Fast learning-rate transfer holds at growing training horizons when T grows slower than sqrt(n), but requires nondegenerate first-order loss sensitivity to avoid spectral-dependent failures.

Learning reusable meta-skills for environment design improves AI test-time performance by 8.95 points over no-skill construction, enabling fixed-weight self-improvement.

Frozen weights do not guarantee safe saturation in agentic coding; auditors must monitor scaffold expansion, not just checkpoint freezing.