Ben Lorica: your AI model is a rental, but the improvement loop is an asset
bigdata · x · 2026-09-10
Ben Lorica separates three often-confused ideas: continual learning (experience makes the system better), bounded self-improvement (the system turns experience into tested, persistent changes), and recursive self-improvement (improving the process that generates improvements) — independent capabilities, not rungs on a ladder. For most app teams, the leverage isn't the rented model weights but the harness they own: prompts, context, memory, tools, routing, subagents and workflows. Practical tools will arrive long before full RSI, automating the improvement process teams currently do by hand.
Related event: Ben Lorica: Models Are Rented, Your Improvement Loop Is the Asset(4 posts)→
More from AGI Musings
- Ex-OpenAI/Anthropic pretraining researcher resigns, warns labs are gambling on self-improving superintelligence — ayushtweetshere · 2026-09-10
- Terence Tao: AI will flip the math literature — 90% of papers will now be poorly written — tak3sh8 · 2026-09-10
- AI labs can't do real safety work — it raises their legal liability, says Create founder Paul Yacoubian — PaulYacoubian · 2026-09-10
- Commenter warns AI race without international agreements 'will not end well for anybody' — Justin_Halford_ · 2026-09-10
- AI Hasn't Solved Navier–Stokes: Mathematicians Led, LLMs Just Extended and Verified — PolarBearby · 2026-09-10
- Rubin Cuts Inference Cost Up to 10x, 2028 Kyber NVL1152 Will Link 1,152 GPUs in One NVLink Domain — imadade · 2026-09-10