OpenAI publishes ten mathematics claims with Lean proofs and no named authors
OpenAI released ten claimed advances in mathematics and theoretical computer science today, produced by an unreleased internal model it calls Astra, with a 249-page manuscript collection and machine-checkable proofs for every result.
The non-sofic group is the one OpenAI claim a computer can check
Chapter 3 of OpenAI's new manuscript claims to have constructed a non-sofic group, settling a long-open question, and ships roughly 34,000 lines of Lean code with no unproved placeholders so outsiders can verify it.
The EU AI Act's transparency rules start today. The high-risk rules do not.
Article 50 of the EU AI Act applies from 2 August 2026, requiring chatbots to identify themselves and generative systems to mark their outputs as machine-detectable - but the much larger high-risk rulebook was pushed to late 2027 and 2028.
llama.cpp ships the fix that lets DeepSeek V4 Flash call tools mid-thought
DeepSeek's new open-weight model shipped without a standard chat template, and local agents kept dying when it emitted a tool call inside its reasoning; llama.cpp release b10217 and KoboldCpp v1.118 both landed fixes on 1 August.
A month after the Hugging Face breach, there is still no lawsuit
Hugging Face says it rebuilt compromised systems, rotated credentials and reported the intrusion by OpenAI's evaluation agents to law enforcement, but the public record shows cooperation rather than litigation, and no independent investigation has reported.
A judge did not rule that ChatGPT users have no rights to their chats
A New York magistrate denied one individual permission to intervene in the OpenAI copyright litigation, and the order explicitly says the data preservation hold was for a possible spoliation inquiry rather than to hand conversations to the New York Times.
Figure's viral ladder climb is a two-hour stair timelapse
A widely shared clip of Figure's F.03 humanoid was reposted as an autonomous ladder climb; the company's own post describes a timelapse of repeated stair traversal that is helping move its robots closer to full autonomy.
Reddit's US daily users slipped while everything else grew
Reddit reported 130.3 million daily users and 61% revenue growth for the second quarter, but its US daily user count fell slightly from the previous quarter and the company said search referrals were choppy.
China did not give away free models. It built a governance body.
Reports that China offered free AI models to the Global South at a Geneva summit describe a discussion session; the concrete instrument came nine days later in Shanghai with the founding of an intergovernmental AI cooperation organisation.
Seven new papers cannot agree what a world model is made of
PhiZero compresses video into a learned alphabet of state transitions, but six competing late-July papers argue the substrate should instead be executable simulator code, explicit game state, paired re-renderings, distribution coverage, architecture, or action intent.
Qwen trained its phone agent on a lab of more than a hundred real phones
Qwen-UI-Agent's technical report describes a fleet of over a hundred physical Android devices running 150-plus real apps, with a scheduler that leases working phone-app-account combinations and blacklists broken ones until a human fixes them.
AI financial advice works when you hand it a full financial plan
Researchers simulated the lifetime finances of 1,000 demographically representative US adults following chatbot advice and found the results move toward a life-cycle benchmark - but only when prompts supply complete data, fixed assumptions and hard arithmetic constraints.