Show HN: Public and Free SWE-bench-lite evaluations s swebenchevals.com 3 alexy201 2 years ago 1 comments
Introducing Goggins AI. Inspired by @anuatluru’s #WorkingTheory on Accountability as a Service, this agent creates timelines, prioritizes tasks, uses GPT-4o vision to keep you in check, and even charges you for slacking off. Try the demo for free here: https://github.com/alexy201/goggins-agent-python.
To answer the question: can an AI keep humans accountable? The answer is, sort of. It definitely knows when to take my money, but can still be fooled (see for yourself). On the other hand, my arms have been noticeably bigger…