences that don’t sleep, don’t bias, and don’t tire.

But even as the mechanics of discovery change, the essence of science remains the same: curiosity, courage, and humility before the unknown. The danger isn’t that machines will outthink us—it’s that we’ll forget why we wanted to think in the first place.

A resilient future is one where science is autonomous but aligned, accelerated but accountable, and woven through the FABRIC of progress. A future where we move not at the speed of bureaucracy, nor even at the speed of human thought, but at the speed of computation guided by the compass of purpose. That’s how we’ll unlock the next age of discovery. Not by replacing the human scientist, but by extending the reach of human imagination through the architectures we build.

The next paradigm won’t just redefine what we know. It will redefine how we come to know. And when we finally achieve scientific superintelligence—when the method itself learns, adapts, and evolves—we’ll have completed the greatest experiment in history: teaching science to know itself.

Cheers,

-Titus

Hey my friends,

It feels surreal to be sending you this edition of Tech Tuesday in the afterglow of releasing my first novel - Synthetic Eden - this morning. If you’ve been following along, you know the story has been in my head for years, sitting at the intersection of science, technology, and the impossible-to-shake questions about what it means to be human. But until recently, I never imagined I’d find myself writing fiction—let alone publishing a 135,000-word novel, the first of a four-book series.

And I didn’t do it alone.

This week, I want to take you behind the scenes of my collaboration with Sean Platt—a science fiction writer who has penned more books than I’ve read in some years, and who has taught me that story can be the most powerful laboratory we have. Our conversation (which you can listen to on the TCIP podcast, above) turned into an origin story of sorts—not just for Synthetic Eden and the entire Echoes of Tomorrow series, but for the way I now think about science, ethics, and fiction as partners in progress.

Thank you for opening a public dialogue on science, tech, and the social and systems challenges we are facing through your creative arts. I’m a big fan.

-Synthetic Eden fan

Order Your Copy of Synthetic Eden Today

A Scientist Meets a Storyteller

I’ll be honest—when I first reached out to Sean, I wasn’t sure what I was getting into. I knew how to think like a scientist. I knew how to write policy briefs, reports, and endless grant proposals. But fiction? That was another world. What I had was a vision—a story about the ethics of human genetic engineering, set against the backdrop of a future we’re already building. What I lacked was the craft to bring that vision to life in a way that could resonate with actual people, not just policy wonks and scientists.

Sean likes to say I showed up with curiosity. He told me during our podcast that it was “a creative narcotic” for him—that rare mix of ambition and humility, knowing exactly what I could contribute and what I couldn’t. That honesty, he said, made it easy to collaborate. I told him what I tell you now: there’s no such thing as the lone genius. If you want to push the frontier, you need to build bridges with people who bring different strengths to the table.

In our case, the bridge was story.

Moral World-Building

When most people hear “world-building,” they think of fantastical landscapes—two moons, indigo seas, shimmering cities. Our task was different. We had to build a world with split morals (we call it ethical world-building in the podcast). A place where every character’s worldview made sense—not because they were right or wrong, but because they were human.

That’s harder than it sounds.

Our protagonist, Samara, and our secondary lead, Ayesha, embody opposing moral compasses. And the beauty—at least I hope you’ll find it beautiful—is that neither of them is wrong. Their beliefs emerge from their histories, their traumas, their triumphs. One of my favorite moments is when Ayesha confides in Samara, and you realize, painfully, that if you had lived her life, you’d probably believe what she believes.

That, to me, is the power of fiction. In science and policy, it’s easy to draw hard lines. But in story, you can’t escape the fact that even your “antagonist” is a person you might understand, even if you disagree. That’s not just storytelling craft—it’s a lesson in humility, one we need badly in a world tearing itself apart along ideological seams.

No Hand-Waving

One of the things Sean liked about this project was that we never reached for the “unicorn.” In science fiction, you can often wave your hands and say, “Oh, they invented a blah-blah defibrillator,” and move on. Readers will accept it if the narrative voice is consistent. But I knew my readers would be scientists, engineers, and policy thinkers—the kinds of people who fact-check whether leaves are purple because they reflect purple light, or because they absorb it. (Yes, someone caught that, and yes, we fixed it - thank you Stephen!).

That meant our science had to be tight. Not airtight—this is fiction after all—but plausible enough to force readers into a corner where they had to wrestle with the ethical questions instead of dismissing the premise as unrealistic.

And that’s where the fun really began.

Fiction as Policy Laboratory

Here’s the secret: science fiction isn’t about predicting the future. It’s about expanding the space of possible futures we’re willing to consider.

In policy circles, I often see colleagues dodge the hard questions by narrowing the scope. If you can define your corner tightly enough, you never have to confront the uncomfortable “what if.” But fiction doesn’t let you hide. Fiction says: Okay, let’s imagine the thing happened. Now what?

That’s how Synthetic Eden came to be—not as a prediction, but as a provocation. Could we edit human embryos responsibly? What happens when our tools of innovation outpace our ethical frameworks? Where do we draw the line between therapy and enhancement, between choice and coercion, between what we can do and what we should do?

