Open weights are turning enterprise AI into specialized intelligence companies can own
bigdata · x · 2026-07-29
Ben Lorica argues that the more important shift in enterprise AI is not just cheaper open weights, but the ability to turn them into specialized intelligence a company actually owns.
- Open models are narrowing the gap with frontier systems, making them harder to ignore for competition, security, and sovereignty.
- The post-training stack is filling in: supervised fine-tuning, reinforcement fine-tuning, environments, graders, verifiers, data tooling, evaluation, deployment, and monitoring.
- More than 25 startups are now building around reinforcement fine-tuning, but the product is usually the full workflow around it, not just the algorithm.
- Teams can increasingly describe a task, define a baseline, identify failure modes, build evals and test data, retrain, and iterate in a semi-automated loop.
More from Companies & People
- Reddit debate: Are AI agent startups making real revenue or just burning VC money? — Pure_Armadillo4339 · 2026-10-16
- Paul Graham on investor rejections: dumb reasons, dislike — or both — dbasch · 2026-10-16
- Big Tech Is Killing Agentic AI — and Enterprises Are Letting It Happen — DavidLinthicum · 2026-10-16
- Pivotal opens Q1 2027 AI Safety Research Fellowship: 15 weeks in London, £15-18k stipend — peterbhase · 2026-10-16
- Iason Gabriel named Principal Scientist and Director of AGI & Society at DeepMind Institute — IasonGabriel · 2026-10-16
- LangChain Guide: How Madrigal, Abridge and Vizient Scale Agents in Healthcare — LangChain · 2026-10-16