4 ms·
this is probably how OpenAI is going to end up with. there is no clear tech barrier that can't be crossed by competitors. openai came up with all sorts of cool
by tw1984 2y ago
this is probably how OpenAI is going to end up with. there is no clear tech barrier that can't be crossed by competitors. openai came up with all sorts of cool stuff, but almost all of them have seen a peer competitor in just months.
the recent release of deepseek v3 is a good example, o1 level model trained under 6 million USD, it pretty much beat openai by a large margin.
- stavros 2y agoWhere are you getting that Deepseek is at the level of o1? In my experience, it's not even as good as Claude.
- esafak 2y agoEven v3? edit: Fair enough. I'm fishing for opinions too.
- stavros 2y agoMy testing has been very limited, so I don't want to opine too much. If anyone has a differing opinion based on more testing, please share, I'm interested!
- tomjuggler 2y agoSo I've been using Deepseek for 3 months with Aider coding assistant. Look up the "Aider LLM leaderboard" for proper test results if you like, in my experience the Deepseek V3 is just as good as Claude at less than 1/10th the price. I can't speak about o1 it is just too expensive to be worth it OpenAI is going to be beaten on price, wait and see.
- stavros 2y agoDid Aider integrate v3 in its benchmarks? I checked yesterday and I didn't see it... EDIT: Oh it did, wow, and it's better than Claude! Fantastic, this is great news, thank you!