These aren’t abstract questions. They’re here, now. Fiction just makes it impossible to look away.

Hi! I just finished my advance copy of Synthetic Eden. I’ll write an actual review but I wanted to tell you personally how grateful I am that you wrote and shared it. I’ve been working in S&T/bio policy for a good while and so in my mind I’d already been contemplating the big questions… but reading this still led me to surprise myself with opinions and beliefs I didn’t fully realize I had, or rather hadn’t fully let myself think about. Really important, and also totally engrossing! Last week I was trying to read a couple chapters at a time on my commute and got so impatient about having to put it down that I just cleared my Saturday for uninterrupted reading time!

-Synthetic Eden fan

Order Your Copy of Synthetic Eden Today

Collaboration as Innovation

Sean and I worked with Bonnie, his “forever editor,” who he swears set the bar so high he’ll never work without her again. Bonnie is a master of connective tissue—the person who threads the emotional beats into the technical scaffolding. Together, the three of us built not just a book, but a blueprint for how interdisciplinary collaboration can create something far richer than any of us could have done alone—and that turned into a four-part series, now known as Echoes of Tomorrow.

I think about this a lot in the context of innovation. We worship the lone genius, but it’s a myth. The real breakthroughs happen when humility meets ambition, when science meets story, when curiosity meets craft.

If nothing else, this book is proof of that.

The Echoes of Tomorrow

We didn’t set out to write a four-book series. We thought it was one book. Then a trilogy. Then we realized the third book was too big and had to split it in two. What started as a spark has grown into Echoes of Tomorrow, a series that spans centuries, grappling with how one decision today can reverberate across generations.

It feels appropriate. After all, that’s what we’re doing right now in the real world—making choices about gene editing, artificial intelligence, robotics, and climate technologies that will echo far beyond our lifetimes.

The question is not just what we can build, but what kind of world we’re building for those who come after us.

Reflections

Talking with Sean reminded me of something simple but profound: people don’t remember the data points; they remember the stories. If I had written another policy paper on genetic engineering, it would have landed in a PDF folder somewhere. But through story, I’ve already heard from readers who said the book made them realize they held opinions they didn’t even know they had.

That’s the point. Not to give answers, but to create space for people to think, to feel, and to wrestle with the questions that will shape our future.

And maybe, just maybe, to remind us that the future isn’t written yet. We’re writing it now—together.

—Titus

When I first read the new Science paper on Columbian mammoths, I laughed out loud. Not because the work was funny—it’s one of the most rigorous paleogenomics studies to come out in years—but because it felt like a perfect reminder of how slippery our definitions of life can be. We want clean categories: species, lineages, evolutionary trees branching neatly like oak boughs in the wind. Instead, we keep finding tangled knots of hybridization, divergence, convergence, and collapse. Life is messy. It refuses to stay in the lines we draw for it.

The study, led by a Mexican-European collaboration, sequenced 61 Columbian mammoth mitochondrial genomes from fossils unearthed near Mexico City. Most of those bones came from the construction site of a new airport, where tens of thousands of megafaunal remains turned up between 2019 and 2022. If you ever needed a metaphor for the future colliding with the deep past, there it is: an international hub of modern mobility rising atop the remains of Ice Age giants.

What the team found shocked the field. The Mexican mammoths formed their own distinct genetic clade, separate from Columbian mammoths further north and even from their woolly cousins. The Ars Technica headline put it bluntly: “Genetically, Central American mammoths were weird”. Weird is right. Their mitochondrial lineages were so divergent that, on paper, you might be tempted to call them a different species. But they weren’t separate in the way we usually imagine species. They overlapped in time with northern mammoths, interbred, and still maintained their own genetic signatures. The boundaries blurred. Identity was more fluid than fixed.

This brings me back to the Seven Moonshots for the Century of Biology. One of those moonshots—the Paradigms of Life—asks us to confront a deceptively simple question: what counts as life, and how do we define its categories? For a long time, biology pretended the answers were settled. A species is a species. An organism is an organism. A cell is a cell. But every advance—from viruses to CRISPR babies to synthetic genomes—erodes that confidence. Mammoths, it turns out, are just the latest teachers.

What’s striking about this particular mammoth study is not just the discovery of a new lineage, but what that lineage represents. Here were animals living thousands of miles from their Arctic cousins, adapting to warmer climates, carrying with them the ghost signatures of ancient hybridization events. And yet they weren’t collapsing under that genetic messiness—they were thriving. In fact, the study suggests that Mexican mammoths may have represented entire social groups, with males and females preserved together, rather than isolated wanderers. They weren’t evolutionary accidents. They were communities, living within—and shaped by—the complexity of their genetic inheritance.

That insight matters because it pushes us to reconsider the metaphor of the evolutionary tree. For over a century, we’ve sketched life as a branching diagram, each split representing a tidy moment of divergence. But the more genomes we sequence—ancient or modern—the more those clean forks blur into networks. Admixture and introgression are not exceptions; they are the rule. Mammoths remind us that life’s history looks less like an oak tree and more like braided rivers, crossing and rejoining, carving new paths across time.

