German Courts Rule Google Liable for AI Overviews Lies
Two German courts now hold Google directly liable for what its AI Overviews say — and the latest ruling carries fines of up to €250,000 per violation.
Breaking AI news and announcements
Two German courts now hold Google directly liable for what its AI Overviews say — and the latest ruling carries fines of up to €250,000 per violation.
OpenCode warns stored OpenAI, Anthropic and Google API keys may be exposed — and the first email came 22 hours late. Revoke and reissue your keys now.
Meta pulled the clause letting it use business websites to “train” AI after experts called it “absurdly broad” — but crawling stays. What owners must do.
GitHub Copilot retires three models on Nov 16, 2026 — Claude Opus 5 and GPT-5.6 Terra among them. What it costs your team and how to profit from the migration.
Replit CEO Amjad Masad said at TechCrunch Disrupt the company is ‘no longer working with Anthropic’ — with ChatGPT now the default engine of Replit Agent.
OpenAI just launched ChatGPT Slides: create, co-edit, and present pitch decks inside ChatGPT — and export to PowerPoint — turning a designer-cost line item into minutes of your time.
Meta’s new Commercial Terms take effect October 30, 2026, letting it crawl business sites and use “any content” to train AI — what owners must do now.
NVIDIA’s Dory 1.0 is a 4B open-weight looped model with a thinking dial: pick 1-4 loops at inference to trade compute for accuracy, hitting 61.8 on SWE-bench.
Google’s Chrome 155 release notes credit an Anthropic researcher ‘assisted by Claude’ with a serious DOM flaw — AI is now a bug hunter, and the bounties are real.
A startup raised $55M with no product. Its first model, Glimpse, zooms into images mid-thought. What it means for visual AI money.