That can be scaled horizontally and is just a matter of building more data centers.
That was a spectacular way of outlining the fact that you don’t at all understand the problems and limits with “AI” without having any awareness or understanding that you don’t at all understand the problems and limits with “AI.”
Ok, explain to me why you can’t scale out existing small and mid tier models horizontally? yes there are current resource limits on chips and energy but we know how to build those out and those types of limits are common to nearly every other industry, it’s just that no other industry has generated such rapid demand/investment for infrastructure.
There’s no O(n^2) problem on number of requests that would make handling large scale rollout impossible like the article is suggesting.
That was a spectacular way of outlining the fact that you don’t at all understand the problems and limits with “AI” without having any awareness or understanding that you don’t at all understand the problems and limits with “AI.”
Ok, explain to me why you can’t scale out existing small and mid tier models horizontally? yes there are current resource limits on chips and energy but we know how to build those out and those types of limits are common to nearly every other industry, it’s just that no other industry has generated such rapid demand/investment for infrastructure.
There’s no O(n^2) problem on number of requests that would make handling large scale rollout impossible like the article is suggesting.