Hacker News Reader

Do GPUs at a reasonable price next and you'll own the ML space.

R2 will be amazing for hosting computer vision datasets and models.

Working on that too: https://blog.cloudflare.com/workers-ai/

Great start but the field is moving on from tensorflow. Torchscript support would be amazing.

If we can get GPUs deployed at the scale we want, adapting to whatever language or framework you want to use to program them will be easy. The challenge is getting GPUs into enough machines in enough places to make this interesting. But, again, we're working on it.

I'm not sure if the GPU for ML space is a fit for cloudflare.

It is for a cloud hoster. But ML does not take advantage of cloudflare's competitive edge ( their SDN ).

In the same way, i don't think docker containers are an interesting/good fit ( i could be wrong on this though)

Edit: i was wrong...

I have an image recognition service on GCP where majority of the latency is the initial upload. Pushing that closer to the user would be nice.

> It is for a cloud hoster.

Which is absolutely where CF are headed, as evidenced by R2.

Don’t forget that S3 was the first AWS product. It’s fitting that it’s also CF’s first real foray into cloud-land.

Well, i thought cloud hoster together with the advantage of the edge.

I thought that ai wasn't interesting on the edge.

Reply on news.ycombinator.com