I have low(er) confidence in any prediction that simultaneously require some exponential trends to continue and yet requires other related trends to stop.
These predictions are all based on exponentially-rising amounts of compute power. Unfortunately, as I’ve mentioned before, this exponentially-rising amount of compute power comes with exponentially-rising cost of fabs. At about 2080 or so at present trends a fab would cost more than the world GDP, which is obviously nonsense.
I would be interested in seeing a similar set of predictions, but with compute power held constant. From what I’ve seen things by and large still do scale exponentially, but with far longer doubling times[1]. (For instance: Proebsting’s Law is an observation that compilers roughly double the performance of the output program, all else being equal, with an 18-year doubling time[2]. And again with an exponentially-increasing amount of compute required to do the compilation...)
...and it’s actually likely worse. The 2001 reproduction suggested more like 20 years under optimistic assumptions, and a 2022 informal test showed a 10-15% improvement on average in the last 10 years (or a 50-year doubling time...)
I have low(er) confidence in any prediction that simultaneously require some exponential trends to continue and yet requires other related trends to stop.
These predictions are all based on exponentially-rising amounts of compute power. Unfortunately, as I’ve mentioned before, this exponentially-rising amount of compute power comes with exponentially-rising cost of fabs. At about 2080 or so at present trends a fab would cost more than the world GDP, which is obviously nonsense.
I would be interested in seeing a similar set of predictions, but with compute power held constant. From what I’ve seen things by and large still do scale exponentially, but with far longer doubling times[1]. (For instance: Proebsting’s Law is an observation that compilers roughly double the performance of the output program, all else being equal, with an 18-year doubling time[2]. And again with an exponentially-increasing amount of compute required to do the compilation...)
With exceptions.
...and it’s actually likely worse. The 2001 reproduction suggested more like 20 years under optimistic assumptions, and a 2022 informal test showed a 10-15% improvement on average in the last 10 years (or a 50-year doubling time...)