# @solarapparition — 2024-10-31

♥4 ↻1 · https://x.com/solarapparition/status/1852032555991568481

still figuring out my confidence level on this one, but preliminarily, o1-preview has been more brittle than i expected once the problem space expands out of its comfort zone. like it's really, ridiculously good at certain things presented in certain ways, but actually gets pretty confused when the inputs are messy. both sonnets handle messiness really well, and i think the "i don't need to babysit" aspect is the biggest factor to me preferring them

tags: author:solarapparition, kind:tweet, model:o1, on:o1, year:2024
cited on: o1