Martin Fowler's May 5 fragments cover several AI-assisted development topics: a new open-source framework (Lattice) that operationalizes patterns for reducing friction in AI coding via composable skills and a living context layer; an update to the Structured-Prompt-Driven Development article with a new FAQ; Jessica Kerr's reflection on double feedback loops when building with AI agents; a defamation lawsuit against Google over false AI-generated content about musician Ashley MacIsaac; analysis of massive AI infrastructure spending by big tech (over $100B, 50-75%+ of revenues); the case for local models being 'good enough' for daily coding work; Apple's potential strategy of betting on local/on-device AI rather than cloud; and Kent Beck's 'Genie Tarpit' argument that AI tools tend to produce low internal quality code, raising the open question of whether code quality still matters in the agentic era.
Sort: