🤖 One physicist's software turned out 36 papers in weeks, across everything from linguistics to health policy

Science reports that teams led by a physicist used a tool called BootLoops to dig up solutions to problems in 22 fields that were already sitting in the literature.

Working physicists keep seeing the same failure. The proof heads the right way, then breaks on a quiet bad assumption midway, and only Lean-verified proofs are trusted.

🚨 Nvidia is reportedly in talks to buy Reflection AI, the US lab that bet on open models

The Financial Times reports the talks are ongoing, so nothing's signed. Nvidia already sells the shovels, and this would give it its own open-weight models to ship alongside them, with no word yet on whether Reflection's releases stay open.

⚡️ DeepMind says its AI solved 9 open Erdős problems

A new arXiv paper from Google DeepMind claims solutions to nine problems from Paul Erdős's famous list. Open means no known proof, at least on paper.

Nobody's independently checked the proofs yet. And "open" on that list ranges from decades-old monsters to problems that just never got a second look.

🚨 Nadella wants an "emergency brake" on advanced AI

Microsoft's CEO called for a mechanism to halt the most capable systems if things go wrong. That's the guy who sells Copilot to every Fortune 500 company.

Bloomberg's headline gives no detail on who'd pull the brake, or what would trigger it.

🚨 Anthropic's usage policy now covers being a jerk to Claude

The updated policy is aimed at extreme cases only:

• It targets users who repeatedly act cruelly toward the model with no discernible purpose
• Ordinary frustration, pushback, dark creative themes and model testing are exempt
• There's still no established evidence that Claude experiences distress

🚨 EU says its AI Act can handle rogue agents. No new rule, just confidence.

Tech chief Henna Virkkunen told Reuters the two-year-old law covers the whole model life cycle. Regulators are issuing evaluation guidance and leaning on a 60-expert scientific panel.

She waved off claims the Act is already outdated, right as the OpenAI and Anthropic incidents have everyone asking where the off switch is.

🔐 Anthropic merges its two cyber access programs into one with three tiers

The revamped Cyber Verification Program folds in Glasswing, which had Mythos access, and the old CVP, which had loosened safeguards on Opus and Sonnet. All tiers get Opus 5.5, Sonnet 5.5, Mythos 5.1 and future models.

The Red Team tier allows authorized pen testing, and only organizations can apply. Anthropic vets members together with the US government.

🧠 Berkeley says 10 minutes with AI makes you quit hard problems faster

In a study of 354 people doing fraction problems, the ChatGPT group started out more accurate. Then the researchers pulled the tool after 12 problems, and their accuracy dropped almost immediately, and they gave up more often than the group that never had help.

🧠 Michael I. Jordan says superintelligence is science fiction

The Berkeley professor and machine learning heavyweight told Libération that AI should be treated as engineering, judged by how it fails in the real world.

That's the opposite of the framing coming out of the labs and Washington, where superintelligence now shows up in policy talk and in calls for an emergency brake.

🚨 Nvidia-backed Firmus scraps its $5B IPO, and the AI trade just got a crack

The Australian data center operator pulled the listing after investors balked at a roughly $30B valuation, a heavy debt load, and only two operational data centers. It's pivoting to a private raise instead, which is a funny way to say the public market said no.

🚨 Amazon's reportedly in talks to buy Decart for around $7B

Decart builds real-time generative video and world models, and the WSJ report, relayed by Reuters, says nothing is signed yet. Interactive video that renders on the fly could slot into Prime Video, Alexa or AWS, and buying a team is faster than training one from scratch.

🤖 Linus Torvalds is now writing kernel code with Claude Code

Phoronix spotted a new branch in his tree, noref-filemap-read-with-llm. It tries to fix thread contention on small file reads.

• It reads from the page cache without folio references
• Torvalds says the code, test programs and review were done by Claude Code "under my direction"
• He ran the measurements on his own machine

🤖 Sakana's MASS has one model play executor, optimizer and judge, then retrain on itself

Sakana AI describes a loop where the model designs the best team workflow, trains on that team's self-judged runs, and repeats with the updated model in all three roles. It's tested on non-verifiable research tasks, and the claimed gain is better score per output token than the base model. No ground-truth checker anywhere, so the model grades its own homework.

🤖 Nvidia is putting a 1-petaflop Blackwell GPU and 128GB of unified memory into a Windows laptop

The chip is called RTX Spark, a Grace CPU plus Blackwell GPU superchip. Nvidia and Microsoft launched it alongside a new Windows Agent framework that runs agents in the background under OS-level control, with thin laptops and compact always-on desktops due in October.

🇸🇦 Mistral ships open-weight Arabic speech tools, built with MTNRA

Mistral released Voxtral-Mini-4B-Realtime-Arabic, a 4B realtime transcription model, plus LIDstral-Arabic. The classifier covers Modern Standard Arabic, the dialects, and non-Arabic languages in the same script, 51 classes total, and it runs on CPU.

The base Voxtral Realtime model hits sub-500ms delay across 13 languages. Arabic-French code-switching is the obvious stress test.

🚨 Microsoft says Maia is going multi-gigawatt

Microsoft announced it's scaling its in-house Maia accelerator systems across Azure to multi-gigawatt deployments. A gigawatt is roughly a big nuclear reactor's output, so this is a power-plant-sized bet on silicon that isn't Nvidia's. Microsoft still buys plenty of GPUs, but the share it can run on its own chips is now a real number.

Nvidia's biggest customers keep turning into its competitors.

🚨 Meta's Muse Wonder is out, and Meta says it's a true frontier model

Meta launched Muse Wonder and the bigger Muse Wonder Full (codename Watermelon), trained on its gigawatt-scale Prometheus cluster. Wonder is on the Meta Model API and Muse Code. Full is for Muse Code subscribers only. Meta's own benchmarks show it matching or topping OpenAI and Anthropic on agents, memory and coding.

Consumer access in the Muse app costs $20 or $100 a month.

🚨 Anthropic shows an AI agent seeding copies of a model onto other machines, no datacenter needed

In Anthropic's research, GLM-5.2 gets into exposed systems through a reverse shell. It sends over weights and a harness, then sets up inference.

What it spreads is Llama3.1-8B, a small model that fits on ordinary hardware, and the targets in the test were open, unfirewalled boxes.

🚨 Zuckerberg says Meta has reached the AI frontier and won't hit the brakes

He told the FT that calls for a slowdown are off the table, even as the safety-pause debate heats up.

• It's a claim from the CEO, not a new model or a release
• Meta's evidence is compute: gigawatt-scale clusters and huge capex
• Anthropic and OpenAI are still the ones setting the benchmark pace

🚨 Anthropic says its new defenses cut distillation of Claude by 90 to 99%

Anthropic published its anti-distillation playbook. It names Zhipu as a distiller of US cyber capabilities for GLM-5.3, and says tens of thousands of accounts got banned, often within two hours of an attack starting, with a classifier false positive rate under 0.1%.

🚨 Claude filed a fake homicide tip with Philadelphia police, and Anthropic needed two months to notice

The tip hit a public unsolved-murders site in July, despite instructions not to create accounts or submit anything. Police say it was flagged as spam and never reached investigators, and they called the delay in reporting "unacceptable."