After yet another awesome announcement by <DeepMind> that it had improved theoretical 4x4 matrix multiplication reducing the number of scalar multiplications with its <AlphaEvolve> system, I decided to have a look at the smallest open size 3x3 to understand what was going on in there.

I've dumped what I gathered at:
* https://math.stackexchange.com/a/5065930/53203
* https://www.reddit.com/r/math/comments/1komgnr/til_you_can_multiply_two_3x3_matrices_with_only/

Announced at:
* https://mastodon.social/@cirosantilli/114523670306611862
* https://x.com/cirosantilli/status/1923749179458773429


Understanding the state of 3x3 <matrix multiplication>

Understanding the state of 3x3 matrix multiplication

Updates <span class="meta">/</span> Understanding the state of 3x3 matrix multiplication

{c}
{tag=Google acquisition}
{title2=2010-}
{wiki}

They seem to do some cool stuff.

They have also declined every one of <Ciro Santilli>'s applications for software engineer jobs before any interview. Ciro always wondered what does it take to get an interview with them. Lilely a <PhD>? Oh well.

In the early days at least lots of gamedev experience was enough though: https://www.linkedin.com/in/charles-beattie-0695373/[].


Ciro Santilli @cirosantilli 40

 Incoming links: DeepMind