Evaluating agents — long-horizon benchmarks, autonomous debuggers, model whisperers, and evals-driven development.
Accessible with the Engineering pass and above.
tba