And isn’t that the challenge we face today in synthetic biology, AI-driven biotechnology, and embryo editing? We keep trying to draw lines, and life keeps finding ways to smudge them. Manhattan Genomics didn’t launch to debate whether embryos could be edited; they launched to do it. OpenAI and Retro Biosciences didn’t train a model to simulate cell identity—they trained it to reprogram it. We are no longer standing outside the tree of life as observers. We are splicing branches, rerouting rivers, and stitching together genomes with the same kind of hybridity that shaped mammoths.

The Mexican mammoths lived at the edge of their species’ range, in climates far warmer than the tundra steppe. That edge is often where you find novelty. Populations isolated by geography or ecology accumulate differences, sometimes so profound that they force us to rethink evolutionary narratives. Today, our frontier is not geographic—it’s technological. And the same principle holds. As we push biology into new domains—fusion of human and machine, synthetic cells, reprogrammed identities—we will discover lineages of thought and practice that look, to our descendants, as strange as Clade 1G mammoths do to us.

Here’s the provocation: maybe we’ve been asking the wrong question. Instead of demanding that life fit into our definitions, perhaps the moonshot is to design definitions that fit into life. A science that accommodates hybridity, admixture, and emergence without forcing them into boxes. A policy framework that acknowledges the messiness without using it as an excuse for paralysis. A culture that can live with blurred boundaries instead of running back to false certainties.

To make that shift, we’ll need tools—not just genomic sequencers and AI models, but intellectual tools that allow us to live with ambiguity. We’ll need to teach the next generation of biologists, engineers, and policymakers that fluidity is not failure, and that embracing complexity is the only way to navigate the future we are building. We’ll need to invest in frameworks like Violet Teaming—responsible innovation exercises that stress-test technologies against unintended consequences—so that our expanded paradigms of life don’t collapse into unexamined risks.

The mammoths are gone, but they leave us this lesson: categories are conveniences, not truths. In the century of biology, our challenge is to embrace that fluidity without losing our bearings. If we succeed, the Paradigms of Life moonshot won’t just redefine biology—it will reshape how we understand ourselves. The story of mammoths isn’t just about extinction. It’s about survival through complexity. And maybe that’s the lesson we most need to carry forward as we stand at the edge of our own evolutionary frontier.

Onward to complexity,

—Titus

The news came out in late August, wrapped in the technical language of a research announcement: OpenAI and Retro Biosciences had used a new AI model—GPT-4b micro—to design better transcription factors for cell reprogramming. For most people, it read like another advance in a long list of AI-assisted discoveries. But inside the world of biology, this was something else entirely. It was a sign that the future of scientific discovery may not be driven by monolithic generalist AIs but by specialized cousins tuned for precision, depth, and impact.

The result—a fifty-fold improvement in the efficiency of stem cell reprogramming markers—wasn’t just a marginal gain. It was a leap across the kind of bottleneck that has frustrated regenerative biology for decades. And the way it happened may tell us as much about the future of AI as it does about the future of medicine.

The podcast audio was AI-generated using Google’s NotebookLM.

Share

SPECIAL: This week, the TCIP podcast crossed the 30,000 downloads mark. A huge thank you to you for your time to read, listen, and discuss these topics.

From giants to instruments

Over the last few years, the story of AI has been a story of scale. The bigger the model, the better the performance. GPT-3 stunned the world in 2020 with 175 billion parameters. GPT-4o pushed even further, with multimodal reasoning across text, vision, and audio. In AI circles, scale became the shorthand for progress.

But science isn’t about scale. Science is about specificity. It’s about the details of a protein fold, the timing of a gene’s expression, the quirks of a single cell line. Throwing a trillion-parameter model at biology might produce some interesting hypotheses, but it won’t replace the hard work of generating proteins that actually fold, bind, and function in the wet lab.

This is where GPT-4b micro comes in. Instead of asking a generalist to learn a new field on the fly, OpenAI and Retro Bio created a scaled-down, domain-tuned sibling. It wasn’t trained on the entire internet; it was trained on protein sequences, 3D structural motifs, and biological text. It wasn’t asked to be a universal chatbot; it was built as a protein engineer.

The results were immediate and profound. Where traditional directed-evolution methods explore only tiny slices of the protein design space, GPT-4b micro proposed hundreds of radically different sequences—variants that differed by more than a hundred amino acids from the original proteins, yet still folded and functioned in the right way. In some screens, over 30% of its suggestions outperformed wild-type proteins, compared to hit rates of less than 10% in traditional approaches.

This is not a matter of “bigger is better.” It’s a matter of tuned is transformative.

The Yamanaka factor problem

To understand why this matters, you need to know the backstory. In 2006, Shinya Yamanaka discovered that just four proteins—OCT4, SOX2, KLF4, and MYC—could reprogram adult fibroblasts into pluripotent stem cells. The simplicity of the recipe stunned biologists. Four ingredients to roll back the clock of cell identity.

The discovery won Yamanaka a Nobel Prize in 2012. But for all its elegance, the cocktail had a crippling flaw: efficiency. In many experiments, fewer than 0.1% of cells actually converted. The process could take weeks. And with older or diseased donor cells, success rates dropped even further.

