2026-09-29 12:56 UTC
Replies (4)
-
@Warl0k3@lemmy.world 2026-09-29 13:04
“But have you considered that it might not be a fundamental misrepresentation of the capabilities of our system but instead: skill issue?”
-
@iLigator@lemmy.zip 2026-09-29 14:28
GPT 3.0 couldn’t create gta 6 using a single prompt therefore all future models are bad. /s Maybe get out of your cave and test it yourself instead of coping in denial.
-
@boonhet@lemmy.zip 2026-09-29 16:24
We didn’t even have proper harnesses 3 years ago. Claude Code which more or less kicked everything off is under 2 years old, before that you would’ve had to use Cursor or something. And we now have local models better than the frontier ones of 3 years ago if you have a good graphics card or an SoC with lots of shared memory. The cheapo Chinese models now beat the super expensive American ones from just a few months ago. Whole thing is changing so fast if you haven’t tried it in a few months, you don’t really know what it entails now.
-
@chunes@lemmy.world 2026-09-29 16:32
And you have been moving the goal posts for three years even harder. You used to laugh about how llms couldn’t count letters in a word, and you refuse to acknowledge that it has gotten exponentially better since then.