auwsom

You can use this bot I'm hosting for the time-being. Or you can host your own in GCP: https://discord.com/api/oauth2/authorize?client_id=1110757452032520264&permissions=326417517568&scope=bot btw, the bot command to turn it on is `/auto-thread on`...

@MarcusOtter sure, no problem.. just to be clear, it makes a thread *only* when the reply *is* used. This is how Slack is by default. Maybe we can make it...

@Junebugging check out my code (and some details in my comments above) I'll repost the link to my fork here in a second.. https://github.com/auwsom/discord-needle-reply One thing i wanted to check...

Yeah, actually doesnt work at all right now (nothing copied), but works on firefox.

I couldnt get the 40b model to load on a 3080 with 4bit quant.. anyone have success with a single 12GB GPU?