AI, Self-Awareness, and the Centroid Problem: Jennifer Hong on Leading in the Age of Machines | E5

So for those who don't know, Jennifer and I met on Substack.

And Jennifer is one of those people where whenever you read or interact with anything of hers, you learn more.

Jen writes a lot about the human connection, AI, and a bunch of different things that I'm going to ask about.

And I want to take us back to one of the first Substack pieces you ever wrote, which is a little over a year ago, if I'm not mistaken.

It's called A Beautiful AI Mind.

You ready?

Yeah, let's talk about it.

Cool.

So it opens with a line that's a real plot twist, and it's that intelligence is no longer ours alone.

And a year now from that piece, having watched AI conversations evolve in real time, do you still think that that line holds? Or has your read on what intelligence even means shifted?

I actually wish artificial intelligence was not called artificial intelligence.

I think it's a misnomer.

The artificial portion is correct.

I just don't think what it provides is intelligence.

by a very pure definition.

But if we set that aside and just say, you know, it's not going to get rebranded at this point in time, then yes, we have to share what we think intelligence looks like today.

And it's not just coming from other humans.

If you want to talk about the arc of how it's moved over the last year, I think it's moved a lot, especially for people who are AI fluent.

The challenge is, And if you define AI fluent by skillful use of multiple AI tools, I think it's like half of 1% or 1%.

It's a very small percentage of population that you can even consider quote unquote fluent.

But I would say that most of those people are now, it's less awe inspiring and it's more, where does it break? Let's actually try to make it break.

It's a lot of stress testing, a lot of people experimenting and publishing the results, which I think is good because I don't know that.

Well, I shouldn't say that.

I know that AI companies are doing a lot of that research themselves, but it's not as if they roll out new models and say, here's your user guide and

here's what to watch out for.

So for those who don't know, Jennifer and I met on Substack.

And Jennifer is one of those people where whenever you read or interact with anything of hers, you learn more.

Jen writes a lot about the human connection, AI, and a bunch of different things that I'm going to ask about.

And I want to take us back to one of the first Substack pieces you ever wrote, which is a little over a year ago, if I'm not mistaken.

It's called A Beautiful AI Mind.

You ready?

Yeah, let's talk about it.

Cool.

So it opens with a line that's a real plot twist, and it's that intelligence is no longer ours alone.

And a year now from that piece, having watched AI conversations evolve in real time, do you still think that that line holds? Or has your read on what intelligence even means shifted?

I actually wish artificial intelligence was not called artificial intelligence.

I think it's a misnomer.

The artificial portion is correct.

I just don't think what it provides is intelligence.

by a very pure definition.

But if we set that aside and just say, you know, it's not going to get rebranded at this point in time, then yes, we have to share what we think intelligence looks like today.

And it's not just coming from other humans.

If you want to talk about the arc of how it's moved over the last year, I think it's moved a lot, especially for people who are AI fluent.

The challenge is, And if you define AI fluent by skillful use of multiple AI tools, I think it's like half of 1% or 1%.

It's a very small percentage of population that you can even consider quote unquote fluent.

But I would say that most of those people are now, it's less awe inspiring and it's more, where does it break? Let's actually try to make it break.

It's a lot of stress testing, a lot of people experimenting and publishing the results, which I think is good because I don't know that.

Well, I shouldn't say that.

I know that AI companies are doing a lot of that research themselves, but it's not as if they roll out new models and say, here's your user guide and

here's what to watch out for.