[behnamoh]

Probably something you can create for yourself using some basic prompt engineering on GPT-4 or even 3.5.

BUT...

Do I want to share sensitive info with an API? No. Llama 2 might be a better option but it's way more expensive to self-host a 70B L2.

[creata]

> Llama 2 might be a better option but it's way more expensive to self-host a 70B L2.

It's free if you're okay with getting one character per second (and possibly melting your laptop).

[Multicomp]

i've been playing with gpt4all for localhost running chatbots, its amazing at how capable it is at such an early state. local LLMs 2 (to say nothing of say 5+) years from now will be much more refined, its early generative AI days.

This will all be quite interesting.

[dcow]

You can run 13B locally and it will do just fine.