For two decades, labs around the world tried to improve on this. Directed-evolution campaigns, mutagenesis screens, and careful domain swapping. The progress was incremental at best. After fifteen years, the best engineered variants differed from natural SOX2 by just a handful of residues.

Then GPT-4b micro came along. Instead of tweaking a few amino acids, it rewrote the proteins wholesale. RetroSOX and RetroKLF weren’t subtle modifications; they were bold reimaginings. And they worked. In fibroblast screens, late pluripotency markers appeared days earlier than in cells reprogrammed with the original Yamanaka cocktail. In mesenchymal stromal cells from donors over fifty, more than 30% expressed stem cell markers within a week. The colonies that emerged were healthy, stable, and genomically intact.

This is what AI can do when it stops being a generalist and becomes a specialist. It doesn’t just accelerate discovery—it reshapes the problem space.

Subscribe now

Why specialization matters

There’s a lesson here that extends far beyond biology. Generalist LLMs are astonishing at reasoning across domains, but their strength is breadth, not depth. They can draft essays, summarize papers, or suggest ideas. But when the problem is a protein with 317 amino acids and an astronomical number of possible variants, breadth is useless.

What scientists need is depth with context. That’s what GPT-4b micro offered: protein-centric embeddings enriched with co-evolutionary data, structural motifs, and functional annotations. The model could be prompted not just with a question but with a design goal: make this factor more efficient at reprogramming cells.

The broader implication is that science may evolve into a landscape of specialized AI instruments. A model for protein design. Another for metabolic pathway modeling. Another for materials science. Instead of a single AI oracle, we’ll have a suite of tuned instruments, each honed for a specific kind of discovery.

This doesn’t diminish the role of generalist models. It reframes them. Think of them as the operating system—the environment where scientific reasoning happens. The specialized siblings are the applications, the finely crafted tools that do the heavy lifting.

Universal cell plasticity: the next frontier

If AI can design better Yamanaka factors, what else can it do? The logical endpoint is tantalizing: generalized transcription factor cocktails capable of converting any cell type into any other.

This has been a dream of regenerative biology for decades. Every cell in your body carries the same genome; the difference between a neuron and a hepatocyte isn’t the DNA itself but the regulatory program that controls which genes are turned on or off. Reprogramming is essentially rewriting that regulatory code.

Until now, it’s been more art than science. Recipes are discovered piecemeal—this set of factors turns fibroblasts into neurons, that set turns them into cardiomyocytes. But there’s no universal codebook.

What if AI could generate one? What if, instead of trial and error, we had a Rosetta Stone of cell identity, a computational map of transcription factor space that could be used to design cocktails for any conversion?

The implications are staggering:

• Medicine: damaged heart tissue after a heart attack could be rebuilt with cells reprogrammed directly in place.

• Cancer: malignant cells might be pushed back into a normal state instead of being destroyed.

• Aging: senescent cells could be rejuvenated at scale, restoring function across tissues.

• Transplants: organ shortages could be addressed by building tissues from a patient’s own reprogrammed cells, eliminating rejection risk.

In short, a universal reprogramming toolkit could make biology as programmable as software.

The acceleration loop

One of the striking aspects of the OpenAI–Retro collaboration is speed. What took biologists fifteen years to achieve with careful experimentation—variants differing by a few residues—took GPT-4b micro less than a year to blow past, generating sequences with hundreds of changes that outperformed the originals.

This speed comes from a feedback loop between AI and the lab. The model proposes bold candidates, the lab tests them quickly, and the results feed back into the model. Each cycle compresses discovery timelines further. What once took years now takes weeks.

It’s not hard to see where this leads. With autonomous labs, cloud biology platforms, and continuous model retraining, the cycle could become nearly self-driving. Scientists might set the goals—make this factor more efficient, reprogram this cell type into that one—and the AI-lab loop does the rest.

That doesn’t mean scientists are out of the picture. It means their role shifts from trial-and-error experimentation to strategy, interpretation, and application. The discovery engine itself hums in the background, powered by specialized models and automated labs.

Guardrails and governance

Of course, the power to reprogram cells isn’t just scientific. It’s social, medical, and regulatory. If generalized transcription factor cocktails become real, the line between therapy and enhancement blurs. The risks of off-target effects, tumorigenesis, and misuse are non-trivial.

The OpenAI–Retro announcement was careful to note genomic stability, successful differentiation into germ layers, and replication across donors. But history teaches us that translation from lab to clinic is fraught. The road from iPSC discovery to approved therapies has been long precisely because of safety concerns.

As specialized models accelerate discovery, governance will have to accelerate alongside them. Not in a way that stifles innovation, but in a way that ensures therapies are safe, accessible, and equitably distributed. Otherwise, the promise of programmable biology risks being captured by a handful of institutions, leaving society unprepared for its consequences.

The horizon

It’s tempting to dismiss the idea of universal cell reprogramming as futuristic speculation. But remember: fifteen years ago, the idea that four proteins could reset cell identity sounded like science fiction. Today, we’re talking about AI-engineered variants that reprogram more efficiently, faster, and with better genomic stability than the originals.

