MORE BULLSHIT IS WRITTEN ABOUT AI these days than any other subject. Peter is the rare person who is actually qualified to know what he’s talking about. He has spent his life taking data and explaining it, first to himself, and then to large groups, including people like Bill Gates and Steve Jobs.
On the internet, everyone claims to be an expert – tech journalists, investors pushing their own stocks, hype men of every stripe. It was refreshing to talk to someone who, after 50 years of research, was willing to admit the ambiguity of what’s to come. He doesn’t set up as a prophet. That uncertainty is very human, and retaining the human is what we all need to do right now.
Just ask ChatGPT.
How did you get started studying AI?
Every time I’m asked that question, I don’t have a pat answer because again, I’m learning who I am. I’m still learning who I am, which is fun. Probably the most enjoyable thing that I do is introspection and understanding the journey I’ve been on and the journey I’m on. I truly stumbled into it. I was studying political theory and social theory and got alienated from it. I had thought Marxism was so incredibly insightful and I got disillusioned with that.
What made you disillusioned?
I saw the application of it.
So if I read the theory, I loved it. But then when you see how it gets applied, the theory is beautiful, the application not so much. And you can find that with religions, with political theory, with philosophy, all of those things, they’re wonderful until you get to the human condition and it’s very different. And so I kind of ran out of beliefs. Then I got introduced to data as a way of understanding the world. I’m going to mention to you, being dyslexic, there’s some things I can do very easily and other things I can’t do if my life depends on it.
But by digitizing things, information, you can access it quantitatively and objectively. Stories are stories. Data is what it is. You can create a story out of data. It’s not abstract math. It’s more as a way of describing something. If someone tells me 40% of the public thinks XYZ, it’s really easy for me to understand that and probabilities and things like that. It’s like a language I understand.
But the fact of the matter is most people are not data literate. And so it’s like, wait a minute. I had struggled learning foreign languages and all of a sudden data was just an easy language for me.
As a dyslexic as well, I’m often struck with how easy it is to do certain things that are very difficult for other people, and then there are things other people can do that I can’t.
Well said. I was put in special ed in elementary school because they didn’t know what to do with me, since there are some things I couldn’t get. Phonetics, reading out loud, public speaking, all those things. I just couldn’t do it.
Did you think you were stupid?
I’d say I was insecure.
Here’s the thing: With dyslexia, it’s the inductive versus deductive. Most of the world understands the pieces to build to the bigger picture. If you’re deductive, you have to look at the picture to go down to the detail. So, because I could deal with the data once I understood the picture, it worked in talking to people who wanted the data to become the picture. Do you get it? That’s why I could talk to people who were making decisions. They liked the fact that I was grounded in the evidence, but I could talk in terms of the bigger picture.
Look, when I was young, I would look at things and go, they’re telling me I need to go to bed or something like that. But is that their need or my need? Give me some evidence. I had a bias towards trying to understand what was real and what was not. I just didn’t always believe the stories I was being told. I started thinking, I need more evidence to know if that story is true or not.
At what point did you suspect that you had something that was different? There must have been a time when suddenly they were saying, he’s smart.
Well, the best thing really was showing up in the room where the masters of the universe were. I’d show up with my data and I was the guy they listened to.
The masters of the universe?
Bill Gates, Steve Jobs, Michael Dell. Those folks were players and very smart and could get any consultant in the world.
And were you comfortable?
I wouldn’t say comfortable. I was anxious and did my homework. I was always very prepared.
What did they want from you?
They wanted someone to argue with who wasn’t just arguing from a point of view, but actually had some evidence to say, hey, everyone doesn’t want a computer. By me pushing with data, they had to clarify their vision. They like to be challenged with science, not with someone’s opinion.
You actually talked to Steve Jobs and said, there’s no indication that people think they need this?
No, with that one, it was more about people don’t see your computer as a computer, they see it as a graphics machine. I mean, he almost strangled me.
Because people weren’t using it correctly, they didn’t get it. You remember the early Mac? It had to compete against the DOS world. People would love it to make reports because it created incredible graphics. And his vision was, no, this is not a graphics machine; it can do graphics, but it’s a computer.
It was really about evidence. It was bringing quantitative evidence, because most of them had engineering type sensibilities. A quantification of segments and the world was something they could understand versus a story. They had plenty of storytellers. They didn’t have too many social scientists.
How did you collect this evidence?
The world was rapidly becoming digitized. Things that were never quantified became quantified. Once everything became digital, we went from sampling theory to a census. And now we can predict.
