Agency in the AI Era

So when you're interviewing for a job and everybody's good because of AI. how do I stand out and get a job that they still need to hire for? I'm fascinated by that Like if Claud is really good at running your company Claude's good at running every other company and there's no variation between them. And generically high quality with no variation means there's no bootes or competitive edge. I think humans who bring competitive edge to this one way or another, just by providing variation, if nothing else is a useful way to think about problems, right? You start to care a lot more about the taste of that person than you do about The entire organization built to deliver the product We're all a little afraid of AI. Even the people who love it are a little bit afraid. Some of us are afraid it'll take our jobs or make us dumber or change everything so fast, we'll never be able to catch up It doesn't help that most AI experts fall into one of two camps They're the Doomers or they're the Zealots. That's why I wanted to talk to Ethan Mllk. He's an AI expert who refreshingly is neither Ethan's a Wharton professor studying how AI, entrepreneurship, and innovation impact our work. He focuses on how employees actually use these tools rather than how they're theoretically supposed to improve our work His New York Times bestseller Co intelligence and his popular substack One useful Thing have become the go to resources for all those people trying to make sense of AI without losing their minds Ethan believes we have more agency over AI than we think. It's our experience, our taste. and our genuine points of view That are things that AI can never replace And that's important Because the choices we make about how we integrate AI personally and at work As a society will shape what it becomes If you like this episode, please remember to subscribe This. is a bit of optimism You are very popular right now What were you I know you teach entrepreneurship and other things. before AI was What were you teaching that was the hot thing? In addition to AI stuff, my other thing I really care about is games and education at scale, two related things. I've done a lot of work on using video games for teaching. So I was, as the brand says, not famous but known in that space of thinking about games and teaching at scale, how do we teach transformationally So that has been something they have worked on for a long time. and the AI work was sort of in parallel to that. Was it your own personal curiosity about AI that sucked you in? or was that you were forced to use AI in the work that you were doing because it had because it made it better. I went to business school and my PhD program at MIT, and I worked with the MIT media lab with their AI group. So Marvin Minsky was one of the fathers of the field. And I was the non technical guy in the group So I was the person who had to explain to other people what AI was, how it worked. when we'd go talk, I have technical conversations. I'd come along. I' been AI adjacent or involved for twenty years but always in the sort of non technical like how do we use this? How do we explain it Rle? And everybody else was technical because it wasn't really working out. Like AI had limited use cases. And so when GPT three came along and started to be make a splash and then chat GPT after that, I was sort of well positioned in this world of someone who'd been thinking about and explaining this for a while in a n technical world The reason I was really interested in talking to you because to be honest with you, I try not to have AI guests on And the reason is is because they generally come in one of two flavors, right? It's the greatest thing that ever happened And it's going to make the world a better place orr it's the worst that ever happened. We're all going to die. And the conversations are for obvious reasons, a little lopsided and one sided, and they usually have some vested interest in one opinion or the other. Y And the reason I wanted to talk to you to be honest with you is yours is just more practical. It's not it's the saavior, it's not, it's going to kill us all, but it's kind of like Well, it's here You know, kind of like the internet showed up What's the best way to use this thing And it's a little more down the middle Yes. It's nice to be in a place where being somewhat pragmatic makes you unusual. That's not usually the place where you get publicity for being the non bombastic version. It's true. But also the other problem with those two opinions is they tend to eat the world, right? If you think we're getting a machine gun who's going to save us all, then you know, all that matters is discussing that, right? justust like any other sort of religious belief. And if you believe that we're all going to die, how can you have a conversation about anything other than, you know, this is gonna to doom us all? And the fact is this is a general purpose technology It's gonna affect everything we do one way or another. It's worth spending some time on that. And some of those effects will be good, someome will be bad I do remember sort of the rise of the internet. I'm old enough to remember it And the conversations were somewhat similar maybe l less dramatic on either side. There was a lot of positive energy at that. Mly positive But there was lots of conversation. likeike I remember people who, you know, there were the zealots who believed that the internet and everything online was going to replace everything. You know what it is? It's the difference between eating to live and living to eat You know, and if you leave it to the technologists, they think all technology is living to eat. I mean, they literally made protein powder shakakes. I don't knowember was soy into your. I remember soy. I remember for people who thought eating was too annoying. whyy are we spending all our time worrying about eating when we could just get it taken care of, right? It's a different view of efficiency. But it's also a view that kind of blinds you to the fact that technology is the most human activity, right? We're making tools for ourselves, how we use them, how we adopt them, how we regulate them Those are going to have big influences. You know AI is more self directed than most technologies, but we still have A lot of agency over what happens next I mean, I know from my own experience, most people are misusing or underutilizing the technologies that are available to us And I'm already getting to the point where I'm turning on you know social media or listening to friends or reading an article And I'm already feeling overwhelmed By all of the advice of people telling me that I should be using it like this and I'd be setting up an agent to run my life and setting up an agent to do my marketing instead of an agent to do my finances, I'm already overwhelmed by all the advice of how I should even be prompting the machine to the point where I'm almost backing off and shutting down because one minute geminiz that you got to only use that one the next minute. you got to only use ChatyBT the nextute only gotta only use Claude all too much already and for anybody who's not predisposed to like Be all in reaction, I think, is pushing some people away, believe it or not Well I absolutely. I mean, look, I am a nerd of the old school. so I like getting into the details of stuff and partially like explaining them, which is I think part of why you have me here. But also it is overwhelming. I mean, part of what's actually interesting is AI's gotten easier, right? Not to be too, you know too evangelistic about it, but like it used to be that stuff mattered. L you had like all those little details, like prompt engineering mattered. So it mattered how I phrased things it matt. said You know, you are a physicist. It was better physics. If I said, think hard about this, that matter. If I offered bribes, that would matter We've been test all that, none of that matters anymore. The model's gotten good enough that if you're good at giving instructions Like a human, do humans, you probably do okay with this. And sudilarly, all the models are getting good quite quickly. So whether you open AI or ChauT, like ChuT or Gemini or Anthropics, they all three of those are pretty solid. There's like one or two hints I'd give people, but otherwises really just use it for stuff. and don't stress how you're using it because you'll figure it out. Also, when we talk about like AI is going to take away all of our jobs, and by the way, those are the technologists saying When One of the things that I also love is the technologists are also very fond of saying, you know, things like eighty percent of the jobs today