Meanwhile current gen models produce routinely produce complex code. Anyone who has worked with Claude/GPT-4o/o1 experiences this first hand every day. The take that their useless because their training set isn't flawless is wild.

3 likes 0 replies