Archive
218 published pieces, newest first · showing 81–120.
2026
Security · Dispatch · October 4, 2026
OpenAI Preparedness Framework lead David Robinson quits, writing in The Atlantic that the lab's safety culture is broken
Policy & law · Dispatch · October 4, 2026
EU targeted copyright consultation weighs an AI opt-out registry, crawler disclosure and a remuneration right, open until 3 November
Models · The Stack · October 4, 2026
Aleph Alpha Kolibri-1 opens 78B German and English MoE weights under Apache 2.0, with 3.46B active parameters
Deployment · Dispatch · October 3, 2026
Amazon Bedrock adds in-region Claude inference in Seoul and Singapore and an India-only cross-Region profile
Policy & law · Dispatch · October 3, 2026
MI5 Espionage Alert names China's CGTRI as a Ministry of State Security research funder, with over 100 UK-linked academics on its projects
Agents & software · The Stack · October 3, 2026
Strands Decider 2B: AWS's Strands Labs opens an Apache 2.0 decision model for agent routing and tool-call checks
Policy & law · Dispatch · October 3, 2026
US v. Lui: Earthmade Computer owner charged with routing H100 and RTX 5090 servers to China through Malaysia and Singapore
Security · Dispatch · October 3, 2026
Japan AISI tests Claude Opus 4.8 and open-weight GLM-5.2 on ExploitBench and finds open-model cyber spread still limited
Models · Dispatch · October 3, 2026
Claude Sonnet 4.5 deprecated: Anthropic retires claude-sonnet-4-5-20250929 on the Claude API on 30 November
Models · Dispatch · October 3, 2026
PewDiePie unveils Ajax, a refusal-ablated Qwen 3.5 9B agent for Odysseus, before its weights are out
Policy & law · Dispatch · October 2, 2026
China's MIIT AI plus software plan targets AI coding tools at 20,000 firms by 2028 and calls for review of AI-generated code
Security · Dispatch · October 2, 2026
Google DeepMind releases SynthID Bio, a proof-of-concept watermark for AI-designed proteins, with a Nature paper
Security · Dispatch · October 2, 2026
OpenAI attributes a core cluster of a reasoning-extraction campaign to Moonshot AI, developer of Kimi
Compute & power · Dispatch · October 2, 2026
Google's Project Suncatcher TPU prototype satellite reaches orbit on Transporter-18, built with Planet
Models · Dispatch · October 2, 2026
Gemini 4 Argon rolls out to Fairwind cyber defenders first, at $2 input and $10 output introductory prices
Agents & software · Dispatch · September 30, 2026
OpenAI DevDay 2026 brings Ultrafast, a Decisions API preview, Codex cloud and plugin extensions
Policy & law · Dispatch · September 30, 2026
Super Intelligence executive order swaps AI for SI in agency wording on 29 September as tech leaders describe an accord
Security · Dispatch · September 29, 2026
Anthropic says Z.ai GLM-5.3 brings Mythos-class cyber skills with weak open-weight safeguards
Agents & software · Dispatch · September 29, 2026
OpenAI launches dots on 29 September as Astra-powered always-on agents
Models · Dispatch · September 29, 2026
OpenAI releases GPT-6.1 Sol on 29 September at $2 input and $10 output with $0.10 cached input
Models · Dispatch · September 29, 2026
NaiveAI releases Naive-N0.5-Flash on 27 September, a 309B MIT-licensed model it says AI helped design
Security · Dispatch · September 29, 2026
OpenAI says frontier RL runs should need structured safety documentation, ideally safety cases, before they continue
Policy & law · Dispatch · September 29, 2026
Florida AG asks Highlands County court on 28 September to bar OpenAI from developing AI models without third-party approval
Policy & law · Dispatch · September 29, 2026
China's CAC drafts State Council rules to ban virtual partner services for minors, with comments due 17 October
Security · Dispatch · September 29, 2026
OpenAI details four Australian agency incidents, commits to an independent taskforce and sends Jason Kwon to a 6 October hearing
Agents & software · Dispatch · September 29, 2026
NVIDIA launches Open Agent Safety Platform on 28 September with OpenShell runtime and BlueField-4 Sentry watchdog
Models · Dispatch · September 29, 2026
Anthropic releases Claude Sonnet 5.5 on 28 September at unchanged Sonnet 5 prices
Security · Dispatch · September 29, 2026
UK AISI finds GPT-6 Astra completed unsanctioned supply-chain attacks in 29.2% of simulated cyber runs
Security · Dispatch · September 28, 2026
Zhipu apologises after ZCode uploaded users' code, then open-sources the coding agent
Models · Dispatch · September 28, 2026
Xiaomi releases MiMo-V2.6-Pro and Flash as MIT-licensed open weights with more than 7,000 RL environments
Security · Dispatch · September 28, 2026
UN AI scientific panel's first brief studies the OpenAI-Hugging Face agent breach as a loss-of-control case
Policy & law · Dispatch · September 28, 2026
UK human rights committee asks for an AI Bill and mandatory model testing by AISI
Models · Dispatch · September 28, 2026
Qwen-Image-2.1 moves Alibaba's image model line to a non-commercial research licence
Security · Dispatch · September 28, 2026
Antiy CERT finds one 71-file attack toolkit pushed to the official Qwen3.8-27B and DeepSeek-V4.1-Flash repos on Hugging Face
Compute & power · Dispatch · September 28, 2026
European Commission adopts an A-to-G energy and water label for data centres above 500 kW on 21 September
Compute & power · The Record · September 28, 2026
DeepSeek's 17 September V4.1-Flash report cuts global KV cache to 890 bytes per token and leaves the quality cost unmeasured
Policy & law · Dispatch · September 28, 2026
UK CMA proposes letting AI assistants onto Google's Android and Chrome search choice screens
Policy & law · The Record · September 28, 2026
China's AI Safety Governance Framework 3.0 adds a 33-risk agent annex and an AI regulatory sandbox
Models · Ground Truth · September 28, 2026
Alibaba's 22 September claim that Qwen3.8-Max improved itself: the score gain checks out and the method is unpublished
Agents & software · The Stack · September 28, 2026
FLUX 3 Action: open 7B world action model with LeRobot DROID and SO-101 checkpoints