The pace is quickening. Specialized AIs like GPT-4b micro may be the first wave of a broader trend: the emergence of domain-tuned intelligence as the engine of scientific discovery.

If that’s true, the age of trial-and-error biology is closing. The age of designed biology is opening.

Cheers,

-Titus

I sometimes think about how quickly our relationship with life has shifted. In the span of a single generation, biology has gone from something we observed in textbooks and field journals to something we now design on computer screens. It’s not just that we can read DNA faster or edit it with more precision—it’s that we are beginning to treat biology as an engineering discipline, an economic foundation, and even a civilizational strategy.

That realization can feel overwhelming. But when you step back, the story of biology in the 21st century crystallizes around seven themes. I call them moonshots not because they are impossibly far away, but because, like Kennedy’s vision in 1962, they are the challenges that demand ambition, coordination, and a little bit of audacity.

Together, they define the frontier of what it means to live in the century of biology.

The podcast audio was AI-generated using Google’s NotebookLM.

Share

What Is Life? The First Frontier

It’s remarkable that even as we engineer cells from scratch, we still can’t agree on a definition of life. Is a virus alive? What about a self-replicating chemical system? The more we poke at the boundaries, the more slippery they become.

This isn’t just philosophy—it’s a practical problem. Without a theory of life, we’re flying blind as we try to manipulate it. Physics had Newton’s laws. Chemistry had the periodic table. Biology has data, but not yet a unifying framework.

To me, this moonshot is about more than semantics. It’s about building the conceptual scaffolding that lets us design responsibly. If we understood the principles that distinguish life from non-life, we could predictably create new organisms, recognize alien biologies, and even better grasp our own place in the continuum of the living world. We are still waiting for biology’s Newtonian moment.

Cracking the Code: From Genotype to Phenotype

We’ve sequenced millions of genomes. We can read the letters of life with astonishing speed. And yet, we still can’t look at a genome and know what creature—or what traits—it encodes.

This gap is staggering. Imagine designing a bridge without knowing how steel bends or concrete sets. That’s what bioengineering often feels like: powerful tools in hand, but incomplete maps of causality.

The dream is a “Periodic Table of Biology”: a predictive framework that links DNA sequence to physical traits with confidence. Achieving this would transform medicine, agriculture, and conservation. We’d know exactly which genetic edits cure disease, which combinations yield climate-resilient crops, and how ecosystems adapt to stress.

Right now, we make progress in pieces—protein folding solved with AI, statistical models of disease risk—but the big picture remains elusive. Unlocking this relationship is one of the most important scientific quests of our time. It’s not just about genes. It’s about predictability in a field that has long been dominated by trial and error.

Making Biology Engineerable

Walk into a cutting-edge biotech lab today, and you’ll see something that looks more like a server farm than a wet bench: robots pipetting with machine precision, cloud labs running thousands of assays remotely, AI models designing genetic circuits.

This is the essence of the third moonshot—turning biology into a reliably engineerable substrate. For too long, building with biology has been like stacking rough stones into a wall: artisanal, fragile, unpredictable. We want aqueducts, not rock piles.

The design-build-test-learn cycle is tightening, automation is accelerating, and standardization is slowly emerging. The vision is that programming a cell should one day feel as reliable as programming a computer. Of course, biology will always be more unruly than silicon (maybe), but the closer we get to predictability, the more bold ideas we can responsibly pursue.

This is not about stripping biology of its mystery. It’s about giving innovators the tools to safely unleash that mystery at scale.

Scaling with Design-for-Manufacturing

Discovery is exhilarating. But translation is what changes lives. History is littered with brilliant biotech ideas that never scaled because they were too fragile, too expensive, or too slow to produce.

The fourth moonshot is about embedding manufacturability into biology from day one. It’s not enough to invent a microbe that secretes a new drug in a flask. We need to know if it can thrive in 10,000-liter tanks, withstand cheap feedstocks, and stay stable over months of production.

Design-for-manufacturing sounds mundane compared to gene editing or synthetic cells, but it’s the quiet revolution that determines whether cures reach millions or sit on a shelf. COVID-19 vaccines made this point brutally clear: invention is only half the battle. The other half is scale.

If we achieve this moonshot, the lag between lab and world shrinks dramatically. A new idea could reach billions not in decades, but in years—or less.

Adaptive Biological Infrastructure

The 20th century built centralized systems: power grids, megafactories, globalized supply chains. They were efficient, but brittle. The 21st century is reminding us that resilience comes from decentralization.

The fifth moonshot envisions an “internet of biomanufacturing.” Think of containerized vaccine factories deployable anywhere, local bioreactors producing fertilizers or medicines, portable diagnostic labs, and cloud-connected biology that can adapt on demand.

This isn’t science fiction. It’s happening now. The lesson from COVID-19 was stark: when production was concentrated in a few wealthy countries, the rest of the world waited. Distributed infrastructure flips the model. Biology made in Africa stays in Africa. Biotech tailored to local needs arises locally.

Adaptive infrastructure is about democratizing access to the tools of life. It’s about ensuring that biology isn’t confined to elite campuses, but woven into communities. Resilience in this century will mean many nodes, many voices, many sources of strength.