didn't exist twenty years ago Which meanans it's fair to say that eighteen percent of the jobs in twenty years we can't even imagine. And to your point about prompting, it wasn't that long ago where the technologisties were telling us prrompting is going to be the thing and people were saying, I'm going to get a degree in prompting. and the technology got good enough in just a few minutes that that literally went away Yeah, and I think that that is part of what makes this kind of interesting, right is we can't imagine the jobs where we'll keep trying. and also not imagining the jobs does get nerve wracking when like you know, AI iss here. and it's like, what is the job? Well, I'm not one hundred percent sure. I mean, I think there's both legitimate reasons to worry about jobs. Th systems are very capable and very competident and they do change they impact real work. This is not electricity in the same way of like we need to figure out a way to harness or use this thing. There's a real impact. I've talked about this a little bit. I actually think that You know, nature pours a vacuum, markets correct. When we have a bubble The stock market will at some point, not of our choosing, correct itself. and all systems seek equilibrium at some point. I'm fond of saying that in the seventies and eighties robots started to enter factories and the blue collar world said, Hey, we're going to lose our jobs and the white collar world said, it's the future, baby. Progress stops for no one. Rkill You know? and, you know, the pendulum doth swing And because your plumber doesn't care about AI, the carpenter doesn't care about AI, the mechanic doesn't care about AI. The people who care about AI are the knowledge workers. Y. And you know It's the future, baby. Progress stops for no one. Rkilled, baby. Yeah, I mean, that's a really set. I mean there is a come upance, but there is another thing, right? Like if you look at the last three industrial revolutions, there's either two or three, I mean, that's all we've got, right? And the reason they worked out wasn't because the technology made everything great alone. It was because there was also labor fought against you know capital and you had a whole bunch of fights happening. Unionization is how the benefits got spread around, right? The technology doesn't naturally do that. What's really interesting is we're going have similar fight here, right? Like AI, we already have good data, right? AI is pretty good at being a doctor It's getting better at being a lawyer all the time. Like most lawyers I talk to see a trajectory where not too long from now, you will be able to get as good advice for not the most complicated issues from the AI. if you're not already, right? And there's ongoing debate about this. But the thing is lawyers are not going to be quiet about losing their jobs. And it turns out a lot of Congress is lawyers and a lot of people don't get any money are lawyers. And I'm willing to bet that you're going to see laws passed in every state that you need to have a lawyer officially, a human, sign off on so that even if they're worse than AI. right? Yeah. So part of this like there is a little bit of's your come up in white color workers but it's also like, no, no They've got the like there their are fields that are not going to go easily, right? And and you know, doctors, lawyers, it's going to be interesting that coders really do not have the protection that doctors or lawyers or actors or other kind of guilds or associations do. So we'll see a lot of conflict over these issues in the near future. Lobying for our own interests is not a new thing. So for example, for years There have been people who try to raise money for the government, you know, where we can get more income. The average lifespan of a dollar bill is about one year The average lifespan of a coin is something like thirty years. And so the proposal to move dollar bills to dollar coins because it would save the government, you know, untold billions per year in printing and all of the rest of it. And the reason we haven't done it is because of the ink and paper lobby, because they like those billions being spent on them to make new dollar bills every year And so we know We're used to doing things that are not to our advantage because of lobbyists. And this is just another thing where those who have access to power and those who have influence will protect their interests, as you said. And not necessarily badly, right? Do we want all lawyers out of a job?' an open question. A lot of people will be like, yes, yes, please. But like, you know, I mean, already actually there's some early evidence that the number of cases being submitted to judges is exploding exponentially where people are pleading their own cases with AI law. like How do you deal with that? It used to be that we had a filter, right? So the secondary thing is you had to convince a lawyer to take your case. And if they did a bad job, they would be punished. And if you got a one hundred page legal brief, a person wrote that. And that was an indicator of real effort. But now it is it. How do we deal with that kind of system? So we're going have all of these ripples and all kinds of areas that require policy changes, regulation changes, societal changes even if we don't have sort of apocalyptic AI event the way we were talking about earlier. So let's go from the theoretical to the practicals. And as we sort of said before, which is a lot of people who aren't technologists are underutilizing or misutilizing this remarkable technology. Some are still using it as a glorified Google. Google does search, this is more of answer machine, but still I mean, we all do that because it is a very simple and fantastic use case. the more nuanced and complex promptps I think I know I'm under utilizing. Like I ask for an answer even if it's something that has some depth. but I don't ask it to write a report or make an interactive, you know dashboard of the data that not I know that I'm not doing that, and I'm already ahead of some in terms of my utilization of the thing I have two questions. One is You're teaching students, you're teaching a lot of grad students at business school. Is it a false belief that just because people are younger, they're all in on this technology or are they also sort of bumbling and fumbling their way as they learn about it? So we actually have a paper on this. So I think what people think I hear this term in the internet, we're talking about digital native all the time. People talk about this, right? Like kids these days are good at using the internet. And indeed, like if you talk to somebody who grew up with TikTok, they will know all these intuitive ways that, you know, as an older person, you will not understand. like, o, that's cringe. There's a bunch of rules and slang and approaches that you want to take. And I think that that model kind of carries through that younger people get the technology it for A Like I talk to like a CHR own' like, the kids these days they're AI native. I they're not AI native. You're just talking to cllauded. They're conduits to cllauded. Like If you ask for a report, they'll give you a beautiful report. They have no idea what's in that report. How could they? They have no knowledge of this. They're just giving you what cllawed says And we actually found some evidence on this when we did a study at BCG at BosA Consultant Group. We found junior employees were often much worse at using AI. They seemed like they were using well, they were adopters. But how could they judge whether something was good or bad? I think this is a rare case where the more experienced you are and sometimes the older you are, the better you're going to be at using AI if you decide to use it. because you can intuitively grasp How do I give instructions? AI works enough like people that if you give an instruction and you're good at a field good, who would be like, no, no, I understand what went wrong there. You're thinking about this, you should be thinking about that. Even though AI is not a person that doesn't think. you'll know what kind of information to expect to get good answers. So I think actually experience really does matter. If you think about our education ser is built, we're actually schools are in chaos. I mean they're always a chaos, right? Universities have always been in chaos. This is not new. You know P peopleople are chich A We actually know the pathway forward, right? Like it's going to be a little bit messy. it's like, we'll in more we did this with cal We' do more in class assignments. Outside of class we use AI tutors, which turn try out to be very effective at controlled experiments. We're