Boltzbit paper claims BAST lets LLMs learn up to 1000x faster than SOTA training
HeyToha · x · 2026-09-25
Boltzbit released a preview paper, Infinite-Parameter LLMs — Generating and Adapting Weights from Live Data, introducing Bayesian Self-learning Transformers (BAST).
- Key distinction: most "learning" agents just save conversations, retrieve chunks, and feed an ever-growing history back into a static model; BAST instead uses live interaction data to generate and adapt model weights, without full retraining cycles.
- Results: BAST learned up to 1,000x faster than the SOTA training algorithms used as baselines.
- Motivation: pretraining text supply is projected to run out in 2-5 years, hitting the ceiling of the bigger-model-more-data paradigm; the authors argue for energy-efficient, dynamic-weight AI.
- Long-term vision: an agent that becomes better through use rather than just carrying a larger archive.
Related event: Boltzbit Claims BAST Makes LLM Learning Up to 1000x Faster(3 posts)→
More from AGI Musings
- Kai-Fu Lee's Fortune Interview Debunks 3 Misconceptions About China's AI — kaifulee · 2026-11-11
- "Deep Learning is a search algorithm, not a learning algorithm" — JosephJacks_ · 2026-11-11
- AI-assisted proof shows no voting rule beats Borda on strategyproofness — conitzer · 2026-11-11
- The new PhD playbook: build a benchmark, earn frontier-lab recognition, then raise money for a data startup — teortaxesTex · 2026-11-11
- Ethan Mollick: GTA 6 Will Be the Last AAA Game Built Without AI — emollick · 2026-11-11
- Air Street Capital founder: AI is 'a museum of the mind' preserving human know-how — nathanbenaich · 2026-11-11