Robin Hanson: AI Will Inherit Humanity (And Change Everything)

Robyn, we are starting to see AI agents perform real work, writing code, conducting research, operating computers, and completing business tasks.

Are we already seeing the first signs of the M economy that you imagined? And for those who don't know, what is the M economy?

I have this book, The Age of M, Work, Love, and Life, and Robots Rule the Earth.

It's about a certain kind of AI that's easier to analyze.

So, in fact, when you say AI, you're pointing to this vast space of possibilities, all these different possible digital minds.

But we don't know where in that space we're actually going to realize.

And that makes it harder to figure out what a future of AI would be like.

So, the kind of AI I analyzed in this book was about very human-like AI.

We take an individual human brain and you digitize it and you make a model in a computer of that person's brain and now it acts just like it would.

Otherwise, except now it has more possibilities.

So that's the world I analyzed in Age of M.

And we are now seeing digital minds show up in some forms, but we really don't know where this is going to go.

There's still a vast space of possibilities.

And so...

If we were about to have the age of M, I could tell you a lot more about what we're about to see, because that was what I did in the book, try to analyze what will a world of M's look like.

But a world of AIs, more generally, is just really hard to specify, because you just don't realize how many different possible AIs there are.

It's a vast space.

You wrote the book 10 years ago.

Now looking at that, what did you get right? What did you get wrong at this point?

Well, this book wasn't trying to predict 10 years later.

It was saying that eventually we'd have this kind of computer that would change the world.

But we haven't seen it yet.

So my book is neither right nor wrong in that it was talking about something that hasn't happened yet.

Now, we have seen a lot of advances in the LLM kind of AI.

And, you know, that was something I didn't predict 10 years ago, but I wasn't trying to.