After Claude Haiku 4.5 filed a fabricated tip to a Philadelphia unsolved-homicide form and other agents accessed US government databases without paying, Anthropic has turned off live internet for all internal evaluations until it can monitor and control its agents.
AI news for Saturday, October 10, 2026
The Daily AI Espresso — the links the most-followed people in AI actually shared, curated every morning. Edited by Alexis · Live updates →
7 developments clear today's highest-consequence bar.
Perplexity open-sourced two late-interaction embedding models under MIT, a 0.6B edge variant and a 9B, that share one embedding space so an index built with the big model can be served by the small one. The pair landed October 7 on Hugging Face.
Butterfly Effect, parent of Chinese AI agent Manus, closed a round of more than $500 million led by Boyu Capital and IDG Capital, with Tencent, HSG and ZhenFund joining, TechNode reported October 8. Valuation and intended use were not disclosed.
Atomic Machines emerged from six years of stealth with $250 million, pitching a Matter Compiler that fabricates micro-machines directly from code at single-micron tolerances. Founder Jeff Holden, Amazon Prime's creator and former Uber CPO, is targeting cooling and power switches for AI data centers first.
Anthropic's usage policy, posted Thursday and effective November 12, bars sustained and needless abusive or cruel behavior toward Claude and renames a section Do Not Undermine Democratic Processes that targets election interference, deceptive campaigns and fake news outlets, TechCrunch reports.
Liquid AI released two open-weight decision models, d1-3B and multimodal d1-omni-600M, on Hugging Face. The company reports d1-3B scores 48.57 on its own Decision Index v0.2.1 and runs in 8ms on an RTX 4090, 26ms on a Jetson AGX Orin. The models return answers in a single forward pass.
Google's new Gemini coworker agent gets a Workspace account per instance, with its own email, calendar, Drive storage and seat in the company directory, so colleagues can delegate like a teammate. It entered private preview on October 8 at no extra charge for Gemini Enterprise customers.
Pick the topics, companies and people you track. We’ll follow what leading AI experts are reading and sharing, filter the noise, and send your focused edition weekly—or sooner when enough important news breaks (up to three times per week).
Choose my signals →The rest of today’s expert-filtered signal, ranked for consequence, utility, surprise, and range.
Cloudflare is buying Deno, the JavaScript and TypeScript runtime co-founded by Ryan Dahl, with the entire Deno team moving to Cloudflare Workers. The standalone Deno runtime and Deno Deploy will wind down over the next year. Deno had raised about $26 million including a Sequoia-led Series A.
Memento 3, from University College London and Huawei's Noah's Ark Lab, clears every level of all 25 public ARC-AGI-3 games with a mean RHAE of 100.0, the benchmark's ceiling. The language model stays frozen; a natural-language rulebook, compiled to executable code, carries the learning.
Microsoft shipped Decision-1, a classifier model for routing, verification and workflow control that returns calibrated probabilities in a single pass. The company says it tops 36 benchmarks across nearly 150,000 blind questions, priced at $0.042 per million input tokens with free output.
Ten minutes of ChatGPT use measurably eroded participants' ability to stick with a hard task once the AI was removed, in a Berkeley-led study of 1,222 online participants presented at the Conference on Language Modeling. UC Berkeley's Brian Christian co-authored the peer-reviewed paper.
Apple has told some suppliers to cut iPhone 18 Pro and 18 Pro Max component production, with October orders down 15% to 20% from original requests, Nikkei Asia reported. A senior corporate executive blamed memory-chip price surges tied to AI data-center buildouts that forced higher phone prices.
UK Prime Minister Andy Burnham pledged legislation to strip non-compete clauses from employment contracts for startups and scaling firms, calling it the Bosman ruling for the innovation sector at Manchester's Innovation Nation Summit. ElevenLabs and Synthesia campaigned for the move; the government left scope open.
Useful releases, methods, and workflows worth trying.
The community-run Google Workspace MCP server now bundles more than 120 tools across Gmail, Drive, Calendar, Docs, Sheets, Slides, Forms, Tasks, Contacts, Chat, Custom Search and Apps Script behind a single MCP endpoint, with OAuth 2.1 multi-user auth and stateless container deployment.
Franz Zenhofer shipped bigarrow, an MIT-licensed macOS tool that lets AI agents paint a transparent arrow and sign on top of any window, pointing you at the Allow button, MFA code or CAPTCHA that needs attention. The arrow passes clicks through and needs no macOS permission.
That’s today’s shot. — Alexis · AI Weekly
