AntonioCao

5 ms·

1.
by AntonioCao 3y ago
tbh for tools like Shazam there's no fundamental difference between a database + hashing algorithm and a self-supervised model; both are great indexing & compression solutions, just for different scales of data.
2.
by AntonioCao 4y ago
Wow. Can't wait for the lean theorem prover integration.
3.
by AntonioCao 4y ago
My first thought exactly. This essay by Hans Moravec is also a great exposition for similar idea: https://frc.ri.cmu.edu/~hpm/project.archive/general.articles...
4.
by AntonioCao 4y ago
This post https://twitter.com/madebyollin/status/1566838643771457536 is a step in such direction. I quite like the author's vision about neural nets eating through the graphics pipeline, given there are work
5.
by AntonioCao 4y ago
Make sense. A fun argument for why brain's underlying "algorithm" should be simple is that evolution didn't take that long to go from creatures with primitive cognitive functions to thinking-talking humans. The sensible
6.
by AntonioCao 4y ago
Honestly you should checkout open source + self-host alternatives like porter ( https://github.com/porter-dev/porter ). I tried it in a project before and the developer experience was surprisingly good.
7.
by AntonioCao 4y ago
Computer Graphics: Justin Solomon (MIT): https://www.youtube.com/c/justinmsolomon Keenan Krane (CMU): https://www.youtube.com/c/keenancrane Cem Yuksel (Utah): https://www.youtube.com&#x
8.
by AntonioCao 5y ago
Eric Weinstein also wrote an excellent analysis of Kung Fu Panda on Quora. https://www.quora.com/In-Kung-Fu-Panda-how-does-Po-end-up-de...
9.
by AntonioCao 5y ago
This is super impressive, would definitely pay for it if more polished. Maybe integrate with tldraw? Anyways, amazing job.