makeaking them work better than we do now. In class, we' be active learning. We'll figure it out. I'm not worried we can' figure it out I am worried about the next stage. I teach people who generalists at Warton and they become a specialists. the same way we've taught specialist for four thousand years, which is apprenticeship Right I send them off to work for whomever, right? They go, you know, they go work at, you know, Bank of America or whatever. and they learn the job and everyone gets a good deal They get a little bit of income, but not as much as they would probably deserve. But they get a chance to prove themselves and they learn the ropes by doing grunt work over and over again. The middle manager assigns them grunt work that they don't want to do anymore and gets to evaluate whether this person iss any good or bad for moving up the ladder. And it's been a great mechanism. And that just broke, right? Because every junior person knows less than Chat GPT and they would rather just use ChatTPT and they be kind of dumb not to use ChatTPT or Claw to give you answers because it's better than what they could do And every middle manager would rather delegate to the AI than a flawed human who takes forever to give them an answer, and it is good. And so every're just doing AI work to each other And I think that that is exactly the problem that you're talking about here, which is the danger is that we lose the talent pipeline There are solutions to it, but they are going to require fairly radical change in how we think about talent pipelines of this is kind of like art. and here's where my brain is going, right? which is I am an art fanatic. It's the thing that I love more than most things and I am totally fine with AI making art. It doesn't bother me. I'm totally fine with AI making music. It doesn't bother me. However, when I hang something on my wall I like knowing that a person conceived of it I like knowing that a person made it because when I buy a piece of art, I'm not just buying the visual thing on the wall. I'm buying the story that goes along with it. Or for example, I was listening to some music this weekend. I was listening to John Bapiste's Beethoven Blue's album which is, if you haven't heard it spectacular. Now, could AI make a blues version of a Beethoven sonata? one hundred percent, it could. But the joy that I got from listening to that music was not just the music that I was listening to, but I was smiling that a person had the creativity to come up with this, and that was part of my joy. When we look at the work product You know, there's two things we're neglecting, which is I like thinking I enjoy debate. I enjoy making my head hurt at difficult things. I enjoy learning The same way a painter likes painting and a musician likes playing music and composing, where is the human desire to want to learn And then will our schools, but especially Our places of work allow for that to happen, where they all become so obsessed with efficiency that we actually, even if we want to learn, you see where I'm going with this. There's ten million directions from here, right? So I want to put a pin on the art thing because it's actually really important and interesting, which is, you know, obviously there might rise of our more artistanal human made things.'s the most direct version of this, by the way, is when you have AI in poetry or long form fiction There's often a lot of things wrong with it. because we're used to if we read something that reads beautifully and it's effortful to read, we assume that there's a purpose behind it. So we spend our own effort figuring out the holes.ike for example, the Aes are very famous at weird analogies, right? So it might say this conversation is like a gap tooth smile. Now that is not meaningful If you spend some time thinking about, you're like, o, how is it like And you will reach a feeling of meaning. Right? If that was, you know, Lasdo Krzi or someone else writing this set of stuff, and I was reading it and I'd be like, oh, this person thought hard about that analogy and I just spend the work to do it If it's the AI doing it You know, in some ways, it's beautiful, but the meaning come from me? and am I being cheated because I have to create the meaning That has no intention. The intention The intention no longer belongs to the artist. The intention now is shifted to the listener viewer. Right? And mayaybe it always has deinitely the novel stuff But like so that's one angle that's kind of interesting, right? And then I think the sort of second one is, you know is I'm thinking about developing some of these kind of, you know, how do you develop intuition? And we actually have a way of doing that. We know how to train people We can teach people to be experts, but the problem is it's effortful And so you you're like There's always been this sort of view that I told you early on, I made games for education. And one of the most depressing things you learn is you could make something incredibly fun, but first of all, it's only eighty percent fun. It's not as fun as actually doing a thing for fun. And second of all, learning is effortful And if you're not doing effortful work, then you're in trouble. And for the few areas that we intrinsically care about, for you, you know, it might be art history or music or maybe for some people, it's math and science, maybe for some people, you know, it's a sport that they care about. whatever we are effortful about And intrinsically motivated, you're like, why isn't all learning like? And the problem is you don't care enough about it.. so but I still want you to learn math, even though you don't want to learn math. I still want you to learn American history. And if you shortcut that through AI giving the answer is you learn nothing. We haveve enough experiments to show that. So making people essentially lift mental weights becomes the problem in a world where there are shortcuts. You know, I find there's a great irony in all of this, which is the problem actually doesn't lie with AI which is we've been on this steady drumbeat, this path to this point where We are you know, discomfort avoidant, the concept of ghosting is a thing where you just avoid a difficult conversation or you see it now particularly among young people where they're more comfortable with quitting a job than having a difficult conversation or getting negative feedback. And then the idea that we've become so end result oriented. You know as capitalism has become short term focused and more focused on a shareholder supremacy, shareholder value over the quality of the product or customer satisfaction or employee satisfaction You start to see we become more results oriented and we've left out the work product. This is not a new concept. AI is just the most exaggerated form being results obsessed at the expense of the effort, the work or the journey to get there Well, and just to take another path from that, I mean, part of this is what makes AI at work so challenging, right? Because if you want productivity gains, just youll get a one hundred times more powerowerPoint R me. My job is producing PowerPoint Like so it requires you to rethink what the work is. And so what the work product is can't be the same thing. I mean, even the most basic way, coders can write a one hundred times more code than they could before. If they're embedded in an organizational process where it takes two weeks to do a product sprint as they often call them, right. So each there's stand up meetings every day and the assumption is the coder will write X number of work, the product manager will do this, the designer will do this, the marketing people will do this and suddenly one person iss one hundred times more productive, what does that even mean for an organization becomes a problem Part of this is our systems, coming back to that theme we've been developing throughout, which is human systems are not built for an AI world. effort know school wasn't built for a place where anyone could write your essays, right? know work wasn't built for people being able to prodduce PowerPoint on demand without thinking about it, right? Fiction wasn't built that I could write as many papers The law clerks and courts weren't built for anyone to be able to bring up a case. That's not a problem. That happens at every industrial revolution. It's just all happening at once everywhere. And sometimes the AI wins, sometimes human systems win, sometimes we both lose. But that's where I'd watch the adjustment happening. I'm going go backwards a little bit here Let's go back to practical. which Based on classroom and in the business world because I know you study that as well What