Real Magic

Do you know how your phone works?
Do you know how GPS works?
Do you know how Netflix works?

Not how to make it work. Not generally speaking. Not at some basic level.

Do you know the details? The circuitry, the RF signals, the semiconductor devices, the software. The weeds. Do you know how all of these actually work?

I don’t. Not all of it. And I’ve spent a good portion of my career working on various technologies in these applications.

This stuff seems like magic. Even to professionals in the field. But it’s not.

Which brings us to AI.

Your phone, GPS, and Netflix are complex, but deterministic. There is a definitive answer to “how does this work?” If you spent enough time and had enough capacity, you could theoretically, at least, figure it all out. If you did the same thing two times in a row with identical conditions, you’d get the same output. You can predict it.

AI doesn’t do that.

AI is a mystery, even to those who build it. Sure, they understand what they built and why it acts the way it does, and how to point it in the right direction. But they can’t predict the specific outcomes.

AI is real magic.