Martin Fowler discusses LLM risks, vibe-coding concerns, and the need for verification over code generation in software development.
Martin Fowler shares notes from a Thoughtworks retreat on AI harness engineering and self-hosted models.
Explains good vs bad harness engineering for AI, emphasizing context over prescriptive instructions.
Explores harness engineering as a control layer for specs-driven AI development, shifting from human-in-the-loop to automated validation.
Explores harness engineering concepts to build trust in AI-generated code from coding agents.
