infrecursion (@infrecursion1) profile picture
Joined April 2020
Replying to @koltregaskes
And it shows haha :)
1
3
111
Replying to @zsakib_
Lol, why didn't you post the nanobanana pro result for this word? It's equally garbage.
1
4
54
Replying to @bcaine
fwiw, I used your prompt on ChatGPT 5.2 extended thinking (probably medium on API) and it gave me this image. It identified more components that Gemini 3.0. But Gemini 3.0 seems to be cleaner. chatgpt.com/share/693b9f85-6…
1
8
1,116
Replying to @simonw
That's strange because gpt 5 mini on ChatGPT gives a pretty good result. Maybe too much rl has ruined the other capabilities?
1
114
Replying to @lpachter
You just need to prompt properly I guess?
1
5
1,047
Replying to @JonasAAdler
Did you actually ship a worse model and that now tops lmarena? As if anyone needed proof how cooked lmarena is lmao
43
Replying to @brave
Maybe don't call out others for shady shit when you did the same in the past? Instead of negative advertisements undermining others just instead show what good things you've got.
20
Replying to @Rainmaker1973
"The development continues for 13 pages recounting notions in electrodynamics and thermal physics. Then it takes a dramatic turn, without any warning of the transition, we find the basic notion of general relativity, the "line element" written at the top of the page." - Bonkers!
71
Replying to @DavidDeutschOxf
I get a very different response with ChatGPT 3.5
1
58
Replying to @sapinker
Bing chat can find both (but after a short confusion):
63