my hot take is that evals are the only part of agent engineering that require real thinking

Harrison Chase argues that evaluations (evals) are the most intellectually demanding aspect of agent engineering, distinguishing them from other engineering tasks that may be more routine or formulaic. The post suggests that designing meaningful benchmarks and assessment criteria for agents requires deeper conceptual work compared to other development phases. This reflects a perspective that evaluation methodology is critical to advancing agent capabilities and reliability.

Source: Harrison Chase (X) | 2026-07-07