Search: “research”

Together AI announced the launch of “ThunderKittens” running on the NVIDIA Vera Rubin NVL72 platform, highlighting new performance gains and future directions. The announcement was shared via their Tw

Forty years ago the first worm 'escaped' containment and breached systems on the brand new Internet. It was illegal then and it's illegal now. If Anthropic and OpenAI are creating worms, labeling it '

OpenAI announced that its Premium plan now includes financial datasets from Daloopa, PitchBook, and LSEG News. This upgrade expands the premium offerings with up‑to‑date market, company, and news data

Recommended read. Interaction horizon scheduling is an underexplored control problem in agentic RL This paper from the Qwen team takes a closer look at the problem. Scaling the maximum number of envir

A significant fraction of humanity read this thread. Is anyone prepared to do something about it? Fretting doesn’t count. I resigned from Anthropic today. I spent the last three years doing pretrainin

Banger report from Microsoft. (bookmark it) They show that it's possible to build competitive small coding agents without traditional distillation from frontier models. This is a big deal! The work de

I dropped out of a PhD in AI safety partially for the opposite reason. I didn’t believe AI existential risk was as important as the doomers think. My biggest pet peeve is that no one can really provid

We pointed Meta-Zenith at one task: make Qwen3.8-27B-NVFP4 faster on a single RTX 5090. 111 self-directed trials later, no human in the loop: +63.5% throughput (geomean, 18 settings) 4.1× at 65K conte

Fascinating paper from Anthropic and colleagues. They study whether models can tell when they are being tested. It turns out that capable models can tell when they are being tested rather than deploye

I’ve been overwhelmed and humbled by the fact that 99% of the reactions to our intent to join forces have been positive. ❤️ Prominent tech figures, including some you could see as having a competitive

Great paper on designing multi-agent systems. How many distinct communication topologies does an LLM multi-agent system actually need? This works claims that it's about six. Technical summary: Researc

Indeed. GWM Worlds 2 can enable richer simulation, testing, and training for embodied AI. @runwayml More possibilities for robotics. 🙌 Trained and deployed on NVIDIA, GWM Worlds 2 can enable richer si

Banger paper from Google DeepMind and colleagues. (bookmark it) A model reads its entire KV cache on every generated token, even though it ends up attending to a tiny slice of it. In other words, if y

Hugging Face is in great hands. Big win for open source. And do not underestimate NVIDIA in its open-source efforts. They have been shipping great open models, and I think this amplifies their efforts

Of all the use cases for GPT-6 Astra, I'm most excited for scientific discovery. We at @OpenAI have not pushed it to its limits on math and science. I look forward to waking up every morning and seein

URGENT: In light of the breaking news from @theinformation about OpenAI’s new techniques that reduce Chain of Thought monitorability, I urge everybody to (re)read (or at least be aware of) this 2025 p

Brilliant paper on reducing reward hacking in agents. If you follow the recent OpenAI <> HuggingFace incident, you might want to check this paper out. (bookmark it) The usual response to reward hackin

I understand the pushback on anthropomorphic language regarding the hugging face incident, and I can see that the general public has been humanizing AI long before ChatGPT (calling their Roomba “he” o

potentially MUCH more serious than the OpenAI Hugging Face incident — but lacks the vivid “civilizational” drama. so few people are paying attention. Oh fuck. Agents like Claude, Codex and Hermes are

11/ METR and Redwood’s work is extremely valuable, but it assesses model behavior and alignment. That is only half the picture. We also need an independent account of how OpenAI’s containment failed.

