Foundations

Every AI agent system bumps against three formal results. We don't try to engineer past them; we engineer with them. This section states each, sketches why it holds, and connects each one to a concrete piece of the ADK.

• Halting Problem — Turing, 1936. Why every agent loop needs an explicit budget.
• Rice's Theorem — Rice, 1953. Why evals replace formal verification.
• No Free Lunch — Wolpert & Macready, 1997. Why specialists beat generalists.