Large language models are already seeping all over and magnifying our intelligence and abilities to do more in less time.

Once they become specialized tools marketed for accomplishing specific goals and integrated with things like calendars and clocks they will have far greater impact as personal assistants.

Probably by next year everyone will have open source language models of amazing quality. Facebook's Llama 65b is very good quality from my tests but the video card to run it is 16k. The open source community is working on LLM optimization, we already quantized it to 4 bits reducing rendering costs.

Once open source models surpass openais closed source ones we will have an insane intelligence explosion that will cost us very little. Personal assistant ais will uplift every human one at a time on a personal level improving quality of life for everyone who uses them.