interesting that it is still 2 bit

Interesting topic, unreadable due to sloppification.

This has been tried by Cray like 15 years ago and the university of Colorado 11 years ago; there were also other initiatives, all failed.

I'm not excluding that now we have more budget/time/AI to make it work (if it can), I'd just say exercise caution reading this (very AI slop) article (and anything that looks too good to be true)

> Photonics

So also 1000x smaller integration