are better simple ways that we could be using the technology available to us. Like I said, most people are miss or under utilizing the tool. and I'm overwhelmed by the people giving me advice as how I should be doing things and what I could be doing. You know, from from just a basic standpoint, how can somebody level up One to ten percent. You can get more than one to ten percent. I take no money for AI esisode. you don'tound like I don'ound like a shill, but you have to end paying twenty bucks a month to one of the big three companies is I'd recommend, either Googles, Gemini, openpen AIs, Chat GPT, or Anthropics Claud. And you have to actively pick the best model available at that point, which is you'll have access to what are called thinking models and those will change over time, but you have to actively select that. It default to a lower one. You will get huge impact improvements just from picking the most recent model and using it The second thing I would say is AI has gotten quite good There's kind of three phases of AI. There's prior to chat GPT where mostly we talked about AI we' be talking about how you use you know, data analysis basically. There wass all this about algorithmic fairness and price mining and all like you know customized pricing, all of that came from prior to chat GPT. Then ChatTP kicked off generative AI and what I will grandiously call because the of my book, my previous book, co intelligence, where you'd work back and forth at the chat bot to get answer, right? Id typed through the chat, would give me answer Now we're in a new phase which is called agentic AI. and it's really just three or four months old practice. What does that mean? What does that mean? An agent is an AI system that can independently go do work if you ask it to. So Aentic is just the adject of agent Agent, right? So Aentic is an AI agent and there's marketing terms around it, but it's an AI that can do work The most important thing to realize is how good the work is and how long the work is. There's this paper and test by OpenAI, so you always take with a grain of salt, but there's been an independent enough assessment and I feel good about it called GDPValve. And what they did was they took People representing five percent of the U.S economy, so journalists and product managers and lawyers and private investigators. And with an average of fourteen years of experience, they had them each create a really hard problem that they face in their field. They hred another set of people with fourteen years of experience to do it. It took them on average seven or eight hours to do the work. And then the AI do the same thing took about fifteen minutes for the AI. Then they had a third set of experts come in and spent an hour evaluating the outputs from each of these, not knowing who's is who's and voting on which they liked better And when this came out a year ago, the best AIs in the world were getting about forty eight percent, forty percent of the time they were tying or beating humans. The latest models as of when we're recording this are about eighty four percent. So eighty four percent of the time, the work that they do, seven hours of human work are equivalent two or better than a human. What that means, going back to the practical piece is you would probably save three times effort and three times cost. If every complex job, you would give it to AI And even if it took you an hour to put it together and evaluate it, even if you had to give up thirty percent of the time, you would still save time and effort. So one of the things I think you're doing is not using AI and giving it hard enough tasks to do. Okay And so that's the other thing, which is the value of the AI, where efficiency is not how quickly it can solve the problem, but how quickly you can evaluate. whether it got the problem right whichich again, brings back to expertise. An expert can look at this right away and be like Not just' wrong, but like often it's wrong because of a specific problem that you should have either specified better or the AI is stupid about something. And sometimes you can instantly get, Oh,'s never going to get this because it's too subtle and I can'tunicate the point. I'm just going to do this myself. But sometimes you're like, oh yeah yeah, this is a rookie mistake. and I should remind it that when it writes articles to not just factually explain everything, but explain it with a story or whatever your thing is, and then it's better, right? So evaluation, feedback, these are things experts are good at, and the AI responds really well to that. Here's the other problem, which is I remember when I wrote my first book Everybody told me, everybody in the publishing world said the most difficult thing for any author is quote unquote to find their voice. Right to have a voice. Now it's a very hard concept to understand. You know, what voice is, Essentially, it's when you read my words, They are of me. they're my personality, they're my point of view. It's not just nicely written but it is It is of me Right? And it's very hard to do for an author and I have found that AI can write beautifully, but it has no voice. And if you ask it to have a voice, it's going to always have voices that are available to it in the world, in other words, published people but not you. And so most writing will start to just sound The same. I mean, I'm already seeing it. I'm getting AI generated emails in my inbox And they're all basically the same email. It's not X' Y, It's doing the heavy lifting here. The thing that keeps me up at night, you know this is a load bearing argument. The Stacado three sent is you know, word, word, word I'm started to just delete them all because they're all familiar. Yes. It's Oone of them stand out. Right. And I push back, I'd say, it's not that it doesn't have a voice, it has a voice Right? a singular voice that is has voice or Ch. And it's actually not a bad boy. Like if I was get if I didn't see it for but it' not your voice. It's of voice. And it's a perfectly good voice, right? It's a little dramatic. Its sometimes is like it loves transitions too much. obbviously loves, you know, M dash is too much, but it's not your voice. And that is another thing that is like Developing your voice. Now, a lot of people can't, right? Not everyone's a good writer. No matter how much we teach them writing, they don't get it. Ghost writers have been around forever, right? I'm glad you know that writing is something I do, and I have established voice. I know plenty of people who use ghost wrriters to do their kind of work I agree on the AI voice. Now I will say you can get it significantly more like you, notot for the kind of long form work of a book A tip here, if you want to do this is give AI a large sample of your writing And then say write two pages summarizing the style of this the instructions of how to write in the style And then you paste that into your custom instruction and you say write in this style It will not it will be slight parody of you, but it will be infinitely better than if you just say, you know, write like this famous person U I mean, I did something recently as an experiment, which is I walked around the living room just talking into Claude. and said write an op ed in style of Simon Synic. Here's the idea And I just walked around the living room for aboutbout three or four minutes. And then gave me an pretty remarkably written article. And thenen I said, fact check it. And it said, Well, that's wrong, that's wrong, that'song that's wrong. I said, OkayK, go offer me what I could say to make it factually correct. The thing that I enjoyed about doing it, which is You know, It takes eighty percent of my time. to make a shitty first draft editing is reasonably efficient and a lot more fun to really just clean something up Most of the time is the first draft, right? And so here I got a shitty first draft in a few minutes And then I sat down and with it, you know it factat corrected, which is so efficient. I didn't have to go do all the research myself. although I did double check all the research just to be sure. As point, the error rates of these things would drop. if you use a modn model, it not making mistakes way It's not making mistakes the same way. I have to say, it was actually kind of fun to edit it in my voice with my sense of humor and the last finishing touches, I realized I could put my voice. I was pretty impressed. I was pretty impressed. Now I could cheat because I have written enough that can know my style. and I wouldn't say it was perfect, but it was It was scary good Yeah. I mean, there's a few things going in there. One of those is This idea of disruption to