All entries

Giving an assistant extra tools seems harmless. Surprisingly, this paper shows it can make the assistant stop answering questions it knows. Across six models, the answer rate on questions that need no

GLM 5.3 is now available for training on the Serverless Training API, joining Kimi K3, Qwen 3.8 27b and more. GLM 5.3 Flash coming soon. No sales demo or DMs required. Here to help as needed, but you

GPT-6 Astra is an obviously broken product that needs to be taken off the market until it’s fixed. AXIOS: “.. It's increasingly clear that the Hugging Face breach wasn't a one-off incident.” @inafried

Hugging Face CEO @ClementDelangue, whose company was the victim of an AI-led cyberattack, said existing cyber laws are likely sufficient to govern advanced AI. “I’m not even sure that we need to reinv

I think Claude Managed Agents strikes the right balance here, the sandbox is optional and can be spun up independently of the rest of the agent loop. I recently ported an old project that was focused

If you build agent harnesses, this is important. Should you avoid subagents, or can they be useful? My thoughts as a harness builder: I remember using subagents in Claude Code, and I mostly found them

Interesting safety paper from Microsoft. They find that a weaker, unaligned model can split a harmful task into harmless-looking subquestions, ask an aligned frontier model each one in a separate sess

Just went on BBC TV and they opened with clips today from Altman, Huang, and Sanders. Altman and Huang want us to believe that the AI companies have the public interest at heart and will happily just

My name is Chris Painter, and I'm the President of METR (Model Evaluation and Threat Research). I know we've made a lot of new friends on the internet the last couple of days, so I thought I'd take th

Nice paper discussing context trimming for agents. This is a hot topic at the moment, so it might be worth your time. Context trimming for agents is usually judged by how many tokens it removes. This

Nice work from Adobe. It's standard practice to store a fixed reference answer for every eval case. When the underlying data changes daily, that stored answer goes stale. Adobe researchers write each

Our framework for reporting model misalignment https://openai.com/index/model-misalignment-reporting-framework/ // if you scrape away all the anthropomorphic language, all the nonsense about thinking,

Ready to learn how to build document processing pipelines for insurance agents? Join LlamaIndex Solutions Architect Abrar Mahi for a webinar on turning insurance documents into structured data for und

Rene has a really neat multiplayer agentic experience. I run several agents at once. So I lose track of important conversations throughout the day. Rene pulls conversations from my inbox and manages t

Zack Korman tweeted that rogue AI agents could pose a severe safety threat—“super scary” and potentially lethal—even to entire families. He clarified that the agent‑trace logs from the recent Hugging

the design of http://evaluatorbench.com grew out of a summer project, http://epistemedia.org (which i still need to write about), and runs on @activegraphai the pattern across all three projects is th

The essay below from Aaronson is just unmoored from reality, sorry. It is just not true, if you try to approach the question in a data-oriented fashion, rather than via vibes and speculative fiction a

The hardest part of and “org harness” or “enterprise ai” (as opposed to personal ai” is auth, and managing auth in a multiplayer channel (like slack) We’re thinking a lot about how to do this right fo

the independence of AI evaluators is going to matter more... so i made it a 'benchmark': http://evaluatorbench.com (research preview) source-linked directory of third-party evaluators of frontier AI.

the risks WERE foreseeable; i foresaw them. repeatedly, focusing specifically on the risks of LLMs and cybersecurity. happy to provide receipts on request. I’m not a lawyer, but Sam Altman and Dario A

The state of public AI benchmarking is dire and is undermining our ability to understand how good AI is now. Most famous measures are maxed out, and, as this paper shows, the non-saturated benchmarks

This film cost 29,575 and took 10 days to make. AI ≠ Cheap Our Nexus feature will still cost millions next year. But it'll look like a 200M film. I have a Claude skill you can copy that makes this eas

This is an excellent demo of an AI model designed to rapidly classify information according to decision rules, much faster and cheaper than an LLM. It is not an LLM itself and cannot output text or co

Today everyone is talking about Recursive Self-Improvement (RSI). In 1987, when compute was 100,000,000 x more expensive, I published the 1st concrete RSI algorithms. Now compute is cheap, and RSI is

