2020 BRN Discussion, page-23721

I was reading on the weekend that Samsung and TSMC are both aiming for 3nm and Samsung 2nm.

We know from the CEO and Peter van der Made three things:
1. You don’t need to use all of AKIDA IP to bring an AKIDA function to your product.
2. Because AKD1000 is digital it is scalable down to 14nm, 7nm etc; and as you go down it will become faster and use less power.
3. AKD1000 only uses 5% of its capacity to perform Covid-19 detection at 94%.

So this leads me to what could you do with 2nm chips that have only 5% of the AKIDA capacity at 28nm when at 28nm that 5% can detect Covid-19 at 94% accuracy remembering that when testing AERO on the 10 gas data set (each set containing 10 gases of that class) the ADE was implemented on only 200 AKIDA neurons and produced state of the art accuracy with 3 second latency. My mind boggles at the possibilities.

My opinion only DYOR.