writing. and I mean, there's costs to everything, right? So one option of writing the crappy first draft is it's your crappy first draft. So I always recommend some crappy first draft because otherwise, the AI's ideas will take over your ideas. It's very good at ideas and like you will find you can't brainstorm. With that said, I find this kind of similar loop of like editing is a weird way of approaching a it's not how we used to do it before, which is like I can get something writtich in my form and then I edit it. You know, I'm sure some writers have worked that way for years, but that's a disruption to writing might be better, right? It might be where it's hard to know. suuddenly you're factory, you know, producing first drafts that you too all the time. and or maybe some people are just really good at editing and they weren't good at draft writing, and suddenly they're more productive than they were before I think this is one of the future jobs that we underappreciate, which is not just that jobs are new, but that the weight of the job will shift. Yes. too your point you know, we've always celebrated the writer and editors have always been like there If you work for public relations or you work in magazines, like the person who's the writer who wrote the press release, they're the person who went to school to write the press release and we just sort of like the editors are just the lower paid, you know, you know, failed writers, you know, quote unquote. But now I think the writers, I think that the balance will shift. I think we're going to take the cross of lots of jobs, by the way, to come back to the job thing. So jobs are many tasks, right? A writer does like as a writer, you're in charge of writing and editing and fact checking all. And the AI does some of that work. It shifts the burden of what you do But it doesn't take away everything. And I think what we're going to see in a lot of jobs is this idea of bottlencks. that the AI is good at some stuff, but bad at other stuff. We call the jagged frontier of AI in our early papers on this, which is it's good at some things, bad at some things you wouldn't expect. Wh it's bad, right? writing perfectly in your voice, getting a joke right. L suddenly, the demand for your labor is higher there Right? And your value was higher. It might have been that your jokes were not what was getting you. like that was not your main deciding factor. But if you're better at jokes now suddenly there's value. Same thing happen in coding, by the way. It used to be that writing really clean code was a really good skill Now the AIs write most of the code, being an architect is good, being an engineer manager is good. The jobs change, what's important and what isn't important changes. And that changes who's good or bad at it too, creating new opportunities and new risks. You may have, instead of one hundred coders, you might have fifty or thirty working on the team, but they're still human beings with egos in securities you know, lack of sleep, all of this stuff. and there's still somebody overseeing the project who has to manage all the messy human stuff, regardless of how good the technology is. I for one believe that doubling down on human is going to become even more important now becausecause we still have to take care of the people who are working on the products with their AI agents Oh abbsolutely. And also when I went something coded in the past, I had to hire a company to do it. Now there might be a coder working for every two person team and more software is being created than ever, right? So Jobs are unimaginable in the future, sort of annoying thing to say. I think it is annoying because I think we actually have some idea of what this looks like, which is not that coders are replaced by know prompt engineers, but that the job of coding changes. The demand for coding shifts from giant organizations where a thousand people work together programming to now dispersed and your car dealership might have a coder building customized software for you. around what the managers want to the team. You might have two developers working for you rather than outsourcing web development that are evolving things, the nature of software and the jobs change. And I think that that is a missing piece of this puzzle also. And I think the other thing we aren't appreciating, which is the more things get good Right? Be it used to be that quality would help you stand out. that if you were smarter, a better coder, a better writer, a better this, a better that, whatever it was, being good at something made you stand out from the crowd, right? If the quality of Let's just say everything gets slightly higher or a lot higher then commoditizes so many products. And so what I'm curious about and cannot predict and don't even have a thought about what happens here Everything just becomes generically good Then how do you stand out in a market now? And we've kind of seen this with the rise of social media. We're in the last generation that has movie stars. I think it's the death of the movie star. Nobody's really buying a ticket to go see a movie because a particular actor is in it. like one battle after another, you know, lots of people went to see the movie, very few went to see it because Leonardo DiCaprio was in it You know, and that's what the movie stars used to do. They used make people go see the movie Now, we'd rather see the franchise when we're interested in Marvel than who's in it. And this is what I mean by commoditization. I'm so curious as everything becomes better and commoditized TV channels. everything's commoditized What's the thing that makes companies, products and people St out So when you're interterviewing for a job and everybody's good because of AI. how do I stand out and get a job that they still need to hire for? I'm fascinated by that I think a few things there's a lot of things there again, right? Part of this, by the way is writ large, right? Like if Claude is really good at running your company, Claude's good at running every other company and there's no variation between them. And generically high quality with no variation means there's no bootes or competitive edge. I think humans who bring competitive edge to this one way or another, just by providing variation if nothing else is a useful way to think about problems, right? Like your sense of taste matters, right? And presumably know why people list Your sense of taste, who to talk to the kinds of questions you ask, you know is similar to the sense like, do you like Rothcoo or do you like Rembrandt? Like there's different tastes that have different kinds of outcomes. Yeah. The second thing is developing taste, right? is a bigger issue, which is how do we get people to develop taste? It's usually a casual lifetime thing. That might be one of the new talents we teach people is developing a sense of taste, which requires, you know, experiencing broad things and making choices having the vocabulary to describe your sense of taste and choices. I think that As people become bigger creators and they can do more, their taste matters more. Like directors m up mattering more than ever because I understand what I'm getting with a Wes Anderson experience, right? And if he can direct a whole thing the way you wanted to do, what would that look like? Yeah We might find the same kind of thing with all kinds of other stuff. There's someone who has a particular taste in ice cream styles. Now you can make ice cream on demand because the will connect you through the APIs to a you know, to a vendor that makes that product for you. So it kind of fits in of enabling one person to do much more. You start to care a lot more about the taste of that person than you do about the entire organization built to deliver the product. So good I mean, I have my own biases and opinions, but I'm curious, is there actually a difference between Gemini Chat GPT and Claude? I know Claude has a much more B to B focus business model That means security is more of a thing because business wouldn't stand for you any lapses in security maybe like customers might. Is there actually a difference? So there are. J a half step back on the boring educational side of this. When you think about AI now, you want to think about three things, the model, which is the brains of the bunch, right? At the time we're recording this, that's Opus four seven from Anthropics Claud model, ChatTPD five point five and Gemini three onene proro. By the time you hear this, there will be slightly higher numbers on all of those things based on how things are going, right? But those are the brains, right? The better your AI model is the smarter is at everything. It's