The most digestible post on SAEs for mechanistic interpretability I’ve come across. Highly recommend!

Baby step to AI editorial parternship NEJM AI : Our popular AI Grand Rounds podcast now has Human-in-the-Loop summaries ai.nejm.org/doi/full/10.10…

Publishing: What is the right mix of editorial human experience & taste & AI assistance? NEJM AI we are at the very beginning of that exploration. ai.nejm.org/doi/full/10.10…

Check out our latest work! 👇We leverage pretrained protein generative models (here Chroma Generate:Biomedicines ) as a prior for inverse problems in protein space (e.g. structure completion, distance contraints, cryo-EM model building). ❄️🏗️ Paper: arxiv.org/abs/2406.04239

In this episode of AI Grand Rounds, Drs. Rohaid Ali, MD and Fatima N. Mirza, a married couple and chief residents Brown University, discuss their innovative work applying AI to health care. Listen to the full episode hosted by the editors of NEJM AI: nejm.ai/ep20

We had a blast! Thank you to hosts Arjun (Raj) Manrai and Andrew Beam for inviting us onto the podcast. For regular listeners, let me tell you: the lightning round is not at all what you expect!

Editorial by Drs. Arjun (Raj) Manrai and Andrew Beam: Human-in-the-Loop AI Summaries of NEJM AI Grand Rounds nejm.ai/45HhFmG #ArtificialIntelligence #AIinMedicine

Excited to introduce RadGraph2: A new dataset for tracking disease progression in radiology reports! 🏥📊 Key features: 800 expert-annotated reports 220k+ machine-labeled reports Captures changes in:Disease onset, progression, improvement & resolution Medical device placement &

What can the history of medicine teach us about how LLMs and other AI models may change care going forward? 'Historian-futurist' Adam Rodman, Andrew Beam, and I share some thoughts here based on our NEJM AI Grand Rounds conversation. Full text: ai.nejm.org/doi/10.1056/AI…

Ever looked at the attention operation and said "hang on, that's a one-point function!"?

We are building one of the largest GPU clusters in biotech and are recruiting ML engineers! If you have experience making GPUs go brr and would like work on exciting problems at the intersection of AI and autonomous science, please consider applying: job-boards.greenhouse.io/flagshippionee…

We’re hiring!

It was great to dive into academia vs. industry, biotech, and the future of AI in healthcare with Vijay Pande and Andrew Beam . Episode out now!

Thank you Andrew Beam and Arjun (Raj) Manrai for a particularly engaging chat on all things AI + Bio & health on the NEJM AI pod.

Great to chat about echocardiography, AI, how to launch RCTs, and ggplot (!) with the inimitable David Ouyang, MD for NEJM AI Grand Rounds. Episode out soon!

Such a fun conversation with Andrew Beam and Arjun (Raj) Manrai! So many friends and interests in common! NEJM AI

Excerpt from a video I just posted on YouTube (about the multilayer perception layers in a transformer, and how LLMs may store facts).

Diffusion is the rising tide that eventually submerges all frequencies, high and low 🌊 Diffusion is the gradual decomposition into feature scales, fine and coarse 🗼 Diffusion is just spectral autoregression 🤷🌈