I believe the fixed split thing is a historical remnant. These days, the OS can allocate memory for the GPU to use on the fly.
I want to know if it's possible. 4GB for Linux, a bit of room for the calculations, and then you can load a 122GB model entirely into VRAM.
How would that perform in real life? Someone please benchmark it!
This is sort of what I feel like the parent was implying without outright saying it, too. But: that's not how Slack works. If the only reaction is an emoji of skin tone A (let's say its a dark skin toned thumbs up), and I click that dark skin toned thumbs up to also react, Slack adds a thumbs up with the skin tone configured in the user's settings (which I believe defaults to no skin tone), not the skin tone of the initial reaction.
Settings → Messages & media → Default Skin Tone