Incredibly, incredibly excited to introduce SpliceAI2, the next generation of splicing prediction! 🧬🚀🧵

Saturation mutagenesis of long-context seq2fun models requires millions of evaluations per gene. We introduce Multi-ISM to make this mapping routine: biorxiv.org/content/10.64898…

So excited for today! 🎉 We open the first New England Computational Biology Symposium at Microsoft Research New England. By chance, it's also 10 years to the day since I started my lab. Perfect way to celebrate this milestone! newenglandcompbio.org #NECB2026 @NewEngCompBio

AlphaGenome (AG) & the AVI score is NOT "AI slop taking over genomics". I will give some concrete reasons why rather than evidence-free ranting (which I also like doing occasionally 😜). 1/

Maybe it’s time to say goodbye to UMAP and other low-dimensional embeddings (at least in many single-cell analyses). We developed MILK to encode single-cell populations into tree representations at unprecedented scales. biorxiv.org/content/10.64898… Here’s what MILK is: 🧵

A free repository rich w/AI engineering resources: bit.ly/4wxecUo v/Chip Huyen

Vercel 开了一门完整的课程：Build Your Own AI Coding Agent Harness 教你从零手写一个完整、能真正干活的 AI 编码 Agent 框架

Claude Science is an amazing tool. But, as @pdhsu has pointed out, these binders weren’t designed by Claude, but rather by domain-specific models, most of which are publicly available, like RFdiffusion. Maybe this is a good time to explain the whole science behind this a bit, as well as the history of these models. A “binder” in this case is simply a protein or a fragment of a protein engineered to stick tightly and selectively to another protein. One example is a molecule that can latch onto a receptor and prevent it from interacting with its normal partner, thus blocking its function. Optimising "binders" is just one small part of the drug discovery pipeline. We already have AI models that are very good at the task of binder design. These models have been extensively trained on relevant domain specific data: int his case, protein structure data. One example is RFdiffusion, a de novo protein design model which was published in 2023. Unlike AlphaFold, which predicts existing protein structure from sequence, RFDiffusion can generate new candidate protein structures, including structures shaped to bind a specified patch on another protein. RFDiffusion was built by taking RoseTTAFold, a protein-structure-prediction model, (so similar in nature to AlphaFold, which was released in 2021, and retraining it as a diffusion model. What researchers did in practice was take experimentally determined protein structures from the Protein Data Bank, added noise to their coordinates, and taught the model to reconstruct them. The original RFdiffusion paper was already demonstrating experimentally validated binders with nanomolar affinities and a 19% hit rate on several test targets. So the core of this technology was not created through Claude. The key models for protein design are now three to five years old. What Claude adds is still amazing useful and labour saving (I've written about how I'm excited about it before), but imo it acts more like a very useful interface and orchestrator.

Yesterday at @ISCB_RegSys I introduced Cherimoya: a lightweight sequence-to-function (S2F) model that achieves state-of-the-art performance. We've been using it internally at the Programmable Genomics Laboratory for several months and are excited to share it with y'all now!

GenPerturb: sequence-grounded interpretation of perturbation transcriptomes using pretrained genomic models biorxiv.org/content/10.64898…

Just want to give a shout-out to David Kelley @drklly who I think often does not get the credit he deserves (outside our core community). I want to highlight why I think he is such a fantastic scientist and leader in regulatory genomics. 1/

PoolParty: streamlined design of DNA sequence libraries in Python biorxiv.org/content/10.64898…

Kudos, thnks for standing up. no need for messy biological insight when parameter count does all the talking. Why bother with interpretation ven u can just scale? once u pass that sacred billion parameter mark, the manuscript basically writes its own acceptance letter to Nature.

New art project. Train and inference GPT in 243 lines of pure, dependency-free Python. This is the *full* algorithmic content of what is needed. Everything else is just for efficiency. I cannot simplify this any further. gist.github.com/karpathy/862…

Explaining linear regression

Okay so, we just found that over 50 papers published at @Neurips 2025 have AI hallucinations I don't think people realize how bad the slop is right now It's not just that researchers from @GoogleDeepMind, @Meta, @MIT, @Cambridge_Uni are using AI - they allowed LLMs to generate hallucinations in their papers and didn't notice at all. It's insane that these made it through peer review👇

Academia isn't a calling—it's a job. Stop glorifying burnout. Clock in, do great work, clock out. Your worth isn't measured in unpaid overtime

New post! There was a lot of innovation in medicine and biomedical research this year, and I've tried to summarize the biggest ones in this blogpost. Medical breakthroughs in 2025: scientificdiscovery.dev/p/me…

I've never felt this much behind as a programmer. The profession is being dramatically refactored as the bits contributed by the programmer are increasingly sparse and between. I have a sense that I could be 10X more powerful if I just properly string together what has become available over the last ~year and a failure to claim the boost feels decidedly like skill issue. There's a new programmable layer of abstraction to master (in addition to the usual layers below) involving agents, subagents, their prompts, contexts, memory, modes, permissions, tools, plugins, skills, hooks, MCP, LSP, slash commands, workflows, IDE integrations, and a need to build an all-encompassing mental model for strengths and pitfalls of fundamentally stochastic, fallible, unintelligible and changing entities suddenly intermingled with what used to be good old fashioned engineering. Clearly some powerful alien tool was handed around except it comes with no manual and everyone has to figure out how to hold it and operate it, while the resulting magnitude 9 earthquake is rocking the profession. Roll up your sleeves to not fall behind.