Building @cultivariumFRX. I like bio+hardware+software. Views can be yours too.
Boston, MA
Joined October 2007
- Tweets4.2K
- Following848
- Followers1.9K
- Likes17.3K
Pinned Tweet
Still Day 1 for our understanding of 4 billion years of evolution.
If you believe the brightest should run toward the darkest places, join us.
Henry Lee retweeted
Largely agree with these thoughts, and have some perhaps discordant points to add.
- Solving biological diseases in the next few years will benefit mostly from the usefulness of the language models, substantially more than molecular ML.
- Proponents of the bitter lesson for biological models (not LLM) aren't wrong, just (possibly pathologically) too early. We don't have enough data, and the data we can collect quickly is probably not the right kind (we may get lucky with biology being easy, but it's usually not).
- I think just scaling models and public (passively accumulating) data is a cost ineffective direction (it's like scaling dense NNs on the internet). Today's model outputs are as good as GPT2. A single bioinformatician with linear regression, sometimes even without any additional data, can beat these models in many applications (people rarely benchmark AI advances against a simple Potts model). The most promising approaches for molecular models are innovating heavily on architecture (e.g. Baker Lab, Alphafold, etc). This may automate/speed up with LLMs.
- Few companies (hopefully Dyno included) are betting that we've triangulated enough pieces from multiple experimental channels to predict (with scale) how things translate in humans. Having worked on translating in vitro things to primates, I'm very cautious with the idea that current models, trained on current public data, can do anything more than manipulating already working solutions, either explicitly or through implicit remixing of what has worked. It may still work, but the underlying bets offer no specific reason why there is any epistemic risk advantage other than being "made with AI".
- Short term, there is more promise in combining molecular models with LLMs rather than scaling them. The two sources of lift in AI discovery elsewhere has been (1) Breaking Silos: AI synthesizes a bunch of knowledge/expertise that is rarely present in the same room (2) Data mining : AI pushes through volumes of data, with an intelligent eye, that humans or simple algorithms simply can't scale or adapt to. I'm curious what will come out of multimodal training of LLMs and biological data, which seems to be something the big players are pursuing. I think latent reasoning in the language (or better yet, + code) domain will reveal a lot more biology than scaling on nucleotides or proteins alone.
- My view is that the next opportunity of combining lab and AI will be massive-scale hypothesis testing with LLMs. Currently a multiplexed assay generally tries to test a handful of ideas because the humans are bandwidth limited in designing properly powered experiments. LLM's can scale this very well.
I have no doubt lots of good news will come on the impact of AI in biology in the next five years. I can't tell what paradigm will win, but my prediction is that it will not be the same paradigm imported from the LLM land (which will get a ton of funding anyways, because it's easy to pattern match for tech VCs).
There’s a really pronounced divide between people who are on the AI side of things and people who are on the biology side of things. Even among people working in AI, the divide persists based on where they’re focused.
The AI side, adherents of scaling laws and the bitter lesson, believe that AI getting better means AI being more capable of novel discoveries in medicine (that go beyond discovering new drugs in well understood targets or speeding up dry lab work etc).
The biology people think the idea of a bitter lesson in biology is kind of ridiculous and that regardless of AI’s capabilities it’s going to be at least 5-10 years until you could hope to reliably simulate human toxicology in silico (right now AI/ML can’t even reliably simulate in vivo studies like binding affinities or functional assays).
Even the most AI pilled of them, who will list all the areas that AI can reduce the time spent on various areas, don’t think the concept of AI discovering novel targets or generating new understanding of causal biology is realistic in the near term. Most, in fact, will readily characterize the assumption that even an “ASI” level model could find cures hidden in our current understanding (without expanding that understanding with significantly more biological data) as patently false.
This was a big part of why we wanted to experience it ourselves by using AI for a (very) small portion of what it can do in biology. It was both mind blowing and revealed some poignant insights into how different the problem of making AI better at something like coding is to the problems of biology.
Even if it unlocked the ability to generate new molecules at a stunning pace, its simulation (in silico) of binding only had 49% accuracy. And if we wanted to improve that, we’d need to feed it back the data from the lab - data that took about 30x as long to get as it took to generate the molecules. And these were tests in a test tube (turn of phrase), imagine how long it would take if we wanted to improve the model with data that required recruiting humans and monitoring their reactions.