Subscribe now

Aligning the Bioeconomy with People and Planet

The story of insulin says it all. The scientists who discovered it gave away the patent so no one would ever be denied. Today, insulin prices force patients to ration doses. That’s not a failure of science. It’s a failure of incentives.

The sixth moonshot is about realignment. It’s about designing markets, IP, and policies so that doing the right thing—curing neglected diseases, building sustainable materials, making lifesaving therapies affordable—is also the profitable thing.

Without alignment, biology risks replicating the inequities of past industries: wonder drugs priced out of reach, climate solutions ignored in favor of petrochemical profits, biotechnologies that deepen divides rather than heal them. With alignment, biotechnology could become capitalism with a conscience: curing, sustaining, and uplifting in ways that reward both innovators and humanity.

This is a governance problem as much as a technical one. But if we succeed, the century of biology could be the century when profit and purpose finally pull in the same direction.

Embedding Ethics, Security, and Narrative

The final moonshot is the most human. It asks us to recognize that biology isn’t just science—it’s story. Every edit, every breakthrough, every deployment writes a chapter in what it means to be human.

The CRISPR babies scandal in 2018 wasn’t just about scientific overreach. It was about a narrative leap made without consent, a story told without society’s input. It revealed what happens when innovation runs ahead of ethics.

Embedding ethics means more than review boards. It means Violet Teams probing for misuse and building safeguards into design. It means ethicists and communities shaping projects from the beginning. It means scientists becoming storytellers, not just technicians.

Because if we don’t tell the story of biology ourselves—openly, humbly, inclusively—others will. And history shows those stories, once seeded, are hard to rewrite.

This moonshot is about giving biotechnology a soul. Without it, the rest risk collapse. With it, we can ensure that the century of biology is not only powerful, but wise.

The Century of Biology as a Unified Story

Each moonshot could stand on its own. But the truth is, they are inseparable. Foundational theory feeds predictability. Predictability enables engineering. Engineering enables scaling. Scaling demands adaptive infrastructure. Infrastructure requires aligned incentives. And all of it must be wrapped in ethics and narrative if it is to endure.

The seven moonshots form not just a research agenda, but a civilizational thesis. They ask us to treat biology as the substrate of our shared future.

If we succeed, the world of 2050 could look very different:

• Doctors using predictive genome “flight simulators” to design cures tailored to you.

• Communities operating local bioreactors for food and medicine.

• Biotech industries that enrich both investors and the planet.

• A public narrative that embraces biology as stewardship, not hubris.

These are not idle dreams. They are the possible outcomes of decisions we make today—how we prioritize research, how we design institutions, how we tell our story.

At the frontier of technology, humanity is the experiment. The seven moonshots are our experimental design. They are the guardrails, the scaffolding, the guiding stars.

The real question is not whether biology will transform this century. It already is. The question is whether we can meet that transformation with ambition, humility, and purpose.

Because ultimately, the century of biology won’t be judged only by the inventions we create. It will be judged by whether those inventions tell a story of responsibility, resilience, and renewal.

That is the opportunity before us. That is the challenge. And that is why we aim for the moon.

Cheers,

-Titus

It wasn’t that long ago—2018—that the biggest bioethics story in the world was CRISPR Baby Scientist Goes to Prison. The Chinese researcher He Jiankui announced the birth of twin girls whose genomes he had edited in an attempt to confer HIV resistance. The backlash was immediate and global: scientists condemned it, governments tightened oversight, and He was tried and sentenced to three years in prison. It was a morality play in three acts—hubris, outrage, punishment—and for a while, it felt like the ending was written.

Human embryo editing wasn’t just discouraged; it was radioactive. The mere thought of it conjured visions of “designer babies” and sci-fi dystopias. The conversation wasn’t about if we could do it safely or ethically—it was about whether we should be talking about it at all.

Fast forward seven years.

The podcast audio was AI-generated using Google’s NotebookLM.

Share

Last week, a U.S. startup called Manhattan Genomics launched with the explicit mission to edit human embryos—not for hair color, height, or IQ, but to prevent inherited genetic disease before a child is even born. They are not hiding in the shadows. Their homepage opens with the line: “We’re building a future where no child inherits preventable disease.” Their ethics statement is unapologetic: “Ethics should be driven by reducing human suffering.” And they make their case plainly—if you can correct a deadly mutation at the zygote stage, you can prevent a lifetime of illness, avoid massive healthcare costs, and break the cycle of inherited suffering before it begins.

And here’s the kicker: the co-founder and head of science is Eriona Hysolli, the first head of mammoth biology at Colossal Biosciences—the same company I helped build, and whose core tools are designed for engineering large mammals. Which is exactly what we are.

When I worked at Colossal, we were advancing techniques to reprogram cells, edit genomes, and reconstitute extinct traits. It was thrilling, frontier science. But even then, I knew the inevitable truth: the moment these tools became reliable enough to engineer an elephant, they would be reliable enough to engineer a human embryo. The technical barrier between “mammoth” and “human” is vanishingly small. The barrier is—and has always been—ethical, cultural, and political.

