Domi

That’s amateur film maker stuff.

In real film making, people become the head rest: https://www.youtube.com/watch?v=kxb9xzAaYjM

Now I need to give One Cut of the Dead another go.

I also stopped 20 minutes in. Twice.

Visual Language Models, like LLMs but they read images and text.

The new VLMs are much better at solving captchas than I am. Especially the older ones with the squiggly text, no way I’m doing those first try.

39 GB is very small, DeepSeek R1 without quantization at full context size needs almost a full TB of RAM/VRAM.

The large models are absolutely massive and you will still find some crazy homelabber that does it at home.