Human biology is very complex, and adding a drug - especially a novel drug and doubly so for a novel drug acting on a novel mechanism - means mapping interactions that have nth-order-effects that we don’t have the data to predict. Virtual cells have made advancements enabled by AI, but the perturbations equate to a small fraction of the possible.
This is before you even get into curing disease, especially uniquely human diseases. Brain function involves cellular identities that are unique to humans, meaning animal studies are of little use.
How long does it take to understand what the side effects of chronic administration of a drug in humans are? AI can speed up analyzing the data, or designing the trial or even getting it approved…but it can’t make a 6 month trial complete in 3 months.
I think its important to point out that significant advancements can be made in AI that contribute to biology, while also being realistic about how much it speeds up something as complex as curing a specific cancer or diseases of the brain.
Henry Lee retweeted
Announcing Ikigai Grants:
Microgrants give early-stage, bootstrapped talent capital. But that’s not enough.
I’ve watched too many grantees become perpetual microgrant chasers, wasting time applying to one $5k grant after another. Worse, most programs encourage this by funding the blanket “smart people doing cool things,” with no real conviction behind it.
I wouldn’t be here without patronage from @tylercowen & @jposhaughnessy. But looking back, the mentors & community were worth much more than the money.
So I built the program I always wished existed:
- $50k total, avg. $5k check
- Matched with mentors & community across @RenPhilanthropy , Fulcrum Science, and my personal network of founders, fund leads, investors & scientists.
Cohort 1 backs emerging talent in areas I believe are deeply underfunded and undervalued. Every application is reviewed by me.
Ikigai Grants is a program of @RenPhilanthropy, with support from Fulcrum Science. Advised by many wonderful friends & colleagues including: @AshleyNAbel @RaianyRomanni @NeuroStats Bonnie EJ Maven @idatin @ashleyz413 @JackScannell13
Apply by Nov 9th, 9a PST:
alexaraki.xyz/ikigai
Henry Lee retweeted
Boston we can’t afford further aura loss
Boston is so uncool that people are mad as hell that a bunch of Starbucks locations closed downtown and now they're forced to go to locally owned coffee shops bostonglobe.com/2026/10/19/b…
Henry Lee retweeted
Of course the flying Tesla Roadster is a gimmick but that sort of misses the point, the tech flex behind it isn't.
Tesla's bulk business is about advanced technology from EV to self drive, but decided to create the most ridiculous manual drive supercar. Supercars are a gimmick,almost by definition, based on how wild can you make them.
The gimmick that can make the car literally fly, briefly, can make it corner like no other supercar and accelerate, literally like a dragster on a test track. It's got nothing to do with flying continuously, but having a car that will outperform every Lambo, Ferrari or McLaren with technology that they cannot replicate because it requires experience from space launches and computation that they don't have now nor will in future.
So Porsche, Bugatti, Aston Martin just became legacy prestige brands, like Rolex, Richard Mille or Patek Philippe. These cars celebrate the bygone age and the baroque, physical complexity of the internal combustion engine just as mechanical watches do for non digital, complex time-keeping mechanisms. But hey won't be faster anymore than a mechanical watch keeps better time than a Casio. They are now, de facto, mobile jewelry and no longer supercars with superlative performance. And this affects their ability to market the bulk business of mainstream European internal combustion engine cars. It's the end of an era of industrial era manufacturing.
I hate to be the one to burst the Tesla bubble today, but the spectacular “flight” we witnessed at the Roadster reveal needs a little context. The car is carrying around 250kg of compressed air, plus the weight of the tanks, valves and thrusters needed to contain and release it.
Based on the published figures, that gives it roughly 1.5–2 seconds of maximum thrust before the system needs to be refilled.
In short, it's a gimmick.
Look at all this technical debt that software engineers left around
Every week someone tells me they vibe-coded their app over the weekend and don't need developers anymore.
