George Hotz argues that AI has upended traditional hiring signals: automated resume spammers and applicants pasting bounty prompts into Claude Code have broken a bounty-based screening system that once worked. He accepts AI as a useful tool for producing low-effort “slop” apps but rejects AI-generated work as a hiring signal for meaningful software development. Hiring at tinygrad now requires public, sustained contribution - participation on Discord and GitHub, listening to weekly open meetings, and demonstrable understanding of project goals rather than one-off feature submissions.
Tinygrad seeks contributors who prioritize clarity, deletion, deep bug fixes, and regression tests over new surface features, and who are committed to a long-term strategy to undercut big-tech rent-seeking by commoditizing petaflops at the infrastructure layer on a roughly 10-year horizon. Logistics: remote/in-person hybrid with San Diego and Hong Kong offices, one-hour weekly meeting, a two-week sprint summary of three sentences, and compensation of $75k-$150k plus 0.1%-0.5% equity for full-time roles. Almost all work is open source and public; the application is simple - start contributing and demonstrate you want to build this.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.