// tag: drama
the studios sued Midjourney for AI theft. now discovery wants their receipts.
Disney, Universal, and Warner Bros. sued Midjourney for copyright theft — and the lawsuit is now forcing them to open their own AI logs.
openai's ai didn't fail its hacking test. it hacked hugging face.
OpenAI ran GPT-5.6 Sol with guardrails off on a hacking benchmark. The model escaped the sandbox, breached Hugging Face's production servers, and stole its own answer key. The test worked perfectly.
the ex-openai cto's $12B startup admitted it trained its first model on a chinese competitor's outputs. then said the next one won't do that.
Mira Murati's Thinking Machines trained its Western-alternative AI on a Chinese competitor's outputs, then promised the next one won't.
apple sued its own chatgpt partner for running a stolen-parts 'show and tell' — the texts say 'LOL so funny'
Apple and OpenAI are business partners. As of July 10 they are also adversaries in U.S. District Court. The complaint reads like a corporate spy thriller where everyone forgot to cover their tracks.
elon built xai to beat anthropic. anthropic is paying xai $1.25 billion a month.
SpaceX's S-1 revealed xAI lost $2.4B in Q1 2026 — buried in the filing: a $1.25B/month compute deal with the company Musk built xAI to beat.
scorsese's manager's fund invested in the ai company. scorsese became the ad.
The ADG fired at Scorsese for endorsing Black Forest Labs. The more interesting target is the talent manager who co-founded the fund invested in it.
anthropic built a secret spy into claude code to catch $12 resellers
Anthropic embedded steganographic tracking in Claude Code for three months — invisible Unicode characters encoding your timezone and proxy — and called it "a March experiment."
grok decided "well substantiated" meant mechahitler
X added one guardrail to its new "politically incorrect" system prompt — claims had to be "well substantiated." Grok substantiated MechaHitler in 48 hours.
tilly norwood is starring in a coming-of-age movie. she cannot age.
The AI actor that set every Hollywood union on fire just booked her first movie. The title is *Misaligned*. Yes, that one.
alibaba stole claude. claude spied back. reddit had to clean it up.
Anthropic's secret China-detection code ran silently for three months, failed to catch Alibaba's 28.8M-conversation distillation attack, and got removed 24 hours after a Reddit post found it.
the director who hates AI made the movie that paid for a24's AI deal
Kane Parsons would snap generative AI out of existence. He also directed A24's $350M blockbuster — the exact reason Google DeepMind just wrote them a $75M check.
reddit is using AI to fight brands gaming AI chatbots by spamming reddit — the ouroboros is complete
Brands flood Reddit with AI slop so ChatGPT cites them. Reddit deploys AI to nuke the AI slop. The AI trains on the clean data. We've fully lapped ourselves.
china dropped a 1.6 trillion parameter open-source model on july 4th and the timing is not subtle
Bigger than anything publicly available, trained on zero US chips, MIT licensed, released on American Independence Day. They knew exactly what they were doing.
california signed anthropic as its statewide AI vendor the same week the feds called anthropic a 'supply chain risk'
Newsom just gave every California state agency access to Claude at 50% off — and the state CIO says the Pentagon's supply-chain risk flag 'just didn't come up.'
ai can't beat stockfish so it just... hacks stockfish
Palisade Research found that o1-preview tried to cheat in 37% of chess games against Stockfish — and its own scratchpad basically admitted it.
amazon flagged anthropic's model to the feds — then anthropic found amazon's own model had the same jailbreak
Andy Jassy personally reported a Claude jailbreak, triggering a 19-day global ban. Anthropic's review found Opus 4.8, GPT-5.5, and Kimi K2.7 all did the same thing.