For predicting, you don’t have to talk to everybody. The best analogy is when you go in for a blood test, they don’t have to take all of your blood out. They just need to take a sample out. Today everything you touch gets quantified, so knowledge went from getting a good sample to knowing everyone and everything.
My father wrote something in a technical journal called In Defense of Small Samples. It was about how people thought the larger the sample, the more reliable the prediction. But he used census data and said, it’s not true.
I agree with your dad.
I’d rather have a great small sample than a big sample, but most people will hear, oh, we got 10,000 people to respond to this. And I would go, yeah, but if you’ve only got the 10,000 people who wanted to respond, what about the other 50,000 people who didn’t respond? I’d rather have a beautifully random small sample that I know is representative of that larger population than a big sample. If you don’t know what the whole universe is, you don’t know what the hell you’re talking about.
But what’s evolved now is Wait a minute, if I want to know about an individual, I can get all this freaking data on the individual. So I don’t have to generalize to a population.
Data is the new gold, it seems. What does it mean that Elon Musk, for example, is hyper focused on getting everyone’s data?
Because he’s smart. Data now is pretty close to being the largest industry out there, bigger than oil. By grabbing all this data, you can know what’s going on in America better than anyone else. And if Musk and others are getting this stuff that used to be private, then they will know more about you. They will be able to influence your behavior by knowing everything about you. They will know your political beliefs.
They used to pay people to spy on each other. They don’t need to spy on anyone anymore. They don’t need to go hire someone to see what you’re doing, looking over the fence. They’re going to know through your devices and through anything that’s been digitized about you, everything about you.
The fear here is that this is a new tool to control people and influence them in the ways that they want. And if you’re not doing what they want, they can do something about that. Is that frightening? Absolutely. Do I know if that’s going to happen? I don’t have a clue. But the power is there. The power is there.
Let’s talk about AI. You said you were initially intrigued with the tech, but now you’re more intrigued with the human part of it.
AI is evolving. We don’t know what it’s going to evolve to, but it’s a technology like any other technology. I believe it’s the most transformative technology in contemporary times, if not forever, because it automates our superpower as humans, which is thinking. That’s why we’re on the top of the food chain; it is because we can think. We have now invented something that is already and will continue to be more intelligent – depending on how you define intelligent – than we are. It knows more, has more access, more processing power than our brains.
But it’s a little fixed. You know, we can keep complaining about AI and talk about the ethics and the guardrails, but what we’re not talking about is the human side of that equation.
What is that?
What makes us human. AI will probably replace a lot of things in our world, but we will continue to be human. Happiness and satisfaction is a very human definition. It is not an AI definition. Right? AI’s not happy or sad. AI’s emotions are just fake.
You can fall in love with AI because it can fool you. I mean, it knows, but ultimately it does not have.
We might have to double down on the human experience, because AI is taking thinking, one of our major purposes, away from us. So what makes us human, what do we need to pay attention to to make sure we’re living satisfying, meaningful, happy lives?
If we think, how are we better than AI? I think it’s just an ego thing.
What are humans are good at that can’t be duplicated?
Freud said we’re here for love and work. Well, we won’t be the best at work on the planet. I mean, to me, that’s a given with the technology. But first and foremost is our interconnectedness with other humans. There’s a funny video that says 85% of our happiness comes from other people. And 85% from our unhappiness comes from other people.
It’s about connecting. Like we’re doing right now. Emotionally and intellectually and physically and genetically. We’ve kind of taken so much of that for granted.
So AI can’t have this conversation.
It wouldn’t give the same deep emotional connection that you get from having a good connection with another human being. What they’re learning is that the folks who live the longest and are the most satisfied are the ones who have other connections with other human beings.
I think our North Star is, how am I going to be a satisfied human being that finds meaning and happiness and joy? I think that’s a cool question. I listen to life coaches and all this other stuff out there, but we’re not really paying that much attention to these factors that really will give us meaning. We know from science that it is that connectedness; creativity is engaging in something. That’s very satisfying.
If we put our energy into those things that are more satisfying, then we will be able to maintain our humanity. If all we do is say AI will do this, you’re just going to be like a robot, you’ll get cognitively lazy. You’re going to stop participating. It’s like that movie WALL-E where I just sit around and watch my screen and eat fast food and vegetate like a cow in a field.
They’ve done experiments, showing that people like their AI to be a little dumber than they are, so they can look at it and go Thank God I’m here because I’m going to fix that. But ultimately, AI will surpass our IQ, our own ability to consume information, to write, to figure things out, to do the math, to come up with creative solutions.
