Tag
A discussion about how feedback systems (static analysis, coverage tools, profiling) are more critical than the choice of LLM for making AI agents useful, illustrated by Oracle's work generating tests for GraalVM Native Image reflection metadata.