I think you're getting a bit mixed up with a local model verses an agent. Ollama or LM Studio are good options for running models locally - in general, you want as many GPUs with as much VRAM as you c

And also, about getting cheap $7/y dealz, the main issue isn't just about hardware but IPs, electricity, networking costs, etc.

Super cheap deals always come to an end. Or in the worst, the dealpool comes.

(Quote)

certain tlds are extremely cheap at regery

Electricity is expensive in SEA for the most part, and transit is not cheap.

(Quote)

do you think that there aren't many cheap servers?

(Quote)

(Quote)

(Quote)

(Quote)

The price isn't cheap:

@Felware is in Milwauke, Wisconsin

Not cheap though, not in LET budget, not 7$/y

(Quote)

(Quote)

(Quote)

I'd be down for A super cheap V6 only high storage, I'd merge it with my existing server making it a storage expansion, Even though I'd have to re-setup the server.

sucks to see OP doesn't care that much about the auction, no rules, chatgpt writing, and no mention of the DC that the server is in/potential colo rate. Would be super interested if OP was more forthc