Yup, I saw it, what is he actually trying to say ? How can analog computer supercede digital, analogue is made too specific for only 1 task

Your CPU has dedicated circuitry for performing CRC32 and AES rounds. That's as specific as any analogue computer...

And your mobile SoC likely has h264, h265 and VP9 dedicated circuits also, so it also works for quite complicated tasks too.

If that "1 task" happens to be performing matrix multiplication (or even merely fused multiply add), you can do a heck of a lot with that. You still need digital circuitry to support the IO, but the key idea is doing linear algebra in a way that is faster and/or generates less heat per unit compute.

We have tonnes of computing power dedicated to repeatedly solving a simply-parameterised problem.

Stupid question... with todays memory capacities....At what limit/size do we stop using matrix-ops and simply use lookup tables todo 'matrix-math' ?

Not a stupid question. Economically, memory density will hit a brick wall soon. Developers should prefer to waste time and save space, since parallel computation will not hit a similar limit in the foreseeable future. Memory-to-core ratio is going to be falling.

TLDR is you can't. For a very simple example, storing the products of all 3x3 matrices * length 3 vectors in Float16 precision would take 2^193 bytes (which is obviously impractical).