And don't forget that for the entire existence and our entire experience of LLM-based AI is based on the LLMs being constantly updated and refreshed. And it is widely accepted (and even confirmed by the labs themselves) that updating the models is undeniably unsustainable.

So, I'd love to imagine what the future looks like once we're stuck with a model (no matter how capable) that can never be updated ever again. What future versions of "do it for my grandma" can never be patched out?