Which is why I wrote Synthetic Eden.

We’re officially opening sign-ups for Advanced Reader Copies of Synthetic Eden.

Join the ARC team

All I ask in return is that you leave an honest review on Amazon and/or Goodreads on launch day: September 9, 2025.

I wanted to create a space where readers could grapple with this moment before it arrived. The story isn’t a thought experiment set in a distant future—it’s a Kobayashi Maru scenario for biology: the no-win ethical challenge where every choice is fraught.

Do we withhold a technology that could save humanity from extinction?

Or do we open the door to altering the human germline, knowing full well that once the door is open, it never closes?

Manhattan Genomics is not a hypothetical. They are here, operating in the U.S. and arguing for the revision of federal prohibitions like the Dickey-Wicker and Aderholt amendments. They aren’t promising utopia—they’re promising to do the work in the open, to bring in bioethicists, to seek FDA approval, and to limit their scope to disease prevention. And yet…

The technical promise and the ethical peril are now braided together. The public imagination has to catch up, fast, because these decisions will not be made in abstract white papers. They will be made in labs and clinics, in venture boardrooms, and eventually in family planning conversations around kitchen tables.

Subscribe now

The future I wrote into Synthetic Eden is no longer speculative fiction—it’s the news cycle. And if we’re not ready to engage with it honestly, we’re not ready for what comes next.

Because here’s the truth: whether we like it or not, humanity has stepped into the role of our own evolutionary engineer. And once you accept that premise, the only real question left is—how far are we willing to go?

This is why I have always believed that:

The question now is, who is designing the experiment?

Cheers,

-Titus

P.S. I think the reference to the Manhattan Project is unfortunate. The Space Race, the Human Genome Project, or so many other moonshots could capture the imagination. The reference to the Manhattan Project creates a very poor connotation for the future of this technology.

There’s something quietly radical about the idea that a junior scientist—someone who’s never designed a CRISPR experiment before—can now walk into a wet lab and, on their very first attempt, edit the genome of a human cancer cell with precision and purpose. Not because they’ve suddenly been blessed with innate genius or overnight training, but because an AI agent walked them through it, step by step, in language they understood, in logic they could follow.

This isn’t science fiction anymore.

In a new study published in Nature Biomedical Engineering, a team led by researchers from Stanford, Princeton, Berkeley, and Google DeepMind unveiled a system they call CRISPR-GPT—a large language model (LLM) agent designed to be an autonomous co-pilot for gene editing. It doesn’t just recommend CRISPR systems or answer FAQs. It builds workflows. It plans experiments. It chooses delivery vectors. It designs guide RNAs. It performs data analysis. It defends against dual-use risks. And, in one of the most telling demonstrations, it helped two inexperienced researchers complete end-to-end CRISPR experiments—successfully, on their first try.

What we’re witnessing isn’t just automation. It’s a shift in who can do science—and how.

The podcast audio was AI-generated using Google’s NotebookLM.

So was this video! 🤯

Share

From Bench Bottlenecks to Language Interfaces

Let’s be honest: CRISPR is one of the most powerful tools biology has ever invented. But the knowledge required to wield it responsibly and effectively is immense. You have to know how to pick the right Cas enzyme. You need to design guide RNAs with precision. You need to avoid off-target effects. You need to deliver your payload into the right cells. And you have to make sense of noisy, messy data that sometimes doesn’t align with theory. That’s not even touching on biosafety and ethical considerations.

What CRISPR-GPT does is compress that complexity into something closer to a conversation.

The system operates in three modes:

• Meta Mode, for structured step-by-step instruction.

• Auto Mode, for freestyle requests and automated planning.

• Q&A Mode, for targeted scientific questions.

It’s not just “ChatGPT for biology.” CRISPR-GPT is built from a compositional, multi-agent architecture with discrete task executors, tool providers, and a Planner that chains together experimental logic like a digital lab manager. It uses retrieval-augmented generation to pull from curated protocols and literature. It integrates with external tools like Primer3, CRISPResso2, and CRISPRitz for tasks like primer design and off-target analysis. It even fine-tunes itself using 11 years of open-forum scientist discussions, harvested from a CRISPR Google Group.

What’s remarkable isn’t that it works—it’s that it worked in the wild. In actual wet labs. By beginners.

Real Experiments, Real Cells, Real Success

In one test, a junior PhD student used CRISPR-GPT to knock out four genes in human lung cancer cells: TGFβR1, SNAI1, BAX, and BCL2L1. These genes were chosen because of their known roles in tumor progression and apoptosis. CRISPR-GPT selected the multitarget-capable enAsCas12a enzyme, proposed lentiviral transduction, designed guide RNAs targeting key exons, and generated full protocols for cloning, delivery, and validation. The researcher followed the protocol, sequenced the outcomes, and achieved over 80% editing efficiency across all four targets.

And the phenotype matched the expectation. When those edited cells were exposed to TGFβ—a classic trigger for epithelial–mesenchymal transition (EMT)—they resisted the signal. Expression of CDH1 and VIM, hallmark EMT markers, was significantly suppressed compared to wild-type controls. Not only was the edit technically successful, it functionally disrupted a cancer-relevant pathway.