Before it touches a paying customer, make sure it includes:
> a schema someone designed, instead of one the AI improvised table by table
> login that still works when someone puts a plus sign in their email
> password reset links that expire and work once
> MFA, plus a recovery path for the user who lost their phone
> sessions that time out, and a way to kill every session for one user
> SSO for the enterprise customer who won't sign without it
> SAML, because that customer's IT team doesn't do magic links
> SCIM, so a fired employee loses access the moment IT offboards them
> roles and permissions deeper than "admin" and "everyone else"
> checks that stop user A from reading user B's data by changing one number in the URL
> tenant isolation, so one customer's bug never shows another customer's records
> an admin "log in as this user" button for support, with an audit trail on every use
> rate limits on login, so nobody can guess passwords at scale
> lockouts that stop attackers without locking out your CFO the night before the board meeting
> database migrations that run forward and roll back
> migrations that don't lock the main table at 2pm on a Tuesday
> indexes, because the query that was instant on 50 test rows times out on 5 million real ones
> connection pooling, so a traffic spike doesn't exhaust the database
> backups every night
> a restore from those backups, tested, because a backup you've never restored is a hope
> point-in-time recovery for the day someone runs a delete without a where clause
> soft deletes, so "I deleted it by accident" is a support ticket and not an incident
> validation on the server, not just in the form
> timezones stored right, so the 9am reminder doesn't fire at 4am
> the customer with an apostrophe in their last name
> emoji, accents and right-to-left text in every field
> money stored as integers, because 0.1 + 0.2 doesn't equal 0.3 in floating point
> currency, rounding and tax rules for every country you sell into
> a data model that survives your second customer asking for "one small change"
> idempotency keys, so a retried payment doesn't charge the card twice
> deduplication for the user who clicks submit four times
> pagination that still works on page 400
> exports, because sooner or later every customer wants their data in a spreadsheet
> imports that survive the CSV someone saved in 2009
> retention rules, and code that deletes what your contract says you'll delete
> secrets out of the code, out of the repo and out of the AI chat history
> rotating every key the AI pasted into a config file last month
> confirming every package the AI installed exists and is the one you meant
> dependency scanning, because the AI pulled in libraries nobody reviewed
> input sanitization everywhere a user can type
> protection against SQL injection, XSS, CSRF and the rest of the OWASP list
> file uploads that get scanned, size-limited and stored where they can't execute
> CORS rules that aren't "allow everything"
> security headers on every response
> encryption at rest and in transit
> least-privilege access for everyone, including the AI agent holding production credentials
> logs that record who did what, without recording passwords, tokens or card numbers
> a penetration test before your enterprise customer's security team runs their own
> answers to the security questionnaire every enterprise buyer sends, with evidence behind each one
> a disclosure process for the day a stranger emails you a vulnerability
> a plan for the day a laptop with production access gets stolen
> SOC 2, which means policies, evidence and an auditor, not a checkbox
> HIPAA, the moment a single patient record touches your system
> a BAA with every vendor that touches that data, including your AI model provider
> audit trails that show who viewed a record, not just who changed it
> PCI scope, and card numbers kept out of your database
> GDPR deletion requests that reach every table, backup and log
> data residency for the customer whose contract says EU only
> quarterly access reviews someone signs off on
> dev, staging and production environments that match each other
> infrastructure written as code, so the setup lives somewhere other than one person's memory
> a deploy pipeline that runs the tests before anything reaches customers
> one-click rollbacks, because Friday deploys happen
> zero-downtime deploys, because your customers don't have maintenance windows
> feature flags, so you can ship the code without shipping the feature
> autoscaling for the Monday morning spike
> a CDN, so the app loads fast in Singapore too
> DNS, SSL certificates, and renewals that don't expire on a Sunday
> background job queues, so slow work doesn't freeze the page
> retries with backoff, so one flaky API call doesn't become a lost order
> a dead-letter queue for the jobs that fail every retry
> cron jobs that don't run twice the day you add a second server
> caching, and clearing that cache when the data changes
> SPF, DKIM and DMARC, so your password resets don't land in spam
> bounce and complaint handling on every transactional email
> storage rules, so you stop paying to keep 2021's temp files
> cost alerts, so a runaway agent loop shows up the same day and not on the invoice (we learned this one for $1,382.59)
> a disaster recovery plan, written down and tested once a year
