02:35
I have started experimenting with gpt-3.5-turbo via api and with some very lightweight usage i have already burned $0.5. It seems costly if you let it run inside loops and folds. I don’t understand why can’t they make it free like it is free to use from the website. It is costing them same money if you use it from web or api. Why so much discrimination against api?
There are now many open source models. I don’t know whether they can run locally on a basic Nvidia rtx card in a laptop.
11 hours later…
14:17
@gwr Also not a mod anymore for at least 2 years… chat.stackexchange.com/transcript/2234?m=58451371#58451371
« first day (4188 days earlier) ← previous day next day → last day (594 days later) »