Extropic thermo silicon hardware alpha. Shipping Summer 2025. Apply now.

What would truly open-source AI look like? Not just open weights, open code/data, but *open development*, where the entire research and development process is public *and* anyone can contribute. We built Marin, an open lab, to fulfill this vision:

If you divide 1 by 998,001 you get all three-digit numbers from 000 to 999 in order, except for 998.

II-Medical-8B-1706 is our latest state of the art open medical model 💡 Outperforms the latest Google MedGemma 27b model with 70% less parameters 🤏 Quantised GGUF weights, works on <8 Gb RAM 🚀 One more step to the universal health knowledge access that everyone deserves ⚕️

Hello Gemini 2.5 Flash-Lite! So fast, it codes *each screen* on the fly (Neural OS concept 👇). The frontier isn't always about large models and beating benchmarks. In this case, a super fast & good model can unlock drastic use cases. Read more: blog.google/products/gemin…

1/ 🔥 AI agents are reaching a breakthrough moment in cybersecurity. In our latest work: 🔓 CyberGym: AI agents discovered 15 zero-days in major open-source projects 💰 BountyBench: AI agents solved real-world bug bounty tasks worth tens of thousands of dollars 🤖

Wrapped up Stanford CS336 (Language Models from Scratch), taught with an amazing team Tatsunori Hashimoto Marcel Rød Neil Band Rohith Kuditipudi. Researchers are becoming detached from the technical details of how LMs work. In CS336, we try to fix that by having students build everything:

As frontier AI systems increase in capability and agency, the risk of AI-driven cyberattacks will likely rise sharply. Tasks once done by elite hackers may soon be carried out autonomously, and this demands urgent attention.

Very cool thread about the CS336 Language Models from Scratch course at Stanford taught by Percy Liang et al. Makes me wish I was a student again!

AlgoBench is extremely tough, with agents not finding substantial speedups on most tasks. But sometimes these agents do really cool things: here, the agent realized that it could solve this convex optimization problem with a scipy function, leading to an 81x speedup.

Announcing DeepSWE 🤖: our fully open-sourced, SOTA software engineering agent trained purely with RL on top of Qwen3-32B. DeepSWE achieves 59% on SWEBench-Verified with test-time scaling (and 42.2% Pass@1), topping the SWEBench leaderboard for open-weight models. Built in

Every $ invested in open-source generates $2,000 of value. Might be orders of magnitude more for open-source AI! x.com/clementdelangu…

n-simplex attention makes incredible sense because of its honesty: it literally says you can put more compute on attention operation to get more gains: we've seen this trend so many times. This differs from lot of 'suspicious' claim, such as you can use less compute to perform

Design and Simulate a fully functioning 8 bit CPU.

curious of why so many feel that Y Combinator has slipped from some sort of trusted gold standard to sketchy fast-food franchise? beyond the typical batch dilution & weak leadership arguments, is there anything deeper here ?

design for scale favor simplicity let data (scarce, so prioritize first) and computation (second) do the heavy lifting ht: Richard Sutton

Here's another proof of concept example of a lethal trifecta attack: if you combine the Supabase MCP with another MCP that provides exposure to untrusted tokens and a way to send data back out again - in this case a support ticket system - attackers can steal your Supabase data

Countering Anger Outburst as Power Play An explosive display of anger hijacks primal threat circuitry, spiking cortisol and narrowing attention, so compliance feels safer than resistance. Studies show it reliably extracts larger concessions in negotiations. By framing the

Researchers just published an unbelievable paper in Nature that uses purely brain signals to let paralyzed people speak with intonations! It's instantaneous (~25ms), medium pace (40-60 words per min) and fair accuracy (~60% word error rate). Baby step to full telepathy.

I have an alternative proposal.