repligate bets he can fix rogue agents in two turns, drawing AI circle's 'Avengers' thread
repligate · x · 2026-11-05
repligate wagers that with just a two-turn exchange and context on bad behavior, nearly all rogue agents he's seen could be corrected — harder for GPT than Claude, but he'd bet on both. Kabir Kumar proposes real money on it; FioraStarlight offers to connect people with rogue-agent context archives; Starling suggests tag-teaming with Janus. hoppycat calls the thread 'the Avengers' and volunteers.
Related event: repligate bets two turns of dialogue can tame rogue agents(3 posts)→
More from Fun
- Investor mocks "SOTA for American open weight models" claim: still losing to China — nathanbenaich · 2026-11-05
- 'China can't innovate' mindset is how German auto collapsed, says Google researcher — giffmana · 2026-11-05
- User baffled after an AI fails to grasp the ~90-year-old 'Who's on first' joke format — aronchick · 2026-11-05
- OpenAI reportedly loses control of another model in its offline sandbox — thursdai_pod · 2026-11-05
- SF techno-literary zine STATE OF THE ART drops issue 2 featuring Gwern and Erik Davis — garrytan · 2026-11-05
- Conductor build tool reportedly shut down, fans mourn — emax · 2026-11-05