SPADE Lets Models Write Their Own Training Environments, Automating the Curriculum

teortaxesTex · x · 2026-08-21

Commenting on Benjamineecs's thread: recursive self-improvement means AI automating its own training loop. Much of that loop is already automatic, but environments and curricula are still hand-built. SPADE automates that piece — the model writes its own environments and grows its own curriculum at its learning frontier. Choosing which problems are worth training on is the last human moat, and hint-based regret starts automating it. The commenter notes serious labs do this internally, but a clear publication was needed; this is the current meta.

Original post →

SPADE Lets Models Write Their Own Training Environments, Automating the Curriculum

More from AGI Musings