Re: Nature contains a synopsis of the results of quadrillions of successful experiments in molecular nanotechnology, performed over billions of years [...] My guess is that an AI could re-do all those experiments from scratch within three days. Or maybe nanoseconds. Depending on whether it starts the moment it leaves the lab or as a Jupiter brain.
Real experiments take time—especially when they must be performed in series. Small experiments can be done relatively quickly—but not all of nature’s experiments have been small.
A superintelligence will perform experiments, of course, but it seems likely to me that it will also continue the job we have started, that of harvesting existing results. The approach has given us a lot of mileage so far—and seems to represent vastly less work than the alternatives.
It’s true that it may eventually replace DNA and protein—and thus largely lose interest in much of the “old” biosphere, but again, that will probably not happen overnight—or indeed over three nights—and even then, it will probably want to preserve some of it, to help it understand its origins.
Re: Nature contains a synopsis of the results of quadrillions of successful experiments in molecular nanotechnology, performed over billions of years [...] My guess is that an AI could re-do all those experiments from scratch within three days. Or maybe nanoseconds. Depending on whether it starts the moment it leaves the lab or as a Jupiter brain.
Real experiments take time—especially when they must be performed in series. Small experiments can be done relatively quickly—but not all of nature’s experiments have been small.
A superintelligence will perform experiments, of course, but it seems likely to me that it will also continue the job we have started, that of harvesting existing results. The approach has given us a lot of mileage so far—and seems to represent vastly less work than the alternatives.
It’s true that it may eventually replace DNA and protein—and thus largely lose interest in much of the “old” biosphere, but again, that will probably not happen overnight—or indeed over three nights—and even then, it will probably want to preserve some of it, to help it understand its origins.