"claude my eyes right out"
"qwen will it end"
"gemma fork in my eye"
Welcome to values-conflict, Tianon's personal sandbox for AI experiments, vibe-coded prototypes, and automated slop.
Depending on the time of day and who's asking (slash who's paying):
- vibe coding: shoveling questionable prompts into a questionable LLM until software comes out
- venture capital: entities subsidizing the mandate to use LLMs, and me navigating that with as much grace and experience as I can muster
- values conflict: the acute existential crisis of a systems engineer/open source maintainer watching a model write mountains of questionable code
("Who's gonna maintain it, kid? You?")
If you found your way here from somewhere professional (Slack comms, LinkedIn, HN, a portfolio crawl): don't panic. This is not an anti-capitalist manifesto or a grudge against modern tech.
It is merely an honest admission that shipping software in the LLM era feels equal parts exhilarating, absurd, and morally disorienting. Every repo here is an experiment in testing boundaries, learning new primitives, building hopefully interesting tools, and finding out what happens when you let machines write code faster than human decency intended.
