“AI for Science” in biology is becoming “Gimme mo datasets and compute.” That’s the wrong move! Across industry and academia, we’re funding pipelines that build bigger atlases, train bigger models, and then declare progress because held-out metrics improve. But biology is not a

“AI for Science” in biology is becoming “Gimme mo datasets and compute.” That’s the wrong move! Across industry and academia, we’re funding pipelines that build bigger atlases, train bigger models, and then declare progress because held-out metrics improve. But biology is not a

There's never been a better time to come to Canada for a PhD, postdoc, or faculty position. The first part of the $1 billion+ investments is already showing up as additional funding for PhD/postdocs: https://t.co/keyFUlpuvr

@GoogleDeepMind has led the way in applying AI to science. Great to see the White House’s new Genesis Mission & excited to be collaborating with @ENERGY by giving National Lab scientists accelerated access to our frontier AI models & agentic tools, starting with AI co-scientist.

Ilya Nemenman and a team of scientists from @EmoryPhysics have developed a 'periodic table of AI methods' - a unified view of the algorithmic methods that may allow design of "new AI methods that are more accurate, efficient and trustworthy.” Full story ➡️ https://t.co/05kZZvVgU9

IT SUCKS to see all the money going into LLMs in the name of “AI for Science,” especially in biology. Like sheep. What an effing waste of a decade. LLMs will have some blip in predictions and we shall celebrate but it shall be a "victorian surgery". LLM-based generative biology

Single brain cell looking for a connection

Two cells feeling each other. This is the kind of dynamic, energetic process that goes on inside our bodies every moment. Once we realize our energetic, communicative nature, our dependence on social connections with other beings comes as no surprise. https://t.co/dzlM4F5UTK

Oh look, ma. Apparently LLMs, as they stand today, are not great for scientific discovery! Who would have known. And who could possibly have been saying this all along for last 3 years! 😇 Receipts attached: Hey VCs, Your Outdated AI Investment Strategy Will Cost You and the

We are NOT building a “virtual cell” at CELL. No one knows what a virtual cell is. CELL, in our case, stands for Consortium for the Equations of Life and Living-Systems. A mouthful and ambitious, but the purpose and direction are clear. For every generation, there is a

Stop Calling It Thinking! We have reached the end of innocence about what we casually call as “thinking” in AI models. In this article, we argue that what people call “reasoning” is a budgeted runtime algorithm, candidate generation plus selection plus verification plus tool

Yup! Just do open science. There are several avenues for publishing today. Stop being idiots. I would strongly encourage pre-prints to start with. Arxiv, BioArxiv… Also @PracheeAC promotes this under Astera, heavily. Profs like @anshulkundaje are pushing for this strongly in

We are NOT building a “virtual cell” at CELL. No one knows what a virtual cell is. CELL, in our case, stands for Consortium for the Equations of Life and Living-Systems. A mouthful and ambitious, but the purpose and direction are clear. For every generation, there is a

Cancer vaccines are tough! We’ve got a murine model with MS validated mutant pMHCs that are also spontaneously immunogenic Tc epitopes in response to implanted tumors. Good antigens! Optimized the vaccine formulation for two years. Treat +/- aPD1; growth curves looking like PBS.

CELL is a serious invitation to mathematicians who are tired of watching biology be reduced to curve fits with no invariants and no right to generalize. We are building a universal configuration space for living systems, a shared state space with explicit operators, boundary

Evolutionary dynamics study how populations of agents, strategies, or traits change over time under selection, mutation, and interaction. They provide mathematical rules—such as replicator dynamics and stochastic evolutionary processes—that link individual success to

Thrilled to launch RamaX: our platform for ultra-fast, accurate, and sensitive screening of massive binder libraries (from 1K to 1B minibinders, VHHs, or scFvs) in just 1-2 weeks. This method has been an absolute game-changer for us at @diffuse_bio. Internally, we’ve completely

True. But nobody can define what a 'virtual cell' precisely means. It's everything and it's nothing at the same time.

So many interesting protein designs are now on the way to the lab for experimental validation, thanks for all the submissions!! Check out all proteins here: https://t.co/wP2xCzi29r

We’re releasing a new eval to measure expert-level scientific reasoning: FrontierScience. This benchmark measures PhD-level scientific reasoning across physics, chemistry, and biology. It contains hard, expert-written questions (both olympiad-style problems and longer

Stop Calling It Thinking! We have reached the end of innocence about what we casually call as “thinking” in AI models. In this article, we argue that what people call “reasoning” is a budgeted runtime algorithm, candidate generation plus selection plus verification plus tool