You know they are 1000% training these models to draw pelicans, this hasn't been... | Hacker News

Hacker Timesnew | past | comments | ask | show | jobs | submit

dakolli 69 days ago | parent | context | favorite | on: GPT-5.5

You know they are 1000% training these models to draw pelicans, this hasn't been a valid benchmark for 6 months +

simonw 69 days ago | [–]

OpenAI must be very bad at training models to draw pelicans (and bicycles) then.

Legend2440 69 days ago | [–]

Skeptism is out of control these days, any time an LLM does something cool it must have been cheating.

dakolli 68 days ago | [–]

they legitimately suck at everything they don't have concrete examples to copy from.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact