it's free in Cursor till Sept 2. My experience is subpar so far

[flagged]

To the people downvoting this comment -- it isn't just that Musk made a couple of very sharp nazi salutes. You may say, oh, that was just an unfortunate similarity, he wasn't doing a nazi salute at all. But he has a history of boosting nazi posts on twitter. Oh, Musk posts so often he can't vet the source of all of his retweets. But if those are mistakes, the fact is he never makes a mistake in the other direction, which strongly suggests it wasn't an accident.

Eg, https://www.msn.com/en-us/news/world/musk-retweets-hitler-di...

	▲	oulipo2 5 days ago \| parent [-]
		[flagged]

▲

giancarlostoro 5 days ago | parent | prev [-]

Its focus seems to be on faster responses, which Grok 3 definitely is good at. I have a different approach to LLMs and coding, I want to understand their proposed solutions and not just paste garbled up code (unless its scaffolded) if you treat every LLM as a piecemeal thing when designing code (or really trying to figure out anything) and go step by step, you get better results from most models.