Ron Alfa

1/ Spatial transcriptomics is among the richest view of human biology that we have: 18,963 genes mapped at subcellular resolution. It's also almost never collected outside of research settings. So we trained a foundation model to generate it from a clinical H&E image alone. Meet TARIO-2. 🧵 noetik.blog/p/tario-2-a-wh…

@kylekuzma AI not just impacting timelines, but improving probability of success by predicting response in oncology.

Three things I’m underwriting hardest in 2026: -AI applied to industrial throughput. -autonomous defense + space robotics -clinical-stage biotech where AI compresses trial timelines. Everything else feels crowded

Published a piece outlining our approaches at OpenADMET where we are trying to build the CASP for small molecule ADMET prediction w/ @fraser_lab @BioSteve @lnhandly @jchodera @Mark_Murcko @wpwalters nature.com/articles/s4146…

It’s easy to forget just how quickly the size of the pie (AI Market) is expanding in nearly every category

This would be a long post but a couple thoughts: -- Optimizing for valid training data: in vivo human tissue or in vivo perturbation only, no cell culture. -- Paired multimodal for every sample + the right modalities for FMs to learn biology and connect to clinic -- Designed for FM training from first principles with control for biases (batch effects, artifacts, etc.) -- We control full stack, so we can learn and course correct (e.g. build custom data to solve issues) -- The goal is to achieve accurate human simulation from world models, optimized in post-training with perturbation -- Scale of data along the right dimensions to achieve scaling laws

@Ronalfa @NOETIK_ai I know some of it, but would you be willing to share the core principles of what data you're making? Direct human obviously. Interventions because causal info. Why cancer? What other decisions are you making that you think will matter long-term?

Biased, but we’re betting heavily on our own human data @NOETIK_ai.

Welcome to the age of Engineered Biology. The 21st century will be defined by our ability to engineer biology at a cellular level, and ultimately systems level (e.g., the immune system). All of it enabled by AI helping to deconvolve ~4 billion years of cumulative evolutionary complexity. We live in amazing times.

A startup idea that only works if there are already a significant number of people using it is not a valid startup idea. There has to be some subset of users who need what you're making so desperately that they'll use it even if no one else is.

We are now at a point where LLM-content is so pervasive that either they achieve true intelligence, or human intelligence reverts to the mean of training data in areas like science. We are already on our back foot. We dump everything into an LLM and uncritically take responses at face value. Science is getting more and more sloppy. AI is used to both read and write content. Science is becoming AI generated, and humans are losing the ability to distinguish what is real.

@patrickc Wonderful project, excited to follow updates.

Tyler and I just published a list of the recipients of the New Aesthetics grants: newaesthetics.art/grants. Thank you very much to all who applied. There were far more applications than we expected. We funded 28 grantees and are excited to see what they create. My reflections on the whole thing: • Though there are clearly selection dynamics afoot, figuring out some route beyond the current aesthetic moment seems to be of wider interest in the art community than I would have guessed. Many applicants described their dissatisfaction with the status quo, some in strong terms. We had to close applications after a few weeks because there were so many. • It's too early to call it, but it seems that both beauty as an unapologetic goal (contra much that is in modernist and contemporary approaches), and ways to channel pre-modern styles into something new for the present era, are of growing interest. • The awards made me reflect on the perhaps obvious issue of how hard it must be for an artist to persistently do something new: schools, galleries, buyers, etc., all have structurally embedded preferences as well. These individual awards made me wonder what form supporting new clusters could take. • Architecture seems to me like the discipline most ripe for new ideas. One correspondent observed: "American architects are somewhat constrained by the association with the academy, in addition to the well known regulation issues. There is a tendency to overthink things so that the designs are formally interesting to someone deep in the conversation, but lacking poetry and magic. There are more firms in Europe, South America and beyond that “just do things” (especially in places where it is easier to build)." This was evident in the submissions. • AI seems to be making people rethink things in a quite fundamental way, just as urbanization/industrialization/popularization of photography did at the end of the 19th century. For some that will mean interesting new forms of AI-augmented art, but the effects of the rethinking will likely be wider. • Arts funding is clearly as precarious and scarce as ever. That's unfortunate, but it probably also means that individual actors can have meaningful impact, and I encourage others to get involved if interested. • There's a lot to know that is not written down, and I'm very grateful to those who have helped and advised me along the way.

Folks on here are missing the point. Everyone can be engineers or researchers or mots, the point is to make titles more uniform and flatten org chart.

if you’re not trying to solve seemingly impossible problems, what’s the point anyway?

@MartinBJensen 100%. No plates here!

I don't understand what @Ronalfa is talking about, what is a plate? Your controls should be in an animal 😉

A leading indicator that we can win at other diseases too.

@RecursionChris Easy VC diligence: just ask for the plate map!

Control strategy for large scale bio datasets serving as substrate for AI is its own science (and one I’d suggest we played a very meaningful role in pioneering). Can’t agree more with Ron. If you see controls in a row, column or on the edge of your plate, rather than randomized across a plate, you know your dataset is not built with ML/AI in mind.

@fulop_dan @dargason @anshulkundaje @gama_search 💯

@dargason @anshulkundaje @Ronalfa @gama_search Best to ensure good data upfront

People are still generating “ML datasets” with all kinds of confounds. If the controls are all next to each other on the edge of the plate, no randomization, ngmi.

@anshulkundaje @gama_search Seems like there’s no real evaluation of all this stuff so basically it comes down to is the work rigorous or not. I don’t even think the journals can review. But at the same time, all the serious technical people I speak with do absolutely know what’s real and not.

@Ronalfa @gama_search A famous PI building scFMs once told me very confidently that their models can automatically correct technical artifacts without even including any covariates in the model & with no constraints on the expt design 😂😂😂

Getting cliche to say now but the AI plans to “cure all human disease” fixate on the CS/tech legible parts of drug discovery which aren’t the overall bottleneck Your awesome lead candidates will still take a decade to get approved Generative personalized medicine is the way

fuxxcancermaxxing

@jackdent @SabrinaHalper 🔥

Fun to sit down with my friend @SabrinaHalper and talk about Chai's journey. Thanks for having me on!