The Human-in-the-Loop is Tired
pydantic.dev · 2,036 words · saved by 1 readers
On reward functions, dopamine, and what it actually feels like when the code starts writing itself
Yet another thought piece about LLMs. I know. Bear with me. This is an attempt to put words around something I think most developers are experiencing right now but haven't had time to make sense of. Programming with LLMs is genuinely useful and genuinely destabilizing. These two things coexist. If we pretend the second one isn't happening, we will all burn out. At Pydantic, we build tools that developers use to validate data, build AI agents, and observe what their systems are doing in production. We are, quite literally, in the business of making LLM-powered software more reliable. And we…
saved by
related reading
- Malleable software in the age of LLMsgeoffreylitt.com
- The vibe coder's career path is doomedblog.florianherrengt.com
- Don't fall into the anti-AI hype - <antirez>antirez.com
- 2025 LLM Year in Review – karpathykarpathy.bearblog.dev
- We mourn our craft | Read the Tea Leavesnolanlawson.com
- crawshaw - 2025-01-06crawshaw.io
- Code is cheap. Show me the talk. - nadh.innadh.in
- Developing our position on AI - Blog - Recurse Centerrecurse.com
- 2x, not 10x: coding with LLMs in 2026obryant.dev
- Here’s how I use LLMs to help me write codesimonwillison.net
- the art of programming and why i won't use llmkennethnym.com
- Your job is to deliver code you have proven to worksimonwillison.net