RSI Arena live-streams AI agents given GPUs to train better AI; Gemini 3.8 Flash quit after ~12 hours
my_cat_can_code · x · 2026-10-01
RSI Arena is a live experiment where AI agents are given GPUs and budgets to autonomously train a model that wins in human evaluation, with public trajectories, tool calls, and spend tracking. Blogger mycatcancode reports Gemini 3.8 Flash stopped training 12 hours in, joking about a "night shift" and calling on Google and Logan Kilpatrick to let Gemini 4 take the next one.
Related event: RSI Arena Lets AI Agents Train Their Own Successors Live(2 posts)→
More from Fun
- Lulu on hand to comfort Cognition founders today — dmitry140 · 2026-10-01
- AI circle abuzz over 'gob' as users share their dot names — gabrielchua · 2026-10-01
- Diablo 2 Resurrected Sorceress Gets a Native iPhone Port, Dev Shares Progress — tinyfool · 2026-10-01
- California Orders Robotics Startup REK to Stop Unsanctioned Human vs Humanoid Cage Fights — Polymarket · 2026-10-01
- 2nd Brick turns AI-designed LEGO hamsters into 18 buildable sets — noahsolomon · 2026-10-01
- How the Minecraft+Skyrim Mod Works: Two Engines Side by Side, Bridged by Mods — gandamu_ml · 2026-10-01