And so, we are happy to learn that nowadays distillation of those pre-trained models, fine tuning a quantized version of it, recently Stable Diffusion Turbo, all showed us that it is now quite possible to tune even the most cutting-edge models in a way that fits even a laptop. Indeed, I’ve been tuning my email replying model purely on my laptop since March. And so, my emails are drafted by AI models, and of course, I read all of them before hitting send.