Dwarkesh Patel The fact humans can usefully absorb that feedback and use it to drive ongoing improvement. Time spent teaching models doesn’t seem to have such good ROI.

I've recently made two bets (10:1 and 20:1) wagering 500 USD against the a measurable definition (DM me for def.) of AGI to occur withing and 2 years. ARC Prize clearly said that ARC-AGI isn't enough. I'm curious about other peoples odds,definitions and wagers=> DM or reply plz

Always fun when you see frontier model apis can leak your email based of your name if you let them fill in a latex template for a papers first draft with slop lorem ipsum (just sent off gdpr access requests to anthropic and openai about this, I'd sure like to see my PII)

🚨 New Paper Alert: Open Problem in Machine Unlearning for AI Safety 🚨 Can AI truly "forget"? While unlearning promises data removal, controlling emergent capabilities is a inherent challenge. Here's why it matters: 👇 Paper: arxiv.org/pdf/2501.04952 1/8

Introducing BenchBB and the community paper of the Protein Design Competition 🎉 We wrote a community paper about our Protein Design Competition, teaming up with your favourite protein designers from both rounds. Excited to announce that this is now on Biorxiv 🙌

Thanks to the efforts of all the amazing people in the protein design community we managed to write up a cool paper after our protein design competitions at Adaptyv. One of the key takeaways was: to design better proteins we need better models and those models need better data.

Really excited about this! It was fun to be a (very minor) contributor on it

Excited to finally share a project I've been working on for the past months! Protein designers have been pushing the boundaries of what is possible, and it's exciting to see a company like Adaptyv Bio empowering designers :)

Anthony was the one pushing us over the edge to create the paper with his call for a writeup and was one of the people keeping up the momentum, thanks a lot for all the help :-)

We had a great time and learned so much contributing to the post-competition analysis of Adaptyv Bio's EGFR binder competition, working alongside an amazing consortium of authors from both academia and industry! 🎉 The preprint on crowdsourced #protein design, featuring

We're delighted to announce that our conference "Protein Evolution, Design and Informatics Edinburgh 2026" will be running from the 13th-15th of May. Register interest here and please retweet! biochemistry.org/events-and-tra…

no kings was the largest political protest in u.s. history

Pro-1, a protein design model by Michael Hla, doesn’t just propose mutations — it explains why it made them. We tested 19 of its FGF-1 designs in our lab and 3 of them improved thermostability while maintaining binding. In this protein designer spotlight we explain how

Technology = power. AI is reshaping power — fast. Today’s AI doesn’t just assist decisions; it makes them. Governments use it for surveillance, prediction, and control — often with no oversight. Our new paper proposes some ML safeguards to resist AI-enabled authoritarianism:

Co-authored with: Isaac Friend, Keir Reid, TheGermanPole, Vincent Wang-Maścianica, Jakob Mökander, philip Julia C. Morse and Robert Trager

We’re grateful to many partners for making this possible: Oracle Cloud for the massive compute that was required; NVIDIA Healthcare for new GPU kernels and accelerating our sequence search; and Sino Biological Inc. and @AdaptyvBio for conducting many wet-lab experiments in the paper.

Chai Discovery just released their new model Chai-2, validated on Adaptyv! Chai-2 enables generation of de novo proteins and epitope-specific antibodies using a multimodal generative architecture that integrates all-atom structure prediction and generative modeling. Chai

Want to learn more? -> Check out Chai's report chaidiscovery.com/news/introduci… Want to validate your own models and test your proteins in the lab? -> Just go to our website and set up your first experiment adaptyvbio.com

Chief content officer Tudor-Stefan Cotet strikes again with another protein designer spotlight. As an added bonus, this time he's teaching everyone about all the X-PO algorithms are about