better negotiations, it's better at poetry, it's better at math, it's better But that's the brains. Then you want to consider apps. Apps are the tools you access these. For most people, when I say apps But they should be thinking of as chatBt. com or cllaud. ai or gemini. google d. com. That is an app But the apps that people increasly talk about when they use AI are things like clawed code, open AIs, codex, notebook LM, which you haven't used for Gemini is free and very impressive for research and gathering data. And those are very specific tools built for specific purposes. And then finally, there's what we call harnesses, which are how the AI can do things, right? So a harness lets the AI write code or do internet searches or make images for you So right now, the three big companies all have roughly equally good. probably you know jockey from position. They're all making very good brains. The models are all very good. Right now Google has the most diverse set of products of apps. Their main apps are probably weaker than anthropic or open A and they have worse harnesses for the main app. So if you want to use AI to do things, right now, the most powerful tools are Clawed coder coworer on your machine if you're using Anthropic open AI' codex tool. And what makes those different is they sit they use your computer So like you can give it access to your files, to your email, can, you know, it can do work for you using your machine, your web browser, whether you like this or not, right? And do work. So because of that, those two are kind of jockey back and forth for the lead, but all three of them are quite good. The models are good across all of them. So now let's talk about security, right? So we're all tired of Ma and all the other companies filling our computers with cookies tracking our every movement on every website even after we've left their website and their product. You know, we've all become very sensitive to turning off cookies and data privacy is now a thing You know, do I want to give any of these AI models? Do I trust any of these companies to have access to all my computer, all my browse history, all my finces eter, etera eter So It's a hard question, right? I mean, there are more secure versions where you can even run your own version of these tools, but they will not be as good as openp AI, Anthropic, and Google's There are a couple kinds of security concerns you might have. One of them is are they taking your data and using it to train their next model If you pay twenty bucks, all of them have an option to turn off that training Is that enough privacy for you? It's hard to know, right? There's open questions about whether or not someone's AI history will be searchable. Is you is it something that lawyers can demand to look at it, right? Discoverable? There's open questions about, you know, what will companies do with this in the long term, even though they sign agreements with you. But on the other hand, Gmail probably has all your email in it, right? Like these look like enterprise software applications at this point rather than sort of invasive individual tools. So do you how much do you trust Google with your information or Instagram with your information? We're in that same kind of boat over again. The difference is As much as I don't want Google to have access to my Google my email. I know that it does I know that nobody can go out onto the web and ask a query in a Google search to read my email and tell me something. I think a lot of us are afraid that somebody could just go onto chat GPT or open you know one other They can't do that They are not no it's just like Gmail in that way, right? There's no bleed over where there's just one giant inbox and you're just barely holding it together, right? It works the same way It looks like enterprise software. So that has its own risk, right? But the basic risk of like, can someone just ask for something and get access to your chat GVT? No, if they log in with it, you know, you have to do all the same things you do, set of factor authentication. Don't leave yourself logged on to a computer. but it works. the analgey I would have is Gmail, right? G Like Google has all this information there obviously processing it and using it for their own purposes, but they're also not going to, you know, they've anonymized it in some way to try and create trust It takes effort to hack into someone's email. It's the same kind of boat, right? Now, whether or not we want you want a company to have even more power over you, those are choices you get to make. But I don't think we should put this in a separate privacy category. The actual risk is if I let it have access to my computer and it could use my web browser, you, could someone convince my AI to send them all my money you know, if it's reading all my emails And that is, you know, hasn't happened yet, but's not impossible. Right. So obviously because you are you know, you teach this, you embrace this You allow your students to use it, I assume I don't know how to ask this, which is how do you ensure that your students are learning? So they are allowed to use these tools to learn? So I went viral first in education with my syllabus right after Chat GBT came at the first version, which is what we call GP three point five you know, that was around for a few months. GBD three point five was pretty flawed. Like it would make up arguments all the time. It would obviously hallucinate. It felt like a you know, like a smart, you know ninth grader or something like that. right? And so I teach college courses, I could tell. So my original policy was use everything you want you You're accountable for the outut That was great for four months until GPD four came along which is now obsolete, wast good for a while. And it was as good as my students across some things, not across all things, but enough that a low effort student was worse than GP four. And I can no longer tell people, just use AI. I can tell because theI was giving them the answers not being the answer We've seen this over and over. There's a lot of studies that show if you' just use ChatTPT to get answers to questions, you think you're learning, even if you're not cheating, you think you're learning and you're not learning because the AI gives you the result It turns out we actually know pedagogically how to solve this problem. We did this with calculators, right? In school, which is we could do in class testing. We can make you use the AI for some stuff enough for oth. So for my classes, I'm luck to of teach entrepreneurship. Output is in some cases, like I gave all my students, for example, what I called the Voit compomp test, which is the name of the Blade Runner human test, but I made my own version of it. And they had to launch their startups using AI, but based around areas they were experts in, experiences they had had, knowledge of the world they had had, a viewpoint they had, which kept them in the picture And then they also had to do a lot of in class stuff, right? We had to have a discussion about these things. I actually had them use AI tutors that ask them questions. They had to use an AI to build a case study with and I set up the AI, so it wouldn't give them all the answers. It would challenge them to come with the case study information. So there's things we can do, but it does require changing how we teach. But the reality is technology does affect our brains So I'll give a real life one to one example, right My my mind, I used to have a steel trap for phone numbers. I knew everybody's phone number. You told me gave me a name. I'll told you their phone And I didn't have to memorize it. I just heard phone number, and I had to steal trapper phone. was just how my brain worked And I in the early days of I bought a Casio digital diary I got it for my birthday. I remember You know, it had two K of memory. I think I upgraded to the six K when it came out. It was like hardcore, right And it was the most remarkable thing and I programmed all the phone numbers from my memory into the device and then slowly added more and more phone numbers as I learned them. And my brain was like, Okay, if that's what you want, fine I can't remember a single phone number anymore. And if we have to remember that they're like the Iliad and the Odyssey were oral traditions. You know, this book that we were forced to read in school that's like in eight hundred pages, you know, go back you a couple hundred years and it was like, son, it's time I tell you the story of the Iliad and you will tell your son the story of the Iliad. It was oral traditions that people remembered But because the printing press, our brains just stopped remembering