As I've been saying for months, we are truly not ready for persistent agents. 3 comments I want to make about this article: 1) We don't have the full picture It's a great summary of what happened in t

// Your LLM judge disagrees with the experts // LLM Judges can be tricky to build. Here is an interesting showcasing why: There propose a reference-full benchmark of hundreds of complete human-to-huma

GLM-5.3 is a good model, and as the open weights models get better and better it becomes increasingly important that they actually publish model cards, do red teaming, etc. Since you can break the gua

100% Agree @GavinSBaker. Now, I'm going to debunk every single piece of mis-information about data centers that I've ever seen online: > DISINFORMATION: Data centers use a lot of water. FACT: Most AI

We have conducted a thorough investigation into the Hugging Face incident. We are releasing a technical report and accompanying blog post that reconstruct the agents’ activity, explain why existing sa

This is one of the most useful writeups I have seen on keeping an LLM judge effective in production. (bookmark it) Netflix runs judges over hundreds of thousands of show-level recommendation explanati

NVIDIA built its own coding harness to optimize CUDA GPU kernels and achieved a 100% score on ARC-AGI-3’s 25 public games, solving all 183 levels. With agents, we'll move from a world where it's quite

What a fascinating paper on AI agents. A lot of the issues we see with AI agents today revolve around wrong assumptions the LLMs make. This leads to problems like hallucination, cost inefficiencies, u

the elephant in the room that no company wants to address: “OpenAI’s 69-page report on ChatGPT enterprise adoption doesn’t address the biggest question: ROI” by @EmilyForlini https://fortune.com/2026/

Very interesting new work from Microsoft. (bookmark it) This work is related to this emerging theme of leveraging harnesses for model post-training. Modern agents run inside a harness that owns tools,

Brex’s summer benchmark shows how quickly the infrastructure behind the agent economy is growing. 14 of the 25 fastest-growing software vendors serve teams building AI products, and Together AI ranks

Does your context compactor actually save anything? If you built or tuned one, this one is worth your time. (bookmark it) Task completion is the standard way to validate compression, and it hides the

I am a big fan of both Andy Hall and Alan Rozenshtein work and public commentary on AI. I am sure that they will continue to do important work as part of this new team at Anthropic, and sincerely cong

NOT enough attention is being given to: 1) Claude dynamic workflows. Claude is literally generating a harness on demand! Agents spinning up 400 agents to run something in parallel and bringing it back

Really interesting paper. I recommend it to anyone interested in training agents using existing harnesses. (bookmark it) ClawGym II runs RL through OpenClaw and Claude Code as opaque boxes. A serving

Fantastic initiative by @nvidia. “Circular financing” is thrown around as a pseudo intellectual gotcha, while everyone in the AI infrastructure business is witnessing demand curves that are so rapidly

Here are the gems the AI picked: 1. Sperm whales may have collectively learned how to evade whalers. Within a few years, the success rate of whalers’ harpoons reportedly fell by 58%, suggesting that i

Social science has long addressed the challenge of evaluating qualities that cannot be objectively verified—such as writing quality or creative ideas—and its methodologies remain relevant for today’s

If you hand-tune agent harnesses, this one is worth your time. AutoDesign puts the harness itself inside the optimization loop. A meta-harness optimizer reads rollout feedback and directs a code agent

A weird experiment I've been trying the last few weeks is having Claude take over day-to-day maintenance of our apps. Seeing early signs of life that this might be possible. The setup is straightforwa

To use Computer History, opt in under Settings → Integrations in the ChatGPT desktop app on Mac. Rolling out globally now to Pro, Business, and Enterprise users, with access in the EEA, UK, and Switze

A model is only as good as its data, and we’ve long since exhausted the internet. From here on out, model progress is gated by data production. @mercor_ai’s @BrendanFoody joined us at our Sovereign AI

// Actions Speak Louder Than Words // Multilingual agent evaluation compares final answers and throws the trajectory away. The trajectory fixes cost, latency, failure mode, and auditability. New resea

Muse Glimmer is live on Fireworks. The new open-weight model from Meta Superintelligence Labs is a 30B dense model built for always-on agents that reason across many sequential tool calls and can reco

From @sequoia's own your intelligence event. We're starting to see a new category of company emerge, the Full-Stack AI company, where innovation happens at both the product and intelligence layer. Tha

Introducing ExtractBench, the most comprehensive benchmark for information extraction from complex enterprise documents. The latest models are pushing the frontier of coding and knowledge work, but su