I would not dare, I know the transistors are still "huge", even with this 'n2-highNA' process, but it _seems_ mass scale production is about to get horribly har

I would not dare, I know the transistors are still "huge", even with this 'n2-highNA' process, but it _seems_ mass scale production is about to get horribly hard in order to have robust chips after 'n2-highNA'. Only some honest guys from chip R&D can answer that. Oh, and GPUs can be parallelized, BY A LOT, to a point the real limit will probably be the electric power to drive them.