stuff. So this has to have an impact On our intelligence, there's no getting around it I mean absolutely I mean, look, my my grandfather was an engineer who built like the fire suppression systems for Cape Canaveral. and his like dissertation was wring doing a single piece of matrix multiplication I have no idea how to do what he just did and he did slid rules to do it. I have no idea to use a slide rule. My kids have not learned cursive, right? Like we give up stuff all the time. The whole idea of technology is on purpose, we give up things that we used to be able to do. Yes, machines, so we don't have to do them anymore Eree Every time we face the same choice about what's valuable and what's not. And what I worry about like the default version of that is bad, right? I mean, we've seen this happening with like, you know, you could argue short forerm video has killed reading because it's more entertaining to do that. and I don't need to spend the effort reading the book to get there. Okay, that was a bad choice. We are going to have a ton of these choices right around AI. It doesn't hurt your brain, but it is a choice that you can hurt your brain with, right? As an educator, part of my job is to get around that problem anyway. Yeah. People can survive a lot without reading very well. They can survive pretty well without doing math. They don't have to learn American history. there's some degree of making this a requirement. But this is a slippery slope, right? Because now now we go down the path of oh, you don't need university. and there's a whole movement that you don't need to go to college. And what we forget is you may not need the subjects that you learn at college, but going to higher ed teaches you to think critically. It teaches you to argue with people who have way more education than you and form strong arguments to take them on It also teaches you adulting. And so my problem isn't that technology replaces that their sacrifices. like I accept that I don't have to have a memory for phone numbers because of technology anymore. I accept that My concern is that thinking The ability to think is the sacrifice here And that's way more damaging than remembering phone numbers or, you know, remembering the liad. So I push back. I don't think it destroys your ability to think. I mean, I think for a lot of people, it gives them even more ability because they have a conversation partner at their level who's willing to discuss a topic. As long as people have any curiosity about the world, Right? All of this is prosthesis for the g. I mean books for prostes.ike someone else came up with an argument for me an opinion. That doesn't mean that there aren't negative effects on that. It doesn't mean we won't give up things we shouldn't give up partart of what What heens me is like we've got twelve to sixteen years of know schooling to try and get some of this right. And if we do it right, AI accelerates some of that. And if we get to make choices as a society, now, will people make bad choices? Yes. And so I do worry about this, right? I've been thinking a lot about what do we give up? How do we stay human? It's going to require effort, just like a lot of other things And there is danger there. but I guess I feel like feels like a big leap to we're not going to think anymore. The AI will tell us what to do. We'll just obey its instructions. I don't think it will stop thinking. I think it'll hurt thinking. Like the quality of thinking, critical thinking gets hurt. And I mean, look, you know this as somebody who studies education. You talk to any college professor and they'll tell you, forget about AI. Just the introduction and distractability of a phone You know, that they'll say that, you know the writing is abysmal these days. Every college professor is complaining about the writing being abysmal. So kids don't know how to write. And when I say write, I don't mean like But I mean, form an argument. and they'll say like the first paragraph was fantastic. secondcond paragraph was fantastic. The third paragraph was fantastic. The problem is the paragraphs have nothing to do with each other because it's clear that they're like getting distracted in between paragraphs. I guess I would say on one hand, you're right, but we've been very bad at education for a long time. Like a bad way to teach is stage on a stage where I go up and give a lecture Right And a hundred people write things down. But we've done it for a couple thousand years because there's a lot of other constraints that make it the way we do work. There is a negative side, but like one of the things that really excites me is AI tutoring. We have some early evidence that has big effects on learning Right. Like instead of me lecturing into a classroom and assuming the height of learning is, I lecture to what the upper part of the classroom people who really know if the middle, you know, I lecture to the person who doesn't know things as much personersalized education is now an actual possibility. Like there is a cure as well as a poison in this thing. And I think that it's worth paying attention to both. If we don't change anything, the effects will be bad on education Right. But that implies that we're all going to sit down and just be like, I guess it's done, you know, like ye Yeahah. And I think for the first time, as opposed a short form video where you had to do this very elaborate thing of like, we'll do TikToks for education and that never works. We actually have a tool that is a pretty good tutor that can talk to you at your level. that can make you get into an argument. That's part of what I do in my classes And we see schools adapting, right? First, they put computers in all the schools and now they're slowly taking them out And it was partially because we just like comes back to the human thing. It's complic like the thing that makes AI interesting is itstands understands in quotes, right For those who are just listening to this, I'm making air quotes in my hands. It understands humans. It has theory of mind effectively And that's what the other technologies don't. Like it can teach to your level. It can understand what you're confused by. It can help you make this interesting for you. If your only interest is basketball or you know basket weaving, it can give you basketball and basket weaving analogies and problems This was the holy grail of education. And I think it's one thing to say, yeah, technologies have lit risk and everything else. I also think we usell some of the impositive impact that we can get from this. The strong argument that you're making here is, and I don't know how many people have done this, which is where you use the talk function where you can actually have a conversation backwards and forwards with the AI as opposed to typing I think the case you're making is the idea is that you can debate with someone at your level. So you're not explaining to somebody who's not at your level. You're not feeling dumb or trying to keep up with somebody who's more experienced or smarter than you, but rather that you can go backwards and forwards and learn the way you like to learn And I've tried this where I'm having a debate or conversation backwards and forwards, backwards and forwards, backwards and forwards, backwards and forwards. And I'll say things like, oh, wait, is what you're telling me this But I think this, that I think is really, really interesting to your point. There's two other tricks there. One is the AI is sycophantic. So if you're having a debate with it's going to agree with you. So you have to to act like a critic, right? L. And then the second is you want to take advantage of the meta piece also, the learning piece of saying, actually halfway through, tell me what I'm doing wrong with my arguments. How can I be more persuasive? What patterns are I missing in discussion Give me some examples of those patterns and how I could have used them. So again, pack the effort piece. If you're willing to do the lifting yourself of asking the questions, like everyone always says, they want to come to office hours and have this debate with professors,' most people don't come to office hours. You sit as a professor and you're waiting for somebody to come and debate to you in the great issues of the day and you sit alone in your office during the hour the people are allowed to come to you because they' other things to do I think that we overestimate how this sort of shining city in a hill where we would sit down and debate and have these discussions. like that's not how most things work. Now we have a tool that can do that. If you're interested, can do that without having its come to my office hours. I