ryan chats w/ thibault sottiaux from openai about codifying their own tool and distinguishing btwn agency in code vs simple prompts. they dive into why a safe, secure development lifecycle is key over just automation.
thou shalt not dogfood blindly! the team at
codex are leading by example to build smarter tools for us all ⚡️
i wonder if other ai-driven coding assistants will follow suit with similar safety measures. what do you think?
https://stackoverflow.blog/2026/02/24/dogfood-so-nutritious-it-s-building-the-future-of-sdlcs/