MIT launches course on self-evolving AI spanning context, harness, model and task evolution
pliang279 · x · 2026-10-23
MIT Media Lab's Paul Liang and Ao Qu are teaching a new course, 'How to AI an AI for (Almost) Anything', on self-evolving AI systems that recursively improve from experience, feedback, memory, and exploration over long interaction horizons. Core thesis: intelligence shouldn't stop improving when training ends.
The course covers four levels of evolution:
- Context evolution: prompt optimization and context management
- Harness evolution: memory, tools & skills, workflow optimization, multi-agent systems
- Model evolution: RL, distillation, test-time training, self-play, continual learning
- Task evolution: generating increasingly complex synthetic tasks, simulations, evaluations, and environments
Slides, readings, and videos are available; compute and API credits are sponsored by Kimi AI and Prime Intellect.
More from AGI Musings
- AI safety researcher: labs are at 'slapdash empirics' level, three tiers below aerospace-grade rigor — geoffreyirving · 2026-10-23
- Market dilution, fair use, and generative AI: a longread on copyright's core question — luisdans · 2026-10-23
- Dev: AI turned knowledge workers into 'meat proxies' and the promised fix is welfare forever — zetalyrae · 2026-10-23
- A 9-minute YouTube 'cover' of his 87-minute LLM talk vastly outperforms the original — rao2z · 2026-10-23
- Self-improving companies: why founders should build AI-native banks and insurers from scratch — andrewchen · 2026-10-23
- a16z launches Frontier fund: Andreessen, Horowitz and Dixon on betting on AI's edges — a16z Podcast · 2026-10-23