Why people run models locally?
retained text
It might be a silly question, but why people are spending tens of thousands of Euros to run models locally? I can understand doing it for privacy, but I don't think that's the reason that drives most people to do it. When GPUs can be rented online for a fraction of the cost, a situation that becomes even worse for models, and when hardware improves at a breakneck pace making current generation quickly obsolete, what makes people think that investing in a local system is the best choice?
document fetched 2026-10-15 00:46 · canonical https://reddit.com/r/LocalLLaMA/comments/1x5rmwo/why_people_run_models_locally