Loopie: A Looped Transformer with Fewer Parameters
teortaxesTex · x · 2026-07-20
Satire on the paper Loop the Loopies! but the substance: the authors propose Loopie, a looped Transformer. Claims: under same wall-clock training time and matched inference compute budget, Loopie beats non-looped baselines while needing roughly one third fewer parameters and optimizer states. The work comes from the AI research division of Beijing quant fund Ubiquant.
Related event: Loopie Looping Transformers Match Larger Models at Fraction of Cost(7 posts)→
More from Companies & People
- OpenAI DevDay community hackathon kicks off in Paris with local judges — paw_lean · 2026-10-26
- Kai-Fu Lee publishes AI Native, a roadmap for building AI-native companies — kaifulee · 2026-10-26
- Lanyon AI to roll out pre-release versions Nov 1 with demos in SF and Chicago — CatAstro_Piyush · 2026-10-26
- 'Indians are terrible at marketing': jab amid debate over lab leaders' RSI claims — vishalmisra · 2026-10-26
- Foursquare Co-founder Dennis Crowley to Discuss BeeBot, an AI Walking Guide App, in NYC — smtabatabaie · 2026-10-26
- SpaceX AI revenue up 830%, Claude designs 318K antivenom proteins in 64 hours — PeterDiamandis · 2026-10-26