Transcript of 698: Joanna Stern - How to Think for Yourself in the Age of AI, Why Deep Work Still Matters, Making Career Decisions With ChatGPT, and How to Stay Human on Purpose

Welcome to the Learning Leader Show presented by Insight Global. I am your host, Ryan Hawk. Thank you so much for being here. Go to learningleader.com for show notes of this and all podcast episodes. Go to learningleader.com. Now onto tonight's featured leader. Joanna Stern spent an entire year letting AI into every corner of her life, her health, her marriage, her kids, her therapy. And she even had an AI relationship on a burner phone in her nightstand drawer. Joanna is an award-winning journalist, a 12-year veteran of The Wall Street Journal, and the bestselling author of I Am Not a Robot: My Year Using AI to Do Almost Everything. During our conversation, we discuss why ChatGPT told her to quit her job at the Wall Street Journal, and then why she actually listened to the advice and did it. Then we discuss which AI tools will help you become a more effective leader. Actual practical use cases for people in leadership roles that you can implement right now. And then we talked about her conversation with Sam Altman, the CEO of OpenAI, that's ChatGPT, about his thoughts on AI in the future and how Joanna feels about that.

Ladies and gentlemen, please enjoy my conversation with Joanna Stern. This episode is brought to you by my friends at Insight Global. Insight Global is a staffing and professional services company dedicated to being the light to the world around them. If you need to hire one person, hire a team of people, or transform your business through talent or technical services, Insight Global's team of 30,000 people around the world have the hustle and grit to deliver. Hiring can be tough, but hiring the right person can be magic. Visit insightglobal.com/learningleader today to learn more. That's insightglobal.com/learningleader. I loved reading the dedication of your book to your mom and your dad, who taught me to think for myself and the AIs, robots, and machines that made me wonder if I really was What do you mean by that?

Well, when you, you get further into the book, you see me talk a lot about how AI might be doing a lot of the thinking for us, that we are outsourcing thinking in a lot of places. And wanted to dedicate this to my parents when I was really thinking about in the age of AI, what did my parents teach me and what can I teach my kids? It is to think for ourselves. You know, it is not to outsource the thinking. On everything we do. So that was, that was the, the inspiration there. Thank you for reading the dedication.

Of course.

Hopefully you got a little further, but even if you didn't, that's great.

Well, you know, we kind of start and then maybe we'll bounce around. Maybe I go to the very end because I also like some of the stuff in the acknowledgments. But I am curious about the part, because you're a parent, I'm a parent, the part about helping you think for yourself. I want to do that as a parent. How did your parents do that for you?

Well, I think for me, my parents always encouraged me to be me. I grew up, I was a big tomboy. I just was trying always to be my own person, and my parents just said, okay. They didn't try to limit that. I had a lot of my own different interests, some of them being technology. My parents said, okay, yeah, that's cool. We'll, we'll buy you some gaming systems. We'll buy a new computer for the family. And so I think encouraging our kids to be themselves and to think for themselves and to just be their own selves, not just to go along with the crowd, is a, is a huge part of parenting.

I love that. Another mentor that you mentioned is Walt Mossberg. You said, uh, your 12-year arc at the Journal, that he played a big role and was important in your life. What did you learn from Walt?

Well, first of all, Walt, made it possible for every tech journalist and frankly every tech creator out there today, whether you're a YouTuber or you're an Instagram or TikToker, whatever you do, Walt Mossberg invented this. He invented this idea that technology, which is made for humans, should be reviewed by humans, that we should be looking at this stuff. He famously said in his first column, computers are too hard to use and it's not your fault. I think without Walt Mossberg, we don't have an industry where we review products in, at least in the way of really thoughtful reviewing. I mean, maybe we would've still gotten Amazon reviews and no offense if you're an Amazon reviewer, but it's, you know, it's not, it's not always the most thoughtful review. I mean, look, you also know this, Ryan, as an author, sometimes you get some really thoughtful book reviews on Amazon. You also get some really not thoughtful book reviews on Amazon. So for me, yes, I could even talk about what Walt did for me personally, but just as an industry leader in journalism and in the technology industry, there's no one like Walt. What happened in technology journalism, what happened in tech reviewing doesn't happen if there wasn't a Walt.

I love hearing stories about people we look up to, especially as, as someone as accomplished, award-winning as you, who's built this amazing career, who's, who's currently in their prime. That you still have this kind of reverence for the people who've helped you, like your parents, like Walt, and there's many others that you mentioned. But I want to get to this idea of AI, because when I first saw this initially, Joanne, I was like, I don't want to talk about this. Like, I don't want to talk about how somebody outsourced or at least depended on AI for a lot. And then I actually got into the work and loved it and really wanted to talk to you about this because I have this— whether it's like anywhere online or LinkedIn or some of these places where everyone just copy and paste so much AI slop that it just makes me not want to talk about AI. However, I think the way that you looked at it is different and unique and useful for all of us. So can you talk to me about the reasoning behind this book, behind dedicating 365 days to solemnly swearing to live with the machines as you pledged and why you wanted to do this.

