logoalt Hacker News

jimbokunyesterday at 3:22 PM0 repliesview on HN

How so?

The models get better and better at understanding the intent of a prompt and doing more useful work with less intervention.