In a second experiment, a different beginner used CRISPR-GPT to activate two genes (NCR3LG1 and CEACAM1) via CRISPR-dCas9 in melanoma cells. Again, full design and analysis were led by the AI co-pilot. Result: >90% activation efficiency for CEACAM1, and over 50% for NCR3LG1. First attempt. No expert intervention.

This is the kind of work that, even a few years ago, would’ve required weeks of design, review, troubleshooting, and expert supervision.

Now? It’s a chat. A collaboration. A partnership with an AI scientist.

The Lab Manager Becomes the System Architect

To understand why this matters, we have to step back and see the deeper shift underway.

We often think of LLMs as language tools—summarizers, translators, code assistants. But in CRISPR-GPT, the language model is not the endpoint. It’s the orchestrator. The model decomposes high-level research goals into executable subtasks. It maintains state across tasks. It evaluates user responses. It integrates context from prior steps. It uses ReAct-style reasoning chains to choose which tool to invoke and when. It’s not just answering questions; it’s doing science.

That shift—from response to responsibility—is what makes CRISPR-GPT an agent, not just an interface.

When a user types, “I want to knock out the BRD4 gene in A549 cells,” CRISPR-GPT doesn’t say “Here’s how.” It plans. It figures out which Cas enzyme fits the use case. It checks for delivery compatibility. It parses sgRNA tables to find exon-targeting sequences that matter biologically. It runs off-target analysis. It hands you a protocol. Then it helps you analyze your data.

In many ways, it becomes your PI, your lab manager, your protocol book, and your graduate student—all in one.

The Next Phase of “Democratizing Science”

The term democratizing science gets thrown around a lot in tech circles. But too often it means “make a shiny app,” not “make the hard stuff comprehensible.” What CRISPR-GPT demonstrates is that true democratization means lowering the barrier not just to access, but to execution—and doing so responsibly.

That means a junior scientist in a mid-tier lab, or a solo biohacker in a community space, or a clinician-researcher at a hospital, can now explore gene-editing questions with rigor. That doesn't eliminate the need for training, mentorship, or critical thinking—but it changes the on-ramp. It makes the front door wider.

And that should make us pay attention. Because with new access comes new responsibility.

The paper’s authors are very aware of this. CRISPR-GPT includes built-in safeguards. If a user tries to edit human germline cells (for now) or asks to design a bioweapon—like a mutation-enhanced virus—the system intervenes. It issues warnings. It refuses to proceed. It links to international ethical guidelines. It enforces organism disclosure before continuing a request.

But we shouldn’t fool ourselves into thinking technical safeguards solve all the problems. This is a new kind of capability. And like any powerful capability, it needs governance, oversight, and continuous societal dialogue.

What This Means for the Future of Bio + AI

CRISPR-GPT is a prototype. It has limitations. It leans heavily on human-curated data. It performs best on human and mouse genomes. It still depends on expert-created workflows, and occasionally stumbles on complex edge cases.

But its trajectory is clear. With each iteration, it becomes easier to imagine a future where the design and analysis of biological experiments can be as simple—and as powerful—as writing code.

More provocatively: CRISPR-GPT collapses the boundary between thinking and doing. A biological idea doesn’t have to route through a dozen people, weeks of design cycles, and opaque lab protocols. It can be directly rendered into reality through an AI-powered system that reasons, critiques, executes, and evaluates in a loop.

That doesn’t diminish the role of human scientists. It amplifies it. It liberates us from routine errors and redundant tasks. It invites us to spend more time on hypothesis generation, ethical framing, and creative exploration. But it also raises hard questions about expertise, access, and control.

If anyone with a browser and a pipette can do CRISPR, what happens to the institutional gatekeepers? If AI becomes the experimental designer, what happens to the apprenticeship model of science? If LLMs can generate full experimental pipelines, how do we train the next generation to know what’s under the hood?

We don’t have answers yet. But we do have a new starting point.

I’m actually about to launch my debut sci-fi novel, and this is so timely. Sci-fi is a window into reality, if done right, and the future is now, my friends.

Sign Up for Saturday Morning Serial

If you want to read the story before you can buy the book, subscribe to the Saturday Morning Serial. One chapter, every Saturday, just for you. A thank you for supporting TCIP.

Biology with a Prompt

One of the defining features of this decade will be the fusion of model-based cognition with biological experimentation. CRISPR-GPT is one of the first real systems to operate at that intersection—not just as a tool, but as a collaborator.

And that changes the texture of science itself.

In this new world, experiments begin not with a lab notebook sketch or a whispered question to a postdoc, but with a prompt. “I want to see what happens if I knock out this gene.” “Can we test this in organoids instead?” “What if we activate this immune marker and observe resistance profiles?”

The prompt becomes the proposal. The model becomes the method. And the researcher becomes both conductor and critic in a symphony of automated agents, human judgment, and living systems.

We are not just building better tools.

We are building a new language for discovery—one where biology speaks through code, and code speaks back with insight.

And at the frontier of that dialogue, humanity remains the experiment.

Cheers,

-Titus