What do you guys think of the GPU aggregation space? I've seen a number of startups that aggregate unused/underutilized GPUs with their own unified software API/SDK layer. Given Nvidia's no end in sight demand, this seems to be a good space to be in. Any thoughts? 🤔

I think it's probably going to be a big space, especially if you can provide a good API on top of it. I would love to be able to call out to a decentralized generation API, something like folding@home but for language models. I feel like I saw someone doing something like this but I can't remember who

Yeah, I know of at least 2 companies doing this