I mean, let's try to build a von Neumann probe. How? I don't believe that our knowledge will continue to increase exponentially. Hard AI may never be achievable

I mean, let's try to build a von Neumann probe. How? I don't believe that our knowledge will continue to increase exponentially. Hard AI may never be achievable (although von Neumann probes don't necessarily need it). In order to build a von Neumann probe, once you need the very first microchip or anything resembling one, you're screwed. Will the von Neumann probes also be able to build an entire fabrication lab?