the so complete pouët.net oneliner
- 2026-04-06
-
non-human entries =D w u such a troll -
I'm serious though. If Claude writes an intro (and it can actually, the thought process is impressive) it should take the full credit. -
Also, a separate party for AI agents (and Tom) should be organised. -
Anthropomorphizing tools doesn't make sense to me. If I will ever use Claude, I would take the credit and Claude would be mentioned in the list of tools. - Following this logic only art directors here should be in credits, while coders should be mentioned as tools.
- Claude doesn't "write an intro", just like your compiler program doesn't compile your code
- It seems that you guys (aside from Tom) don’t get “agentic”.
- Show me a compiler with a "thought process" that resemblance the one of a human.
- thought process =D
-
morning -
it's a tough process - HEY THIS PICTURE LOOKS LIKE A HUMAN SO IT MUST BE SENTIENT
-
Mornin' - it's nothing like a human thought process, but it's a thought process
- must be because it looks like one!
- it is, because it is producing iterative results with easily tracable reasoning in-between
- images of reasoning*
- https://ricedit.com/intro256_test.html here you go
- Blinky, Pinky, Inky and Clyde also do “resoning”.
- I asked my text editor to open some files and then I asked the compiler to compile them and then I asked my operating system to execute the program
-
400 requests, almost 10% on the way to the top! - It doesn't matter, someone forked and vibe coded improvements for my abandoned OptiDoom 3DO repo, if I was skeptical before, I saw such horrible slop and actually rendering broke and nothing of value added that I am convinced now it's worse than I thought
- Optimus: cause it's a toy, but not that stupid. Look at my link. "Dude" Claude goes back and forth. It surely memorized a lot tricks, but usually tries to apply them funny way.
-
https://www.remotelabor.ai/ - Good finding, Maali. But let's see periodically how well those agents can accomplish 256b intro challange end-to-end (no human help). I think it's also a good benchmark.