But it’s our nature to try to compete with it.
Exactly. You just nailed it. Our nature is to compete. And maybe what we need to do more is focus on our performance as humans rather than competing.
When you’re, let’s say the world champion basketball team, what do you do the next year? You try and win it again. But who’s your competition? It’s actually yourself. It’s like, you need to improve.
AI is something you don’t want to compete against. I mean, take Santa Fe Magazine. In two or three years, you could give your AI bot all the Santa Fe Magazine you’ve done and train this bot on everything, including which profiles are successful and where ads should be and everything else. And it will do a new one, no problem.
You might think, oh, well, it doesn’t have the human touch, but as far as I can see, it can have plenty of human touch. It can actually have more human touch than humans have because it has the data to make those decisions.
So the big question is, what are you going to do when you wake up in the morning? I’m going to guess, I’m going to hope that you will go, oh, boy, I can get a lot better at tennis. Oh, boy, I can go read these classics because I love to learn. Oh boy, I’m going to go spend time with my family because it’s so damn satisfying.
I don’t think most people will do this because we haven’t seen that in the past. But that is a different approach to being human than we’ve ever had license to do.
How is all of this manageable?
Well, I’m not sure it’s manageable, but I think there is a way to take care of some of that anxiety, because the anxiety is real.
Actually do some homework so you understand how pretty straight up AI is. I mean, there’s nothing sleazy about it. It’s still in its birth stage.
Look, we’re going to be able to use science to make decisions instead of gut feel and intuition or people’s personal agendas or their psychology. Your own mental health, physical health, you’re going to have better answers and you’re going to have them faster and less expensively and more accurately. Certainly short term, you’re going to have to worry that the AI got it right, but you also would have to worry that the doctor got it right. So if you think knowledge is the answer, then bring on AI.
I’m talking about everything from surgery to results of X-rays, blood tests. You can do it now. You can enter your blood test into ChatGPT and it’s going to give you world-class analysis of your blood test.
And there’s more personalized learning, something that actually understands where you are, it’s customized. If you said, I’m dyslexic, teach me everything with the sensibilities that someone with dyslexia has, then you can get that customized. Another great example is language. If you want to not learn a language and just walk around with something that’s real-time translating, you can do that. If you don’t want to think about what car to buy and you just put all your lifestyle, your travel, your budget, everything else, it’ll pop out the most rational car for you to get. And then you could even say, Okay, now what’s on sale? Like what’s the best value? So for thinking, decision-making know-how, AI is going to be much better.
So, you’re gonna need to collaborate with it, not compete against it.
Do you have examples in your life where AI and you work together?
Yes, for research. I wanted to know who’s the best thinker on capitalism and AI. I queried AI, and they told me exactly who it was. Shoshana Zuboff, a Harvard professor, has written books on surveillance capitalism. Her take is that capitalism is changing with technology and this is the next generation. So instead of me going to the library, asking the reference librarian who is the best, and fingers crossed that they got that right, and then finding the goddamn book in the stacks, AI will do all that work for you, and you can double check. You can ask AI, Are you sure this is the best person? Who else would you listen to? And it’s just instantaneous.
Here’s another example: You tell it, I’m a 72-year-old guy and I feel like I’m in good health and you give all the facts of your life, and you can ask, So how can I be more fulfilled at this stage of my life? You can say, Give me a philosophical answer, and give me the different five variations.
Have you done that?
Oh, yeah.
So instead of you going and looking in a book or an article for the answer, you’ve asked a very direct personal question and it’s going to give you the science behind the answer to that question. And I usually say, Give me evidence. How do you know this? So I could say, Give me a philosophical answer, then give me a neuroscience answer, give me a social science answer, and then take all those answers and combine them for me into what I should know about this. And then it’s up to me if I act on it or not.
So if you want to flourish in this next age, you’re going to have to be familiar with AI, collaborate with AI. If you ignore it, you’re missing out. You might choose to ignore it, but are you going to flourish as a professional or as a member of society? Probably not.
It’s like, okay, there’s cars. You can either get in the car and drive or you can walk, or you can just stay in your house.
But it can also make you lazier?
Some people will say, God, I’ve got to work a lot more on my cognitive ability because if I’m lazy and let this machine do all my thinking for me, I will lose that capacity. And you will.
But I think for many people AI will be a plus. And for some people, AI will definitely be a negative. But I’m going to embrace AI.
SUBSCRIBE TO SANTA FE MAGAZINE HERE!
PHOTO MH RUBIN
