2 ms·
local models are getting pretty close. they still require some serious hardware to run at a reasonable rate but they're about smart enough to be useful.
by jayGlow 1mo ago
local models are getting pretty close. they still require some serious hardware to run at a reasonable rate but they're about smart enough to be useful.
- queenkjuul 1mo agoYeah i just don't have the hardware to run them at a reasonable rate and am not ready to invest that much at today's prices. My tinkering with DeepSeek last year was promising but the minutes-per-response made it unusable.