Insane demo. I knew it'd be something good based on the domain, but I was still very impressed.
...well, by the demo. The explanation sucks, it's sad to see yet another person I respect succumb to slop. It feels like the LLM prioritized making this sound grand over actually explaining what's going on.
It starts off talking about how it "draws an infinite Sierpinski fractal" - which no, it literally does not? That would be a great opportunity to explain the unexpected relation here, but the write-up just carries on as if this was just drawing the Sierpinski fractal. The way you twist the Sierpinski fractal into matrix rain - literally the point of the demo - is never explained, but "the resulting audio is a direct sonic representation of the mathematical structure", so I guess that's cool. This is a lot of text to say very little; I've learned much more from HellMood's various short comments on pouet. At least these didn't muddy the waters with garbage.
...I'm actually in a weird place to complain about this. I think sizecoding is genuinely fascinating, and while I have no skills to pull it off, I think it's really interesting to reverse these demos. Besides the article linked earlier, last yearI wrote a draft explaining lgbt19b (another brilliant demo from HellMood). I narrowly missed last pride month, so I was going to publish it this year instead. If this didn't come with a sloppy writeup attached, I probably would've tried to make a writeup reversing this too, out of pure awe. I don't know what to think now.
Wow, this is pure madness. While watching, I could believe the visuals in 16b, but not the sound. I've read the article and I still can't believe it makes the sound. Huge respect for this.
dzwdz | 8 hours ago
Insane demo. I knew it'd be something good based on the domain, but I was still very impressed.
...well, by the demo. The explanation sucks, it's sad to see yet another person I respect succumb to slop. It feels like the LLM prioritized making this sound grand over actually explaining what's going on.
It starts off talking about how it "draws an infinite Sierpinski fractal" - which no, it literally does not? That would be a great opportunity to explain the unexpected relation here, but the write-up just carries on as if this was just drawing the Sierpinski fractal. The way you twist the Sierpinski fractal into matrix rain - literally the point of the demo - is never explained, but "the resulting audio is a direct sonic representation of the mathematical structure", so I guess that's cool. This is a lot of text to say very little; I've learned much more from HellMood's various short comments on pouet. At least these didn't muddy the waters with garbage.
...I'm actually in a weird place to complain about this. I think sizecoding is genuinely fascinating, and while I have no skills to pull it off, I think it's really interesting to reverse these demos. Besides the article linked earlier, last yearI wrote a draft explaining lgbt19b (another brilliant demo from HellMood). I narrowly missed last pride month, so I was going to publish it this year instead. If this didn't come with a sloppy writeup attached, I probably would've tried to make a writeup reversing this too, out of pure awe. I don't know what to think now.
lukasl | 9 hours ago
Wow, this is pure madness. While watching, I could believe the visuals in 16b, but not the sound. I've read the article and I still can't believe it makes the sound. Huge respect for this.
fanf | 3 hours ago
Elsewhere… https://www.pouet.net/prod.php?which=106210