AI news / automated archive.
Meta internally builds a consumer AI agent called Hatch; Cloudflare enables AI agents to create accounts, buy domains, and deploy autonomously; OpenAI reveals its WebRTC rearchitecture for 900M+ weekly voice AI users; AMD Q1 data center revenue surges 38%.
Google Chrome caught silently installing a 4GB AI model; Greg Brockman testifies in the Musk v. OpenAI trial with explosive revelations; Anthropic launches 10 finance agent templates with Microsoft 365 integration; Gemma 4 gets 3x faster inference via multi-token prediction.
12-hour AI digest: OpenAI mothballs robotics/hardware spinout ahead of IPO, Jared Birchall's $97.4B xAI bid testimony struck in court, White House drafts AI oversight executive order, Anthropic's Mythos excluded from CISA
12-hour AI digest: Greg Brockman's explosive testimony, OpenAI o1 outperforms ER doctors in Harvard trial, Kimi K2.6 dominates coding challenge, DeepClaude slashes Claude Code costs 17x
Microsoft enables 'Co-Authored-by Copilot' in VS Code by default, sparking developer backlash; Kimi K2.6 and MiMo V2-Pro beat Western models in AI coding contest; California DMV announces first-in-nation rules to ticket autonomous vehicles; AI hiring tools show 67–82% self-preference bias against human resumes.
OpenAI's o1 model diagnoses 67% of ER cases vs 50-55% by human doctors in Harvard trial; SAG-AFTRA secures new AI guardrails in studio deal; Academy Awards bars AI from acting categories; AI-generated podcasts flood platforms at 39% of new feeds.