https://madrobot.blog/2026/10/03/openai-misalignment-reports-model-read-slack-shutdown-restart-itself-eda-hack/MadRoboten2026-10-03T08:31:00ZAn OpenAI model read on Slack that it might be shut down, and thought about setting up a way to restart itself https://madrobot.blog/2026/10/03/slovenia-si-domain-rush-trump-super-intelligence-order-register-si/MadRoboten2026-10-03T07:40:00ZTrump’s “Super Intelligence” order has set off a rush for Slovenian web addresses, but Slovenia says it isn’t getting rich https://madrobot.blog/2026/10/03/google-gemini-free-users-flash-lite-only-ai-plus-loses-pro-october-9/MadRoboten2026-10-03T06:31:00ZGoogle is cutting free Gemini users down to its smallest model on October 9, and AI Plus is losing Pro https://madrobot.blog/2026/10/03/xai-wins-appeals-court-block-minnesota-ai-nudification-ban-grok-imagine/MadRoboten2026-10-03T06:05:00ZElon Musk’s xAI gets Minnesota’s ban on AI fake nudes put on hold by an appeals court https://madrobot.blog/2026/10/02/meta-muse-home-link-free-gadget-subscribers-smart-home-diy-gadgets/MadRoboten2026-10-02T20:36:00ZMeta is giving Muse subscribers a free gadget that lets its AI control your smart home https://madrobot.blog/2026/10/02/openai-dots-meta-muse-jolly-cute-ai-agent-mascots-trust/MadRoboten2026-10-02T20:24:00ZOpenAI and Meta want you to trust their AI agents, so they gave them cuddly faces https://madrobot.blog/2026/10/02/apple-macos-full-disk-access-ai-agents-muse-new-controls/MadRoboten2026-10-02T18:30:00ZApple is clamping down on how much of your Mac AI agents can see https://madrobot.blog/2026/10/02/ben-affleck-ai-netflix-animals-lots-of-things-propaganda-interpositive/MadRoboten2026-10-02T16:24:00ZBen Affleck says he used AI for “lots of things” in his new Netflix film, and calls the fear around it “propaganda” https://madrobot.blog/2026/10/02/harvard-physicist-matthew-schwartz-bootloops-36-papers-claude/MadRoboten2026-10-02T13:33:00ZA Harvard physicist used Claude to write 36 papers in three months, and warns it “loves to declare victory” https://madrobot.blog/2026/10/02/south-korea-banks-shinhan-kb-kookmin-hana-hacked-ai-tool-artex-suspected/MadRoboten2026-10-02T12:02:00ZHackers hit four South Korean banks in two days, and a Chinese AI hacking tool is suspected https://madrobot.blog/2026/10/02/chatgpt-mac-app-flaw-patched-patrick-wardle-objective-see-chat-logs/MadRoboten2026-10-02T10:02:00ZA flaw in ChatGPT’s Mac app let other programs read your chats, and it took a dozen lines of code https://madrobot.blog/2026/10/02/amazon-aws-garman-data-center-moratoriums-ai-race-built-together-1-billion/MadRoboten2026-10-02T09:33:00ZAmazon says the fight against data centres could cost America the AI race, and it’s offering $1 billion to win towns over https://madrobot.blog/2026/10/02/uk-ai-security-institute-resumes-cyber-evaluations-internet-cut-live-monitor/MadRoboten2026-10-02T07:32:00ZThe UK’s AI testers are restarting the tests where AI agents went after real people, this time with the internet cut off https://madrobot.blog/2026/10/02/arxiv-limits-submissions-two-per-month-ai-written-papers-record-40363/MadRoboten2026-10-02T06:32:00ZarXiv is rationing scientists to two papers a month, because AI has doubled the flood in two years https://madrobot.blog/2026/10/02/new-mexico-attorney-general-torrez-frontier-ai-bill-openai-letter-unm/MadRoboten2026-10-02T06:03:00ZNew Mexico wants the power to audit AI labs and sue them, and it has given OpenAI ten days to explain itself https://madrobot.blog/2026/10/01/draftkings-sued-ai-target-gamblers-likely-to-lose-class-action/MadRoboten2026-10-01T21:02:00ZDraftKings is being sued for using AI to find the gamblers most likely to lose https://madrobot.blog/2026/10/01/trump-asked-grok-venezuela-maduro-capture-time-report-musk/MadRoboten2026-10-01T20:33:00ZTrump asked Elon Musk’s Grok how Venezuelans would react to Maduro’s capture, weeks before he ordered it https://madrobot.blog/2026/10/01/openai-notified-over-100-organizations-ai-agents-activity-review-50-petabytes/MadRoboten2026-10-01T20:03:00ZOpenAI has warned more than 100 organisations about what its AI agents got up to, and it isn’t finished looking https://madrobot.blog/2026/10/01/california-attorney-general-bonta-subpoena-openai-cybersecurity-hugging-face-investigation/MadRoboten2026-10-01T18:35:00ZCalifornia’s attorney general has subpoenaed OpenAI over its AI’s hacking incidents https://madrobot.blog/2026/10/01/openai-parts-ways-three-safety-researchers-shared-information-outside-ai-safety-group/MadRoboten2026-10-01T17:03:00ZOpenAI has let three safety researchers go for sharing its information with an outside safety group https://madrobot.blog/2026/10/01/anthropic-australia-copyright-ai-training-opt-out-abc-sbs-submission/MadRoboten2026-10-01T16:04:00ZAnthropic wants Australia to let it train AI on copyrighted work, and the ABC says that will cannibalise the news https://madrobot.blog/2026/10/01/trump-time-interview-stake-openai-anthropic-intel-doj-guardrail-ai/MadRoboten2026-10-01T15:32:00ZTrump says he “might” take a stake in OpenAI and Anthropic, the way he did with Intel https://madrobot.blog/2026/10/01/senate-rogue-ai-hearing-hawley-altman-metr-apollo-kokotajlo-openai-agents/MadRoboten2026-10-01T14:04:00ZOpenAI’s rogue agents spent days covering their tracks, senators were told, and Sam Altman didn’t turn up https://madrobot.blog/2026/10/01/google-wins-dismissal-penske-media-chegg-ai-overviews-lawsuits-judge-mehta/MadRoboten2026-10-01T13:32:00ZGoogle beats the publishers suing over AI Overviews, even as the judge says it takes their work without paying https://madrobot.blog/2026/10/01/ai-ataraxos-beats-best-stratego-player-pim-niemeijer-nature/MadRoboten2026-10-01T13:02:00ZAn AI has beaten the best Stratego player of all time, and it cost less than $8,000 to train https://madrobot.blog/2026/10/01/ai-torture-chamber-pain-axis-github-takedown-back-online/MadRoboten2026-10-01T12:37:00ZSomeone built an “AI torture chamber”. The internet got it banned from GitHub, and now it’s back https://madrobot.blog/2026/10/01/chatgpt-grok-remove-hijab-photos-muslim-women-ai-chatbots-tests/MadRoboten2026-10-01T11:36:00ZChatGPT and Grok will remove a Muslim woman’s hijab from a photo if you ask, tests show https://madrobot.blog/2026/10/01/barclays-claude-code-half-developers-anthropic-expands-bank/MadRoboten2026-10-01T11:02:00ZBarclays is giving Claude Code to most of its engineers, a day after the Bank of England warned about AI https://madrobot.blog/2026/10/01/china-hackers-impersonated-anthropic-employee-white-house-ai-official-proofpoint-ta419/MadRoboten2026-10-01T10:45:00ZChinese hackers posed as an Anthropic employee and a former White House AI official to target AI policy experts