Comment by laughing_man
20 days ago
I suspect major algorithmic breakthroughs would accelerate the demand for GPUs instead of making it fall off, since the cost to apply LLMs would go down.
20 days ago
I suspect major algorithmic breakthroughs would accelerate the demand for GPUs instead of making it fall off, since the cost to apply LLMs would go down.
Sounds like the Jevons paradox. From https://en.wiktionary.org/wiki/Jevons_paradox :
> The proposition that technological progress that increases the efficiency with which a resource is used tends to increase (rather than decrease) the rate of consumption of that resource.
See also Wikipedia: https://en.wikipedia.org/wiki/Jevons_paradox
Some changes to the algorithms and implementations will allow cheaper commodity hardware to be used.
There will always be an incentive to scale data centers. Better algorithms just mean more bang per gpu, not that “well, that’s enough now, we’ve done it”.