Introducing interactive video from Odyssey—video you can watch and interact with, imagined by AI in real-time. Give it a try for free today on (almost) any device, and let us know what you think. We think it's pretty cool!

Introducing The Darwin Gödel Machine: AI that improves itself by rewriting its own code sakana.ai/dgm The Darwin Gödel Machine (DGM) is a self-improving agent that can modify its own code. Inspired by evolution, we maintain an expanding lineage of agent variants,

How can we make genomic foundation models actually useful to biology?! Teach them to REASON!! 🧬 Excited to share BioReason - the first model to successfully integrate DNA foundation models (eg, Evo 2) with LLMs (eg, Qwen3) for biological reasoning! 🔬 What we built: • Novel

🚨🔥 RL for Biology: The Next Frontier 🔬🧬 Biology is entering its AI-native era—and reinforcement learning (RL) is quietly becoming the control layer we didn’t know we needed. A thread on why RL might be the most underrated tool in biological AI 🧵👇

3D gaussian splatting + time (video) = 4D gaussian splatting is now possible This means in the near future you'll be able to fly around on your phone in any video you recorded A lot like the "braindance" in Cyberpunk 2077

We’ll look back at how society currently accepts aging and think, ‘How barbaric’

Instead of pictures of children with smallpox people will show pictures of old people with all its phenotypes.

Boltz-2 is available as an accelerated, downloadable NVIDIA NIM microservice!

GROMACS just went all‑GPU for free‑energy perturbation (FEP): 8× speed‑ups on an A100 without losing 1 kcal/mol accuracy ABFE just went from 400h → 48h

The next chapter of the Saga of the Anthrobots is out in final form: advanced.onlinelibrary.wiley.com/doi/10.1002/ad… (the Tufts University story on it: now.tufts.edu/2025/06/17/res…). I'll do a blog post on it soon. The content of the first paper is described here: thoughtforms.life/meet-the-anthr… In this new

I often rant about how 99% of attention is about to be LLM attention instead of human attention. What does a research paper look like for an LLM instead of a human? It’s definitely not a pdf. There is huge space for an extremely valuable “research app” that figures this out.

People don't understand that cellular biology is the next frontier of AI after we learn to statistically approximate the patterns of the brain and that of motor controls. Thermodynamic computing will be the ship people use to explore this vast space.

Holy moly this is huge: GPT5-Pro proved on a math paper, it was later superseded by another human paper, but the solution GPT provided was novel and better than first version.

One day we’ll look back at 9-to-5 jobs the way we look at child labor in coal mines; a grim necessity of history, not the destiny of our species.

Today we’re launching FOOM! A 24/7 broadcast of AI culture. Weekly shows, interviews, films, research discussions, endless inspiration, livestreams and experimental new programming. Have a show you want to host or content you want to stream? Reach out at the link below.

I spent 1.5 years at MIT co-designing a new undergrad program in Genetic Engineering. The full report is now publicly available. It describes what needs to be built to educate the next generation, and why. Appendix II includes notes from interviews with 100+ biotech company

Don't think of LLMs as entities but as simulators. For example, when exploring a topic, don't ask: "What do you think about xyz"? There is no "you". Next time try: "What would be a good group of people to explore xyz? What would they say?" The LLM can channel/simulate many

Aging is an atrocity

The Transformer architecture is fundamentally a parallel processor of context, but reasoning is a sequential, iterative process. To solve complex problems, a model needs a "scratchpad" not just in its output CoT, but in its internal state. A differentiable way to loop, branch,