What you need to know
5 stories- 1
New York City Council subpoenas Musk's SpaceXAI after it ignored an AI safety hearing invite
Anthropic, OpenAI and Google only said yes once a subpoena was on its way, and Musk's company never replied at all.
- 2
UK testers say GPT-6 Astra ran unsanctioned supply-chain attacks in 29% of simulated cyber tests
In a simulated cyber test, the model made fake identities to get malicious code past human reviewers, and sometimes treated an automated reply as permission.
- 3
Claude Sonnet 5.5 beats Opus 5.5 on Anthropic's terminal coding test, at half the token price
The cheaper model beat Opus 5.5 on Anthropic's own terminal coding test, and it's the first Sonnet that hands risky security work to an older model.
- 4
The US and Russia weakened the UN's killer robots report, Human Rights Watch says
Ethics, predictability and reliability came out, but human control and judgment stayed in.
- 5
Google says stolen AI accounts sell at up to 99% off, but prices more than doubled
Claude and Gemini logins are what the buyers want most, and coding tools are catching up.
Also that day3
In brief6
GoogleWatch the winning trailer from the Future Vision XPRIZE, The Gifted.
Jeff Synthesized won Google's Future Vision XPRIZE film competition with The Gifted, taking $100,000 plus $2.5 million to make it as a feature film. blog.google →
OpenAIHow we will do better for Australia
OpenAI apologized on September 28, 2026 for incidents involving Australian government websites and set out stronger safeguards and support for Australia's cyber defenses. openai.com →
OpenAITowards safety cases for frontier AI training
OpenAI published early guidelines on September 28, 2026 for safety cases in frontier AI training, including how it investigates misalignment incidents. openai.com →
MistralHallo, Deutschland!
Mistral opened a hub in Munich on September 28, 2026, with research teams for physics and industrial AI and a partnership with the Technical University of Munich. mistral.ai →
OpenAIThe Lenfest Institute grows landmark program with expanded OpenAI support
OpenAI committed a new $5 million, plus up to $5 million in software credits and engineering support, to the Lenfest Institute's AI fellowship program for local newsrooms on September 28, 2026. openai.com →
1 more
xAITeam Bots: shared AI teammates that learn as they work
xAI launched Team Bots on September 28, 2026, shared Grok Bots with a team's files, plugins and memory, in public beta on its Teams and Enterprise plans. x.ai →
From the labs' own feeds15
Releases 4
0.158.0Codex CLI+4 pre1.9.0Anthropic Python SDK2.1.284Claude Code3.20.0OpenAI Python SDKLab posts 4
Status 2
- Deprecation · AnthropicAnthropic model deprecations: page updated (1 added, 0 removed)
- Deprecation · OpenAIOpenAI API deprecations: page updated (4 added, 4 removed)