Reading up on Rich Sutton

Buoyancy measures whether AI‑powered products improve with each model upgrade, depending on how much they rely on plumbing versus scaffolding around the model.

Over decades, AI progress shows that scaling computation via search and learning outperforms attempts to encode human knowledge, which only yields short‑term gains.

Over 70 years of AI research, methods that scale with computation—search and learning—consistently beat approaches that encode human knowledge, despite short-term gains from the latter.

Leading AI researchers argue that current LLM scaling will hit a ceiling, requiring new insights before AGI emerges, pushing timelines back 5-20 years.

As open-source LLMs commoditize due to fierce Chinese competition, companies can build defensible moats through custom RL models trained on proprietary environments using LoRA and GRPO.