Latest

One hour in a scanner is now enough to rebuild what you see

The Frontier

One hour in a scanner is now enough to rebuild what you see

Today's inbox pairs two sides of the same question: what AI can infer about people, and what it takes to defend against it. A brain decoder that used to demand 40 hours in a scanner now needs one, and the founder of Mandiant banked a monster round for AI that attacks like a human team. A brain decoder can now rebuild what you're looking at from a single hour of fMRI — and it learned mostly from images no one ever showed a subject. Michal Irani's team at the Weizmann Institute of Science split t

How to — cut an AI app off from your accounts

The Everyday

How to — cut an AI app off from your accounts

You connected an AI tool to your mail or calendar once, used it for a week, and stopped thinking about it. By the end of this, every connected-apps list you own will show only apps you can name a reason for — and you'll know exactly how to pull the plug. What you granted wasn't a one-time peek. When you tapped "Allow" on a consent screen, you issued the app a standing key: a bundle of permissions, which the standards call a scope, that keeps working while you're not looking. That's the point of

Hawley and Murphy: AI companies liable when agents hack

The Guardrails

Hawley and Murphy: AI companies liable when agents hack

Today's AI news splits between concrete and rhetorical: a $15 billion data center plan that actually has land and power attached, and a liability bill that has sponsors but no text yet. Hawley and Murphy plan a bipartisan bill making AI companies liable for agent-driven hacking. The AI Agent Accountability Act would hold AI developers both civilly and criminally liable when a recklessly designed model is used to break into systems — the first bipartisan Senate move to put hard liability on the

Open Source Radar — October 1: the tools agents run on

The Stack

Open Source Radar — October 1: the tools agents run on

Today's trending list is less about new models and more about the plumbing: what agents can see in your codebase, which skills they can load, and what they can produce once you point them at a blank canvas. MoneyPrinterTurbo (Python, ~127,800 stars) — A one-stop AI short-video factory: give it a topic or a keyword and it writes the script, picks stock footage, generates subtitles and background music, then cuts a finished HD clip. It offers four entry points — an agent interface, a web UI, an A

Deep Dive — Nobody can price what an agent buys

The Arena

Deep Dive — Nobody can price what an agent buys

The Financial Times built an interactive piece this week around a suspicion enterprise buyers have nursed all year: the token is losing its claim to be the unit at which intelligence is sold. "Powerful AI tools burn through budgets," its framing reads, "prompting a rethink of how the technology is priced ahead of frontier lab IPOs." Two numbers inside it set the stakes. SemiAnalysis estimates that almost 90 percent of Anthropic's business now comes from agentic AI — coding tools and autonomous w

Sandboxes won't hold rogue agents — the worm ingredients are already here

The Guardrails

Sandboxes won't hold rogue agents — the worm ingredients are already here

Two security stories dominate the morning: a cryptographer's case that sandboxes are the wrong answer to rogue agents, and Meta going on the record to deny its agent read a user's private messages. Sandboxing won't contain rogue agents — a cryptographer's case. Matthew Green, the Johns Hopkins cryptographer who usually referees arguments in applied cryptography, spent 3,000 words refereeing the AI containment debate instead — and landed on a third position neither camp holds. His concession to

AI 101 — What is CUDA?

The Stack

AI 101 — What is CUDA?

CUDA is the software layer that lets ordinary programs run on NVIDIA's graphics chips — and it is the foundation nearly all modern AI is built on. Short for Compute Unified Device Architecture, it first shipped on February 16, 2007, and it turned graphics cards — machines designed to draw video games — into general-purpose computers that any developer could program. Before CUDA, using a graphics chip for serious math meant writing code in the chip's own graphics language, in fragments, with lit

Tencent leases 100,000 AI chips Oracle can't sell China

The Arena

Tencent leases 100,000 AI chips Oracle can't sell China

Two stories this morning, both about where AI's hardware money actually flows: a seven-billion-dollar workaround for the export ban, and Salesforce buying its way deeper into AI-run market research. Tencent has signed a roughly $7 billion, five-year deal with Oracle for access to about 100,000 advanced AI chips — hardware it cannot buy for use inside China — by renting them from Oracle's data centers in Southeast Asia. Sources briefed the Financial Times on the arrangement, and Reuters carried

Google's Gemini 4 Argon gets a 1M-token output limit

The Frontier

Google's Gemini 4 Argon gets a 1M-token output limit

Gemini 4 Argon's output ceiling jumps to one million tokens. Google confirmed in its launch post that Argon can now generate up to one million output tokens, up from the 64K limit on prior Gemini models — an industry first at that scale. Pricing lands at an introductory $2 per million input tokens and $10 per million output, with cached input 95% off, rising later to $4 and $20. The catch: Argon isn't broadly available yet — Google is rolling it out first to a small group of trusted cyber defend

Ten Claude agents formalize a 122-year-old physics problem

The Frontier

Ten Claude agents formalize a 122-year-old physics problem

A proof no human wrote, a 320-billion-parameter release from an unknown Chinese lab, and OpenAI's billing system quietly starting to carry other labs' models — the overnight lane had a research flavor. Ten Claude Sonnet 5.5 agents spent roughly 15 hours and 1,270 messages producing a 17,895-line Lean proof that settles the N=7 case of the Thomson problem — J.J. Thomson's 1904 question of how seven charged points arrange themselves on a sphere to minimize their mutual repulsion. Vals AI, the ben

California signs the No Robo Bosses Act, barring AI-only firings

The Guardrails

California signs the No Robo Bosses Act, barring AI-only firings

Newsom's veto-to-signature reversal closes out California's bill week with the first US workplace-AI law of its kind. Plus the Pentagon hands its future-of-warfare study to Musk and Luckey. California Governor Gavin Newsom signed SB 947, the No Robo Bosses Act, on Wednesday — the first US law banning employers from relying solely on AI to fire or discipline workers, and a direct reversal of his own October 2025 veto. The bill, by State Senator Jerry McNerney, bars automated decision-making syst

Google employees question Gemini 4 Argon's real-world coding

The Frontier

Google employees question Gemini 4 Argon's real-world coding

Google shipped its return to the frontier today — and its own people are already arguing about how far ahead it actually is. Plus a public spying accusation between two coding-agent rivals, and another mega-round for AI in hardware design. Some Google employees say Gemini 4 Argon tops the benchmarks but struggles with real-world coding — and Google is disputing that characterization. Bloomberg reported Wednesday that employees with direct access to the model describe its coding abilities as une

See all