want to double click on the two points you made because I think they're really valuable, which is remember that the AI is a sycophant and you've got to tell it to criticize or critique and Ask it to evaluate your thinking and help make your thinking stronger. Those are two brilliant, brilliant prompts that I think more of us should remember to improve the quality of our interaction with the technology. We were talking earli about AI in writing. A piece that was missing in your conversation. You talked about using your fact checking.'s very good at that. I would use it more for initial research. All of the AI models have a deep research mode that's quite good and we'll actually do research for you. But the thing you're also missing from that is when I write something, I have the AI evaluate from different perspectives. So I will have the AI read it through as a reader who doesn't understand much about this topic and tell me what I need to change read this through as an expert who, is out to get me on social media, whereere would they nitpick my arguments Right? Am I being irresponsible anywhere? Did my humor fall flat? So giving the AI, the personas don't change the AI's ability. So say you're good at physics doesn't make a good at physics, saying you're physist, but it does make it talk like a physicist, right or parative physicist Talk like a cynic, talkal like a critic, talkal like a naive person. You will get answers you couldn't get without going through a wide range of readers. Also true in entrepreneurship, by the way, get feedback from the AI in different personas about your idea. This is very practical and very good. are you actually afraid of I it think we're in for a period of chaos, right? let's say the industrial revolution works out like the last three did, like the AI revolution Living through it still sucks, right? Like Charles Dickens is basically just a story about how miserable the industrial revolution was, right. you have haves and have nots, you have social change, even if everything works out fine. Now we have better tools as society, but I don't see a lot of action. You've led this conversation by saying People either doom and gloom you know or everything's going be great. I find policy making is in the same place right now. Either it's all going to work out great or we have to stop this whole thing And Neer of those are realistic outcomes. How do we cushion people if they're uninsured. turns out training programs for new jobs never really work. Is there something we could do better this time around to you know reskill people? We're going to have negative effects on inform on, you know, deep fakes are going to be everywhere. How do we deal with whoo we trust for information. There's a thousand little good and bad things that are gonna be happening all at the same time They're going to be very complicated. and they're going to get boiled down because of how social media and everything else works to either AI all bad, in which case you have a list of all these things that are mix of real things, you know and fake things about AI water use or whatever it is and it's going to be AI is bad or AI is great. And it is a thing. It's a technology and interacts with people You know, technologies are neither good nor bad, nor are they neutral. They have effects on our world And I worry that we're not taking this seriously. The other thing I worry about is people don't know how good these systems are better than you think, right? I have a doctorate. you know, I'm a professor for a while, I publish in journals like The AI writes a pretty damn good academic paper now, not just a parody of academic paper, if you give it a data set to work from It is proving math at a level where you really need to be one of the best math person in the world to know whether the system is right or wrong, it's often right at this point It is doing really good images and marketing work that beats most marketers in studies that we have of this. It be like these are really good systems. their development is not slowing down. We have to start thinking about what we want the world to look like rather than just assuming it's all either going to work out or not And how much agency do we as the general population have, or are we just the subjects? Are we just the pawns in this game between these three major companies, Microsoft, Open AI and Anthropic? So I think that We have there's two levels of agency we have. Level of agency number one is societal, right? Like there's a reason why people are floating you data center bands, right? Because they think that'll be popular. The usual mechanisms of policy making of organizing, of writing letters to your congresspeople, those still work The second is where I think there's even more agency, the A labs are full of coders and they have found an unreasonably effective way of making a tool that mimics human thought. Like It's weird the large language models work as well as they are. we know they work technically, but we don't know why this is so unreasonably good. Like how can it do poetry and offer know and you know interior decoration and write discount of cash flow analysis and you know a pitch deck about, you know, the Gettysburg address. Like it shouldn't be able to do these things. It does all of this stuff. So we give them too much credit. They don't actually know much about how AI is useful in non your field. Remember, there's a jaggon frontier's good at some stuff, bad at some stuff The biggest source of agency is actually using it positive use in your own job and work. Like a large part of what I post is like this is a way to help humans thrive with AI if we use it this way rather than just automating away human work. And I think our biggest sense of agency is, okay, you have access to these tools,, Simon, how do to use that to expand your business, to make sure that all the people who work for you have spoken to some amazing folks. They're all really smart. How do they do more than they did before? How do they do more satisfying jobs? There's a lot of agency there. And then if you talk about it through your platforms That changes things. And a lot of what I do is talk to executives and leaders and companies where I'm like, have you have to show people how augmentation, how this can be used to make humans thrive, how can make your business thrive rather than the default plan of like, if I fire everyone and replace them with AI, profits will be higher. Like that's the dangerous thing. So to me, the real agency right now is let's find positive examples and tons of them out there, use them and build them to make AI make the world a better place and not worse I really appreciate this. Like you've given me, you've enriched how I can use this product. I'm going to take you on. I'm going to have the agency that you recommend I think this is a moment for transformation Yeah And I think people aren't being ambitious enough. Everyone's like, what if I record my butook? likeike it's not, you know, how would you reach every one of your audience members separately if you could do that? And why don' you just build it? rather than waiting for it to happen? Like I'm not sure I'm going to use it that way because I like the artist I take pride in the fact that when somebody's talking to me that it is actually me, my opinions. I don't I don't think it's about automating Simon, like creating a Simon clone. I' never like that. Like there's people who createed Ethan Botts. don't I don't think that's the way as you're talking to You know a fake version, a parody of yourself. I am saying what do I want people to accomplish in this world? Like how do I build a tool for everybody Y that I believe in belie. And like I said, I would do assimon AI with a very specific application that it lives alongside. but I like people knowing that when they see me and they think it's me, it really is me. And I agree. I mean, it's the same my writing. I write all my own know my own Twitter posts and everything else You know, it's important to keep the muscles alive for nothing else Ethan, such a joy. Thank you so much for taking the time. I really appreciate it. Thank you, it's a pleasure As always, thanks for listening And if you liked this episode, please do remember to subscribe to a bit of Optimism wherever you enjoy listening to podcasts. And remember new episodes drop every Tuesday A bit of Optimism is a production of the optimism company lovingly produced by our team, Lindseay Garbinius, Phoebe Bradford, and Devin Johnson. And if you want more cool stuff or just to find out what I'm up to, visit Simoninic dot comot Until next time, take care of yourself, take care of each other.