We're well on our way of being able to speak to it at least though. With all the text-to AI's coming out, we've been successfully training it exponentially over the last year. I'm an absolute potato, but I'm even able to make stable diffusion do my bidding, and it's all with natural speech. 3d printers will need a couple of generations to build up to replicators and the rest of the... you know, forming reality in a way you can feel it.