Architecting for AI: interfaces and docs

Measuring the impact of AI on software engineering – with Laura Tacho

Topic 7 of 14

Published Jul 23, 2025 · Clip 35:52–42:20 (6:28)

Laura describes two trends in preparing codebases for AI: recommitting to clean, well-defined interfaces between services (an AWS 'everything is an API' model), which benefits agentic tools and humans alike, and writing AI-first documentation—code examples, no visual dependencies—because developers now implement tools via AI assistants in the IDE rather than docs pages or tutorials. She cites Vercel and Clerk as strong AI-first documentation examples creating a flywheel. The host notes this helps onboarding and API vendor adoption, and they agree the AI-agent/human interests form a beneficial Venn diagram that also fixes long-standing documentation neglect.