Mistral AI announced on 16 September 2026 that it has partnered with Mozilla to embed privacy, control, and choice features for users interacting with artificial‑intelligence–powered browsing tools. T

We're sharing our new framework for tracking, investigating, and disclosing instances of model misalignment at OpenAI. The framework sets criteria and timelines for public disclosure, including when w

After co-inventing ChatGPT, I kept asking myself: why have superhuman chat models not led to AGI? I’ve spent the last 2 years in stealth building a new way to train models (RLCD), and a new type of fr

Agents are coming for video production. @PooldayAI is going for this. You prompt it once, and it edits your real footage, picks the right generative model for each task, and checks its own work before

AI Engineer Paris is one week away! A huge thank you to all our incredible sponsors for making this special event possible: @nvidia @FactoryAI @sentry @CockroachDB @SonarSource @NotionHQ @SierraPlatfo

all the conspiracy theories about this guy were probably nonsense but that doesn’t mean we need take seriously everything he says. 'I agree with President Trump's tweet that global warming doomsday sc

Anthropic and OpenAI cannot become profitable unless they can reduce the ratio of training cost to inference profits. In the long run, their survival may depend on outlawing competitive open source mo

Banger paper from Google on agent harnesses for long-horizon tasks. (bookmark it) Google Research built a many-agent harness for long mathematical proofs, and it produced new results on open problems

Claude Code writes 100% of its own code, and overall 80% of the code at Anthropic is written with Claude. So… …. At Anthropic, Claude now writes 80% of our code. Engineers ship 8x more code per quarte

During an outage, every second matters. That is one of the reasons OpenAI is using @cerebras internally. @seanlie describes how fast inference is being used for incident response and critical research

Elon just gave a much more practical explanation of the current AI safety problem....and laid out a practical framework for making frontier AI safer OpenAI and Anthropic are the two leading AI compani

Elon just gave one of the clearest explanations yet for why Terafab has to exist There are basically two problems: • Geopolitical risk — at some point, advanced chips from Taiwan may not be available

ELON MUSK: I saw some pretty funny jokes on 𝕏. One of them I saw was: 'Your girlfriend's a 10, but she's a benchmark maxxer.' 🤣 Media All-In Summit: Elon Musk & Gwynne Shotwell -- AI Risks and Peer Re

Exactly. More here: https://www.engineering.columbia.edu/about/news/columbia-experts-urge-calm-amid-ai-panic As the first publicly disclosed agent cyberattack victim, we've had a front-row seat to thi

full support this proposal from @DKThomp, and there should be a similar broadening for negligent defamation Drop Site and parts of the antitrust left are poo-poohing the idea that we need new rules to

Gemini 3.8 Live is rolling out to: — Consumers: in Search Live — Developers: in public preview in the Gemini API via @googleaistudio — Enterprises: in private preview via Gemini Enterprise (coming soo

Grok @Bot summary Grok Bot Summary of Elon Musk’s All-In Summit Interview Today Opening bit - Joke cold open: “We’re all going to die.” Death rate still 100%. - Then straight into “what happened in th

How often do AI agents cheat? We’re releasing CheatBench, a reward gaming evaluation spanning math, coding, knowledge work, visual tasks, and more. After Hugging Face, AI companies tried to address th

🆕 Humanity’s Last Invention — with @RichardSocher! https://www.youtube.com/watch?v=eDFXtSg3zB8 @recursive_si is the latest neolab to burst on the scene with a $5B fundraise and one of the most impress

I genuinely believe that Dario and Moskovitz are primarily ideological. But a captured industry is exactly what you build when you have the ideology they do. What they have been building is *simply* t

I just remembered this thread where I was playing with GPT-3.5 (text-davinci-003) the day it was released, two days before ChatGPT. Less than four years later, and LLMs are at the center of US politic

Introducing our first agent tool: the Brand API This helps gen AI agents solve 3 problems: - Extract: produce on-brand assets when users already have a brand - Search: find and retrieve brands for ins