byrneml

This is a much better solution than having vertical focused GPU clouds. You need other compute primitives for non-GPU workloads, Northflank offers the best of both worlds

Good alternative to HuggingFace Datasets linked in this tweet. Today shows why we need more than one company in this space!

LLMs are time dilators for knowledge work

Could you apply the same technique for real-time video editing?

Pretty comprehensive guide on doing weird things with LLMs on CPUs.