Cryptocurrencies never did this with the entire computing industry because it got its act together and efficient blockchains arrived without the need to constrain the supply of CPUs, GPUs and memory chips to the point with drastic price increases, and we have faster blockchains handling billions of transactions a week.

Just look at what AI (in the form of LLMs) is doing to the rest of the computing industry because of throwing insurmountable levels of debt into data centers instead of researching efficient methods for running 1TN+ parameters language models locally or even to gain the same performance, intelligence equivalent without such large parameters.

It just tells you that AI is at the point where personal computing is going to price out a lot of people if it doesn't get cheaper. Until there are viable efficient methods in running 1TN+ parameter models or a smaller model performing at the equivalent or better than frontier models, we will continue to see more of this in the future.