This is some cool shit:
https://arstechnica.com/information...

Now I want to learn how to AI at least enough to understand what they are doing.

People worried about AI replacing programmers when it was the math people software has been replacing.

this is really cool! love it

https://nature.com/articles/...

In layman terms:

Hey... AI... We have this game for you: Try to find the solution for this - least number of steps win. We only limit you by number of moves.

Now go and fucking brute force the shit out of it.

That's why AI will always find solutions that might not have been thought of before - it is basically an solution generator without human bias.

(Edit: Yes, it does far more than that, but I reduced it to the basic gist. Strictly speaking it's not brute force either, as Alpha Tensor has a shared "knowledge"… So it is more like "prediction of all futures, playing them out and finding the best future possible, then restart based on that point in time". Easiest, though not fully correct description I can formulate)

@IntrusionCM I bet this is playing out in kinetic and non-kinetic warfare as well. Best/fastest computer wins?

Ahhh I can't unsee the pixel imperfection in the example image.

( The numbers aren't actually centered in the circles )

@Demolishun I'm not an expert on that part.

Though I think that AI for warfare is... Far more complex.

I really don't want to imagine it, but spontaneously my brain says... You will not need the best computer, rather the most accurate data paired with the heaviest (possible unknown) server centres combined.

Just think about the exponential number of variables involved... Easiest to imagine is terrain, but then there's weather, weight of troops / equipment, enemy troops / equipment, movement predictions etc... If there is an AI for that, I really don't want to know it's size.

It must be ridiculous obese...

@ElectroArchiver I only read it for the article, not the pictures.