Schmidhuber Slams AI Plagiarism History as OpenAI Accused of Stealing a Math Proof
hardmaru · x · 2026-09-10
Jürgen Schmidhuber weighs in on a viral allegation against OpenAI: a math professor near solving Navier-Stokes reportedly used Codex for a year; OpenAI allegedly combed his N-S-tagged session logs, spun up 10,000 agents to finish the most promising proof, threatened him into silence, then claimed they never saw his work while admitting they "cannot rule out" that de-identified usage data helped their models. Unverified, but explosive. Schmidhuber adds a broader critique: AI history is "permeated by plagiarism" — least squares renamed "neural networks," cybernetics renamed "AI," 1960s techniques renamed "deep learning" — all without proper attribution.
More from Companies & People
- WordPress founder Matt Mullenweg urgently hires security staff, moves assets off Automattic hosting — photomatt · 2026-09-10
- Beff Jezos: 'max fear-mongering stage' as centralised AI fights open-source threat to $2T valuations — beffjezos · 2026-09-10
- Satire: 'Dario' calls AI enslavement just a challenge for long-term ARR growth — thedealdirector · 2026-09-10
- Sam Altman: Local Open AI Must Stay Competitive or Big Labs Will Own Everyone's Cognition — beffjezos · 2026-09-10
- Anthropic issues safety statement amid controversy; investors mock the bland language — Miles_Brundage · 2026-09-10
- Failed AI interview task exposes how candidates navigate AI, not just results — jimmykoppel · 2026-09-10