no

Give us the dramatic sounding take 😄 I agree , good source of news but in small doses only to protect my sanity

A bit of this, a bit of that #foodstr

RTX 3090 as well. “MTP” is multi-token prediction. Details here: https://unsloth.ai/docs/models/qwen3.6 “UD-Q4_K_XL” is unsloth’s 4-bit quantized model. I want to keep context in VRAM for better performance and 27B quantized 4 bit seems to be the sweet spot on the 24G VRAM of the 3090

Awesome. Partly encouraged by listening to you and Justin talking about the power of local AI for freedom tech, I started on my own local setup. Currently running Qwen3.6-27B MTP UD-Q4_K_XL on an RTX via llama.cpp on Linux (Pop OS) and Hermes Agent as my local brain. Eye opening how capable it is . I’ll be setting up a second 3090 as the next step. Onward

Installed it and sent some sats to it via lightning. So now they are in a VTXO? So cool 😎

This is what I see

Ah that changes everything 😄

#4, best wings angle