chocrates

Jeffries seemed to indicate it’s not a priority

They need 2/3’s. So 67 i guess

Fortnightsexual

🤣

I love him as Avery Bullock in American dad

Might be able to find a clever way to store state outside the model. Also probably building hype to fund more AI.

I don’t know much about world models but they are touted as having a higher peak “intelligence” than LLM’s

I have a menu bar plugin that shows it, don’t remember where I got it?

I’ve heard that GPU’s for this work are exceedingly hard to get ahold of as well. Glm 5, which i think is near frontier model capabilities, takes around 1.5 Tb of vram un-quantized.

I hate it of course. I need to find an exit strategy

I spent $200 in one session the other day, watching the clanker design, implement, and review the change. Still have no idea what I did but my name is on the commit

I’ll do some digging, it doesn’t work on my Ubuntu Wayland as of last week

does screen share work on Wayland yet?

No not at all. I feel like I need to sell a kidney, but I got to 64gb vram

How big is un-quantized 2.8t param model?

I just got some hardware but I still don’t have a ton of vram :(