> a second region, if your contract promises uptime one region can't deliver
> structured logs you can search at 3am
> error tracking that groups the same bug instead of sending you 4,000 emails
> uptime checks from outside your own network
> alerts that page a human, tuned until they stop paging for nothing
> dashboards for latency, error rate and throughput
> tracing for one slow request across every service it touches
> an on-call rotation, and a person who picks up
> runbooks, so the engineer on call knows what "queue depth critical" means
> postmortems that change something
> a status page customers check before they email you
> unit tests written by someone who knows what the logic should do
> integration tests for every third-party API you depend on
> end-to-end tests for signup, checkout and the other flows that make you money
> load testing before the launch email goes out
> test data that looks like production without being production data
> regression tests, so fixing bug 41 doesn't bring back bug 12
> testing on the five-year-old laptop your biggest customer's ops team still uses
> accessibility compliance, because the enterprise contract requires it
> subscriptions, upgrades, downgrades and proration
> failed-payment retries and dunning emails
> invoices with the customer's PO number on them, because procurement won't pay without it
> sales tax and VAT
> refunds, credits and partial refunds
> usage metering that matches what you bill, to the cent
> annual contracts with custom terms your billing logic never planned for
> revenue reports your finance team can reconcile
> webhooks that verify signatures
> webhooks that retry, with a log of every failed delivery
> a versioned API, so you can change it without breaking your customers' integrations
> API keys customers can create and revoke
> API rate limits that protect you without blocking real integrations
> a plan for the day a third-party API changes its response format without telling you
> OAuth token refresh, because tokens expire at the worst possible moment
> a sandbox where customers can test their integration
> protection against prompt injection in every field that reaches the model
> customer and patient data kept out of prompts that leave your network
> a model spend cap per customer, so one heavy user can't run up your bill
> evals, so you find out when a model update makes your answers worse
> a fallback for the afternoon your model provider goes down
> a migration plan for the day your model gets deprecated
> prompt and output logs for debugging, minus whatever you're not allowed to keep
> a human approval step for anything the AI does that can't be undone
> pages that load on hotel wifi
> empty states, loading states and error states on every screen
> error messages a customer can act on, instead of "something went wrong"
> forms that keep what people typed when one field fails validation
> emails that render in every mail client your customers use
> notifications people can turn off
> undo
> keyboard navigation and screen reader support
> translations, with dates and numbers formatted for each locale
> a back button that does what people expect
> code a human can read
> code review on every change
> a branching strategy, so two people can work without overwriting each other
> monthly dependency upgrades, so you're never two major versions behind
> framework upgrades before the old version stops getting security patches
> deleting the dead code the AI left behind from its first three attempts
> splitting up the 2,000-line file the AI kept appending to
> architecture decisions written down, with the reasons
> a README that gets a new engineer running locally in under an hour
> documentation for whoever inherits this
> an admin panel for your ops team
> support tooling, so customer success can answer questions without pinging engineering
> a changelog customers can read
> terms and a privacy policy that match what the code does
> every account, domain and repo in the company's name, not the name of whoever built the prototype
> a bus factor higher than one
Oh, and besides keeping all of this working every day, forever… your vibe-coded app will still need people who know what they're doing:
> reading every line the AI wrote before it ships (a senior engineer)
> designing the data model before your second customer breaks it (a backend engineer)
> explaining why the AI's fix for one bug created two new ones (a senior engineer, again)
> deploys, rollbacks, servers and the cloud bill (a DevOps engineer)
> answering the page at 3am when payments stop (whoever's on call)
> patching the framework the week a vulnerability goes public (security)
> walking an auditor through how patient data moves (security and compliance)
> testing what the AI says it tested (QA)
> running a migration on a live database without losing a row (a backend engineer)
> saying no to the feature that breaks three others (a tech lead)
> turning "works on my laptop" into "works for every customer at once" (an architect)
> debugging the code when the AI that wrote it can't (a senior engineer)
> remembering why it's built this way, a year from now (someone who was there)
…wishing them the best 🫡
Yet all the dialogue and funding is about mechanizing more of what we’ve done, or railing against it to preserve the artisanal.
