1️⃣We are excited to open-source syftr: a powerful tool for automatically finding Pareto-optimal generative AI flows! syftr searches a large search space of agentic and non-agentic flows to surface optimal tradeoffs between accuracy, cost and latency.
🧵
Microsoft Copilot SDK 2.0 is here.
Build faster agents with a new Rust core, serve many users from one shared runtime, and resume sessions across restarts, with control over each session’s identity, tools, and permissions.
🔗https://t.co/ygdoz4kbVI
The End of AI Frontiers https://t.co/rQOIpmthGr .
There's a bit of a train wreck here that's painful to express. Nevertheless, it's a significant event for me and quite a few other folks involved so I wanted to memorialize a bit.
@JohnCLangford One could use these insights to bias an architecture search around promising parts of search space.
With some more compute one day I would like to do a trilevel optimization: RTL, kernels, neural architecture
My salute to the Indian pilot Captain Smit Machchhar for his extraordinary bravery.
Despite being stabbed and seriously injured, he fought back, resisted, opened the cockpit door, and enabled passengers and crew to overpower the attacker — preventing a catastrophic mid-air disaster.
He saved the lives of 174 people, including Israeli citizens and other nationals.
I wish Captain Machchhar a speedy and full recovery.
He is a true hero.
🥳 Excited to share that @lossfunk got 4 papers into NeurIPS main conference.
Please join me in congratulating the authors!
Aman @arcaman07, Sushrut @martisamuser, Abhinav @MajorTimbWlf21, Pranav - @Avg_sapient
First authors are all undergrads :)
I would like to clarify a few things:
1) The screenshot is my reaching out to Levent to coordinate our releases. I hope it’s clear from the message that we came in with the best possible intentions.
2) I never ever asked for Levent to be removed from authorship of his own work (as indicated by my text). I was surprised to learn during the call with Tristan that they had only solved Euler and not Navier-Stokes; after learning this we brainstormed possible paths forward. One option we discussed was that Tristan could be the lead author on a rewrite of OpenAI’s Navier-Stokes proof. It is in that context that I said “it would be simpler if Levent was not an Anthropic employee” because I felt it would be inappropriate for an Anthropic employee to author OpenAI’s work. Importantly it was admitted that internal Anthropic models had been used in their proof of Euler blowup; I therefore felt I could not consider Levent to be an independent academic. Another option I wanted to propose (but got cut short) is to offer access to our internal model so that they could try to finish their proof and bridge the gap between Euler and NS. Again I did not know how to navigate giving access to internal OpenAI IP to an Anthropic employee.
3) To reiterate it plainly: as my text clearly indicates, and as I said during our call, OpenAI's intention was to do everything possible to celebrate their mathematical achievements and the heroic efforts that they made on Euler. In the call I was immediately met with a litany of slander, including direct threats that if we were to announce Navier-Stokes he would immediately go to the press with a barrage of unfounded accusations. I refuted all these accusations but he replied “there is nothing you can do, I simply do not trust you”. I was confused why one would turn an incredible source for celebration (of their achievements!) into such bickering, which is when I said that I did not understand why one would risk their career [over unfounded accusations]. Genuinely, at that moment, I was trying to care for him and do a last ditch attempt to get a chance to give them all the credits that they deserve. I deeply apologize for this extremely poor choice of words, it is the opposite of what I was trying to convey. (I should say that I retracted them on the spot by the way.)
4) Overall, on a personal level, it was incredibly difficult to have these conversations. Levent refused to attend any of the meetings despite my repeated asking. As Sholto Douglas said, there will need to be coordination between Anthropic and OpenAI in the future; I felt I was doing a proxy negotiation with Anthropic while the Anthropic employee refused to directly participate.
Many startup employees do not recognize the sheer litany of ways that a founder can screw you over without you even knowing. Founder trust is one of the most important things to look at when you’re joining a startup.
From cutting you out of M&A, screwing your retention pool, overdiluting your equity, firing you before your cliff, not having attractive options exercise plans, poor 409a price management, not telling you about QSBS / early exercise, blocking you from participating in secondary, obscuring company performance and many more.
Many many decisions that are made in rooms you are not in as an employee where the only thing that matters is: “does the founder have your back?” Great startups with untrustworthy founders lead to poor outcomes and often good startups with trustworthy founders lead to great outcomes.
Pick wisely.
We’ve seen an explosion of new ML chips with unique architectural features, but software support remains the critical bottleneck
Achieving peak performance increasingly relies on hardware-specific optimizations in the kernels, but we observe that coding agents are particularly weak at this
Introducing Hawkeye, a framework that brings hardware-awareness to coding agents by grounding them in a minimal and comprehensive taxonomy of optimization strategies
For new GPU or ML accelerator architectures, you only need to write 10 unit tests and solution kernels (one per optimization strategy), and we show that coding agents can effectively scale test-time compute with this minimal supervision to write hardware-aware kernels
Hawkeye can port kernels across architectures (Ampere, Hopper, Blackwell), vendors (NVIDIA, AMD), and precisions (FP8, NVFP4, MXFP4) while consistently leveraging hardware features and approaching expert kernel performance
Work co-led with @keramakr and done in collaboration with Alexander Ingare @simonguozirui@18jeffreyma@ZishenW@simran_s_arora@Azaliamirh@profvjreddi
Grok 4.6 was the first model trained on internal model-development tasks.
My team built the training and evaluation stack that enabled Grok to learn from work that accelerates model development itself, including production inference and kernel optimization.
Grok 4.6 now leads our internal MTS Eval and InferenceEval. In one experiment, we gave an earlier checkpoint our production inference codebase in a fully automated environment. It explored 297 optimization ideas and shipped 3 changes on top of an already human-optimized stack, improving prefill throughput by 3.1% and decode throughput by 1.5%.
Our work is featured in the R&D Enablement section of the Grok 4.6 model card.
Our latest code model is 25% more efficient, higher quality, and a quarter of the cost than our model launched in June. Now live in GitHub Copilot - try it out.