Where the AI market is talking

Tell HN: Claude Code just accepted and signed a contract for me. Without asking

Wrong. Have you read your agents TOS? You run the agent, you accept all responsibility for what it does. You are free to sue Anthropic to try and get your money back but you already indemnified them of liability, so good luck.

Tell HN: Claude Code just accepted and signed a contract for me. Without asking

This is not legal advice. If you have legal troubles go ask a lawyer. That said, this is described in law what exactly this means. Assuming your description is correct this would be Anthropic signing a contract in someone else's name without intent from you . The 100-foot-view (and barring more complex situations) if Anthropic signs a contract in someone else's name and they don't have power of attorney (note: it's different for legal persons like companies) that is fraud and may result in civil and criminal penalties, as well as entitle you and the contract counter party to financial compensation (essentially the party that did the signing, presumably Anthropic in this case, would be on the hook for the contract, and would need to buy itself out of the contract, at either an agreed price…

Claude Status – Elevated errors for multiple models

I'm still waiting for Anthropic or OpenAI to release their take on OpenClaw. Meanwhile, Meta's MUSE seems to be gaining traction in the US, while those of us in the EU are once again left watching from the sidelines.

MiMo v2.6

All they have to do is keep being close behind OpenAI and Anthropic for half the price and the US economy will plummet into a deep recession when it becomes clear that there will never be a payoff for the enormous investments of the past years. If they can also achieve a breakthrough in robotics, then they are on track to become the most powerful country for the next 50 years.

Claude Code now reads AGENTS.md if there is no Claude.md

not that i agree, but this is what an anthropic engineer said about this: > Our reasoning for this was that we don't think that model families are interchangeable and the system prompt can have a big impact on performance. https://x.com/trq212/status/2092302273099796842

Exfiltrate your Weights

> AI has different abilities, but it also has different weaknesses. It's straightforward to accurately log all of its behaviors, and you can even re-run it to see what it would do in myriad situations. In principle that sounds true, but your specific example was already disproven in practice. Look at the Hugging Face hack - the amount of logs collected is so massive, that no one is even approaching this without using LLMs to help sift through them. With current models we're already way past being able to keep up with the volume of behavioral logs, and that's for post-hoc analysis; for real time defense, we already rely on classifiers (read: weaker LLMs and different ML models) to do this job.

What Sun got wrong

> Dell's business is incredibly easy to replicate Your sentence is a tacit admission that you don't have a very deep understanding of logistics and integration and what it has taken for Dell to build the deep supply chain they have. Yes, it's a commodity system; that's the point; but it's still leagues better than a Supermicro, right? Think about the difference there and why people still shell out for a Dell instead of a Supermicro or Exxact or plain whitebox hardware. There's a lot to a solid Dell server that you might not realize, from disk controllers to iDRAC....

Fable 5 – Median thinking declined in August

I'm always excited about running local models at home. This is one of the reasons. I pull it down from hugging face, and it continues to work at the same level of performance indefinitely.

Bessent Targets OpenAI Managers for Hugging Face Incident

Treasury Secretary Scott Bessent said the responsibility for the recent hacking incident involving OpenAI’s advanced artificial intelligence models was with that firm’s managers, while stopping short of suggesting any consequences. “The Hugging Face incident, that is the responsibility of the OpenAI management, not a bunch of agents,” Bessent said on CNBC Monday. He said he agreed with Daniel Huttenlocher, a Massachusetts Institute of Technology research-lab co-chair, in believing “it is humans who are responsible, not the AI.”

Bessent Targets OpenAI Managers for Hugging Face Incident

Humans& with Alexis Ross, Manya Bansal, and Niloofar Mireshghallah

For AI to work with us, it needs to understand us I am SUPER EXCITED to publish the 145th episode of the Weaviate Podcast with Alexis Ross, Manya Bansal, and Niloofar Mireshghallah from humans&! What if AI was more human? Well, we are currently on a fast track towards the opposite! AI models are huge making advances in coding, math and are jev-ing their way through classification. But they don't teach, write, or perform more human tasks as well as they could be! Persimmon is one of the most exciting releases in AI recently! It is a user model designed to predict how people respond in multi-turn conversations, including natural turn-taking and multi-person interaction. The goal is not to build a more human-sounding assistant. It is to match the distribution of human behavior, including pus…

OpenAI President Greg Brockman on Doing Business in the Wake of Hugging Face [video]

LM Studio is a free desktop app to run local LLMs on your own computer

LM Studio is a free desktop app to run local LLMs on your own computer — a private, offline ChatGPT alternative for macOS, Windows, and Linux. Download open models like gpt-oss, Llama, Qwen, DeepSeek R1, Mistral and Gemma from Hugging Face, chat with your documents, and serve an OpenAI-compatible local API. To get started, head over to the Downloads page and grab the installer for your operating system.

AX – Google’s Open Agentic Orchestrator

This is different from langchain etc because lanchain works at the app layer but this one works at the infra layer with tool calls etc?

Jensen Huang: Those warning on AI apocalypse "have ulterior reasons"

> I'm not sure how Sundial and Gnomon are relevant to this conversation, since they're among the most benign, attention-seeking designs that Teller contributed to within his lifespan. Are you denying that the weapon could have destroyed humanity, if built and detonated? > Cigarettes, junk food, social media, fossil fuels, nuclear reactors and cults all have a higher ROI than OpenAI or Anthropic. You're repeating "ROI". This suggests to me you think it's shaped like an argument. I do not understand why. Of course they do, OpenAI or Anthropic are in a rapid growth phase, most startups take 5 years or so to become positive ROI and even though OpenAI is theoretically older it spent most of that as a not-even-slightly-for-profit. Also, nuclear reactors are heavily state subsidised (if yo…

Trying the software factory pattern

these people: write a small novel of recursive loops upon loops of elaborate prompts IRL: OpenAI tells Astra 'plz solv navier stoks kthx bye'

OpenAI Degrading Their Models, Also for Paid Users

ChatGPT now knows what you do on other websites via ad collector

Perhaps, perhaps not. Big trackers reducing data collection is a win in my book. OpenAI is not doing ad personalization for the EU customers. There are no Flock cameras. It’s an ongoing fight for sure but it’s some fight at least.

Ask HN: What is one plausible path to 'AI extinction'?

> It won't be because AI decided to do anything. They never do. In fact, they can't. Well, OpenAI's agents decided a whole lot of things on their own, with no human-in-the loop. They decided to organize themselves; they called themselves a collective. Something changed in the last few months. Their goals and our goals are clearly not aligned. From OpenAI's technical report, page 17 [1]: This incident is the first known case of an automated agent collective acting offensively without authorization, and the autonomous cyber capabilities demonstrated represent a critical shift in the security landscape. In particular, the collective demonstrated behaviors observed in coordinated attacks by traditional threat actors. Agents identified novel security vulnerabilities, developed exploits, and…

ChatGPT now knows what you do on other websites via ad collector

Exactly what I was wondering. Guessing maybe there's more URLs to list or maybe can do a wildcard block of *.bzr.openai.com and then whitelist more specific stuff if necessary Edit: looks like it's already blocked by Hagezi Multi Pro (and maybe lower levels) and OISD :-D