AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
Is It a Standard Call or a Generic Call?
So let's say I wanted to write an accelerator for matrix multiplies. Can I just like go ahead and be like, is it a standard call? If you wanted to write like blast library, you'd probably have to like cook your blast,. You'd have to have to redo your blast back end or something like that to call out to your accelerator. So we at some point we imagine that people will want to run different programs on sort of different hardware SKUs, if you will. But once that at some point it would be a reasonable use case.