[AndrewKemendo]

Self host your own LLM

[singleshot_]

Why do you think this would be less discoverable than hosting your own email server?

[QuadmasterXLII]

If you use a stateless client (like just rawdogging cli llama.cpp) there’s nothing to discover. Setting a program with an option to have logs to not do that could conceivably get you in trouble but using a widely used program that never had logs seems like it has to be fine. Maybe they could nail you for googling “which local llm approach generates logs?” also, don’t get nailed by your bash history!

[kevin42]

Because you don't keep logs.

[AndrewKemendo]

Because nobody would know about it unless you told them for some reason

[nerdsniper]

That might fall under the “unethical” part of my question. Could “probably” get away with it if done carefully, but I’d rather be fully in compliance.

[dadoomer]

Why would self-hosting for privacy reasons be unethical just because the query would be subject to subpoena in principle?

[nerdsniper]

If you have records, and the discovery is worded in a way that covers the records you have, then deleting or not producing them would be unethical.

More importantly, I’d rather use SOTA models over self-hosted ones if possible.

[nvr219]

You’d need to hand that mac mini over if subpoenaed

[AndrewKemendo]

Can’t hand over something that doesn’t exist if it’s running in a VM container and gets destroyed every 12 hours