synthetic data is fine if you can ground the model somehow. that's why the o1/o3... | Hacker News

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

		nialv7 on Dec 22, 2024 \| parent \| context \| favorite \| on: GPT-5 is behind schedule synthetic data is fine if you can ground the model somehow. that's why the o1/o3's improvements are mostly in reasoning, maths, etc., because you can easily tell if the data is wrong or not.

dartos on Dec 23, 2024 [–]

That makes a lot of sense.

Binary success criteria has very little room for bias.

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