Thematic, Cross-Asset Investment Research
New York, NY
Joined February 2021
- Tweets44.1K
- Following2.4K
- Followers299K
- Likes42.2K
Our Agentic Utilities basket has outperformed SOXX by about 38% over the last six months. Its drawdown from YTD highs is -9% versus SOXX’s -27%.
Considering the overwhelming pressure at the time to focus solely on AI infrastructure which went up every single day…pretty good.
Some really strong results from a pair of our “Agentic Utility” names today ($BAND & $TWLO) - playing the infrastructure angle on AI has been great, but it’s pretty nice to see some mentions of AI driving earnings beats outside of the semis complex.
citriniresearch.com/p/agenti…
Citrini 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.
Citrini retweeted
The more I work with powerful AI systems the more optimistic I become about the future of biology. As described in this great post from @citrini it will take time to see the impact of AI across the full complexity of human biology. My view is that drug design is where we are seeing and, in the near term, will continue to see the most visible impact. Drug design is to drug development what coding is to software development. It is the most verifiable step with only weeks or months needed to get results. Solving drug design is necessary but not sufficient and at minimum progress here will reduce the time it takes to bring drugs to patients. My more optimistic vision is that AI will enable us to create drugs beyond our current imagination.
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.
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.
What I’m confident about is that there will be a shocking (to some) degree of acceleration in advancing new measurements, automation and any other kind of technology that makes gathering that useful data easier. I have no doubts that AI will reduce the amount of time it takes medicine to cure diseases.
I just think that if you think that reduction will happen to aspects of the process that are irreducible, and overweight the impact of those new technologies, you’re going to end up disappointed. And expecting a cure for Alzheimer’s in 2 years is doing that.
Citrini retweeted
What is AI really capable of in bio?
To separate hype from reality, we spent 3 months working firsthand with @AnthropicAI's internal research model to see if it could create an experimental scorpion antivenom.
Introducing Citrini Therapeutics.
citriniresearch.com/p/citrin…
Citrini retweeted
Today, we’re announcing an open call for guest investigators in life sciences. Over 8–12 weeks, guest investigators will work directly with my team and other Anthropic researchers to produce analyses and tools for their work. We’ll run our models and harnesses, at no cost to them. We’ll work together to open-source the resulting data, artifacts, and code, and write preprints about what we find.
We want humans to remain at the center of scientific discovery, deciding what questions to pursue, and how to pursue them. Those decisions are not ours to make alone. This open call is part of our ongoing efforts at Anthropic to support researchers and accelerate scientific discovery.
I’m looking for principal investigators and their research groups at academic institutions, nonprofits, or government labs, working in causal human biology, structural and computational molecular biology, medicinal chemistry, drug discovery, and clinical science.
Applications are open for about 14 days: anthropic.com/news/open-call…
Avoid Okta catching a stray in a convo about cybersecurity Challenge (Difficulty Level: Impossible)
Matthew Prince on PHP2Shell, Duo on Arrival, 100 Really Difficult Phone Calls, Norwegian Gold Scheme xitter.cf/i/broadcasts/1NGarogqv…
While eighty thousand new experts in the AI trade were minted in the past 8 months, all extremely confident in the names everyone else was already in, there have been accounts from the old days of fintwit generating alpha in various corners of the mkt.
This quoted post is unavailable.
I’m kind of worried that we are gonna have massive exploits like every other week
Some stats on #php2shell. In the 10 hours before it was publicly announced, across the @Cloudflare network we saw thousands of attempts to exploit it. That jumped to tens of thousands in the 45 minutes post announcement, hundreds of thousands within 2 hours, and millions now.
The rate at which the economics of AI molecule design are getting better is insane. Probably a cumulative 50x reduction in token use from idea to ready for synthesis & experiments over the past 6 months.
Mr. Ricks is a legend.
Well said Vas. AI will help us conquer disease, but this will take time, effort and lots and lots of data that doesn’t exist, yet. I would add that robotic labs will begin to make scaling data more tractable. lnkd.in/p/g8qvP5qX
Citrini retweeted
New on our Science Blog: designing protein binders against scorpion neurotoxins.
Scorpions sting about 1.2m people every year, and antivenoms are often costly or in short supply. Claude has made an early step toward finding new treatments.
anthropic.com/research/claud…