“Second, we realized that one cannot predict what future laboratories will look like simply by choosing the technologies that seem optimal today.”
Big news! Our new book, "Making the Modern Laboratory," is now available for pre-order. It is a stunning, 504-page hardcover book that explains how the research laboratory, and everything in it, came to be.
Did you know that the modern micropipette was invented by a German postdoc so frustrated with mouth pipetting that he ditched his lab for two days until could return with a prototype?
Or that an 18th-century scientist sewed tiny taffeta pants for male frogs to determine whether or not semen was essential for fertilization?
These stories and hundreds more are told in the book, which has a hard cloth cover and is printed in full color on Arena Natural Rough, a textured, uncoated paper made by Fedrigoni. The book is also finished with colored endpapers, thread-sewn binding, and a squared spine. It is priced at $49 and worth every penny.
Learn more and order: press.asimov.com/articles/mo…
Henry Lee retweeted
If you care about biotech and drug development watch this interview with Matt Gline, the CEO of Roivant, a biotech firm that is now worth $24 billion.
He talks about why he thinks predictions about AI from people like Anthropic’s Dario Amodei are overly optimistic.
“If Dario is right and I'm wrong, and he cures every disease in five years, man, are we all the winners from that outcome,” Gline said.
“I think if you work in this industry, it is very hard not to be skeptical of any of those claims. Because, and I've said this in other settings recently too, biotech is a wood chipper that like is built to like ingest souls and shred them. It's just like such a hard business that like someone comes and they're like, oh, but I've built a computer and I'm going to push a button and all of that goes away and it just works. Like that is probably not the way that's going to go.”
We are so back synbio4.0?
I'm excited to share something I've been working on...
NeoGenesis Capital is a new fund focused on the intersection of superintelligence and synthetic biology.
I believe some of the most extraordinary breakthroughs of the coming decade will happen at this intersection.
I'll be sharing more on @moonshots_pod soon, including why I'm so excited about what's ahead!
Incredible read, beautiful story.
“They do not record that because of his mind, a premature infant trembling in a NICU ward can be given the exact measure of a medicine that was never tested on such fragile flesh.”
Earlier this year, I came across a peculiar email written by the late Lewis Sheiner, a UCSF physician who built the mathematical methods for safely testing drug doses in people. This field is called "pharmacometrics", and it remains a cornerstone of modern drug development.
In this 20-year-old email, he emphatically assigns credit for the foundation of this work to an *economist* and *quantitative trader* named Barr Rosenberg -- who, driven by a personal tragedy involving a mis-dosed medication, had reached out to him. How did an economist get entangled in the math of drug dosing?!
This mystery seemed too fascinating to ignore. So I spent this summer tracing the story backward, speaking with Barr's former colleagues, clinical pharmacologists, programmers, and even the first director of the FDA's Center for Drug Evaluation and Research, Carl Peck(!) to learn about Barr's unlikely contributions to this important field.
This is the story of Barr Rosenberg, the pioneering pharmacometrics software -- NONMEM -- that his math inspired, and the curious inner workings of how a ragtag team of clinicians, statisticians, and programmers shaped the way we test drugs today: "Borrowing Strength".
Henry Lee retweeted
I'm looking for technical scouts to find the world's best researchers whose dreams don't fit into academic lab- or venture capital-shaped boxes
It is, first and foremost, an ideas job. Flexible on work arrangement. Technical vision must be first-rate.
freaktakes.com/p/help-wanted…
Henry Lee retweeted
The main limitation in AIxBio may be our ability to observe biology, not intelligence.
Increasingly excited about new technologies that let us see and understand biology at much higher resolution.
So let’s do it.
The entire startup landscape has changed postAI. Building products and acquiring distribution have become mostly commoditized, pushing founders to the fringes of deep tech, physical AI, and things AI can't automate.
At the same time, distribution is shifting from paid ads and growth hacks back to hardcore, face to face sales.
I believe the next wave of great founders will be founders who can sell door to door, creatively acquire their first 100 customers, and build real, hands on tech that is difficult to replicate.
We're going back to the early days of computing in the 80s and 90s, when founders had to get their hands on actual silicon, build hardware, and sell through door to door pitches, catalogs, and personal relationships and networks