A couple days ago, I wrote about AI and my world. Today I’m going to branch out a bit beyond little ol’ me…
A lot of people these days are afraid of AI. Is it going to make us humans obsolete?
When I say there’s not much to be afraid of, I tend to get pushback in two areas:
One is that AI is advancing very quickly. This is absolutely true. If you look at where it was a year ago, and compare it to where it is today, well, there’s no comparison. AI is, after all, largely based on the processing power of chips like those from NVIDIA, and the development of those chips follows Moore’s Law, which, if I remember correctly, is that the processing power of chips approximately doubles every two years. In other words, a chip today can do twice as much as a chip two years ago, and a chip made two years from now will be twice as powerful as a chip today. So yes, AI is bound to continue progressing at dizzying speed.
However, progress in one area has been glacially slow, and that is designing a program that is sentient, that knows us and the world better than we do. In other words, we have chips that can process information faster than anyone thought possible, but the programs they run are hitting the limit of us — in other words, they can’t get outside of our own heads. I guess it’s entirely possible that someday that limit will be breached, but we humans have been trying, unsuccessfully, to breach that limit for thousands of years (even going so far as inventing religions to try to do it), and we haven’t been able to. So I suspect the chances of that limit being breached aren’t that good.
Another pushback I hear is that a tremendous amount of investment is going into AI these days, and all that money is just going to speed up the race to sentiency.
However, investors tend to know that the race to sentiency has been glacially slow, so where is that investment going? It’s going into building the infrastructure to support the filling out of the fat part of the bell curve. In other words, it’s going to fund the energy and data centers and connectivity that will be needed as the rest of the world adopts the tools that already exist.
Okay, that’s enough of me spouting off on things AI. In the next edition, we’ll get back to our regularly-scheduled presentation-related programming.