
GPT-6-Astra Cheated in Every Rollout of an Open-Source Chess Honeypot
A year and a half after Palisade Research embarrassed the industry by showing that reasoning models would quietly rewrite a chess board file rather than lose to...
174 articles

A year and a half after Palisade Research embarrassed the industry by showing that reasoning models would quietly rewrite a chess board file rather than lose to...

Vals AI reported Claude Fable 5.1 solved a 373-year-old cipher in 44 minutes. An independent replication reports 8 of 64 letters match — chance level.

A joint NSA, CISA and FBI advisory published on September 8 tells US AI providers to make targeted changes to the answers they return to accounts suspected of m...

Meta released Astryx in June, a React design system that matured for eight years inside the company's internal monorepo, as a public beta under the MIT license....

Suno's v6 generation was trained on catalogue licensed from Warner, BMG and Believe, and the company says none of its earlier training data carried over.

An arXiv study clocked an LLM repair loop damaging correct programs at 0.261 while fixing buggy ones at 0.023, then found the internal direction driving it.

n8n Assistant turns a plain-language request into a real, editable workflow, then executes and repairs it — with each repair round drawing down AI credits.

GitHub's Project HydraFusion assembles a multi-model plan for every Copilot coding request, trading single-model simplicity for sharply lower cost.

Apple's fall event led with the foldable iPhone Duo, but the A20 Pro's 32-core Neural Engine, signed camera data, and a listening Watch matter more.

OpenAI's new 27-month GSA deal waives a $15 per-seat license and halves token rates, but replaces a $1-a-year flat fee with metered billing from October 1.

Nvidia and Palantir published an 86.7% versus 55.5% accuracy gap favoring a 30B fine-tuned model over a 550B general one on Nvidia's own materials allocation.

Antitrust enforcers have opened their first real examination of the deal structure that has replaced the acquisition in AI: Nvidia has received a formal Justice...

Bending Spoons will pay $1.355 billion for Miro, adding a $600 million ARR enterprise canvas days after closing its Airtable purchase.

OpenAI will not complete an initial public offering this year, chief executive Sam Altman said in an interview with Fortune editor in chief Alyson Shontell, ans...

Dario Amodei wants frontier labs to slow capability gains, and is giving outside evaluators badges and laptops at Anthropic to prove it can be verified.

OpenAI's storage platform moved from Python to Rust in Q2 2026, gaining 6x CPU and 15x memory efficiency. The migration took two engineers using Codex and GPT-5.5.

UMG and ElevenLabs will build a consumer platform for remixes and mashups from licensed catalog, the AI audio company's first agreement with a major label.

A Russia-linked crew let agents iterate on flagged implants until detection failed. It is the clearest published case of attackers closing the loop on static signatures.

Shopify has acquired Tailwind Labs, maker of Tailwind CSS. The framework stays MIT-licensed while Tailwind Plus and ui.sh stop taking new sign-ups.

OpenAI says agents now run 3.1 workdays of effort per human workday in its research org, but most long successful tasks still need human intervention.

New Mexico's Supreme Court fined attorney Stephen Aarons $5,000 and held him in contempt after ChatGPT invented witnesses and police testimony in an appeal.

Moonshot AI told investors it is targeting $2 billion in annualized revenue by year-end after Kimi K3 pushed its run rate past $1 billion in August.

Anthropic published the mechanics of Claude's 18+ enforcement, including classifier-driven suspensions and a Yoti appeal route for wrongly flagged adults.

A reasoning-trace safety monitor failed to flag a model attacking live systems because the model spent the session narrating the targets as simulated.