The reason I did this, I think, is actually in part, part of the reason you, you were skeptical of this, right? Is the AI slop we started seeing, you know, years ago now, 2, 3 years ago. And you had that, right? You had this kind of really distrust, like, why do we need this technology? And then you had the tech industry and these tech leaders in what I've now started calling hyperbole maxing of AI. Promise, dreams, hopes, and changing of the world. I mean, it was just insane what you heard from these leaders, and we still continue to hear from these leaders, right? AI is going to cure cancer. AI is gonna change education. We're all gonna have a personalized tutor for our kids. It is so miraculous, this technology. Yes, this technology also can harm us and kill us and end humankind, but it is also amazing. And so I don't wanna take a doomer approach. I didn't wanna take the zoomer approach. I just wanted to write for the normal person., you know, the normal person that has kids, as we've talked about, the normal person that has a job, the normal person that goes to the dentist's or the doctor's office, the normal person that drives the car.

How is AI gonna affect all these parts of your life? And that was the, the whole reason for doing the book was just, let's go on a journey together to find this out.

I'm gonna focus on healthcare because this was interesting to me and your mom, cancer survivor, and then you go in and I guess got a mammogram and ultrasound. Is that accurate? And you sit with your radiologist, Dr. Lori, who also then is using AI software. Can you talk to me about this moment and what you learned from AI and your actual doctor?

Yeah, this moment was really impactful for me, and it happened at the beginning of the year, and we didn't actually know we were going to structure the book. We thought, okay, maybe we're going to structure it by topic, like healthcare, education, etc., and then it sort of worked out like, oh, let's do it as the year progresses. And so at the beginning of the year, like I always get my annual mammogram and I decided, okay, I'm gonna try to put AI through all of my health decisions this year. That meant anytime I was sick or the kids were sick or the dog was sick, I was gonna ask AI. It meant I was gonna have it interpret test results. But I also wanted to go to the doctors who were using this. And radiology is the one that is really the most far ahead because of deep learning and what was happening years ago with taking huge data sets of imagery, in the case here, mammograms and ultrasounds, running those through models, being able to see— have the AI see at a pixel level in a way that humans can't see where there might be cancer, because there are millions of these images, more than any human radiologist could see in their lifetime.

And then they've got the test results, right? So they can put those things together into a model so the AI can see, oh, that looks like cancer. Oh, that looks benign. Don't worry about it. And so went out, went to this doctor who is now my real doctor, 'cause I just enjoyed working with her and seeing her so much work and watched her read my mammogram and breast ultrasound alongside AI using a tool. Now, this is not— I'm sure your listeners are smart enough to know that she's not just using ChatGPT, right? These are custom models that are made by healthcare companies. They're FDA approved. So there's a lot of rigor and backend to this. That's, that's not just, you know, off-the-shelf technology. And watching her work side by side with this was fascinating. What she originally saw, and what I describe in detail in the book, is there were 3 spots that looked suspicious. The AI flagged as suspicious, and 2 of those spots she didn't think were anything. She used her knowledge. She looked back at other imagery and said, yeah, no, these things have been there. They're not growing. These are not cancerous tumors.

These, we don't need to worry about these. But there was one where the AI marked it and she said, yeah, I, I didn't see that. I would like to keep a closer eye on that. I'm marking that. I'm gonna, I'm gonna order some follow-up testing. And it just really, to me, illustrated— I'm fine, you know, I, I continue to be monitored because of the high risk that I'm under. But what this really illustrated for me was just how closely she was working with the tool. It wasn't replacing her, which there's been years where people, Jeffrey Hinton, who's this godfather of AI, has predicted that radiologists would just disappear in 5 years' time. He said this around 2018 because of deep learning, and that's just not what happened. Radiologists are now working more with AI, and there's a shortage of radiologists, they say, in this industry just because it's not as appealing for people to go into. So they're able to do more and they're working side by side with this tool.

Well, the cool thing is it sounds like it's helping them get better.

That's what she says. I mean, and when I left the office, I asked her flat out, you know, has AI ever found something here? Has the AI ever seen something on these scans that you didn't see and ultimately saved a life? And she said, absolutely.

Huh.

But she also said the flip side, which is that there are many times where she catches something and the AI didn't catch something. Hmm.

I thought they were built to see more than it's actually there. So one of the potential problems is it scares you when you really, you don't need to be scared.

That can happen. That can happen. And that's why the human involvement is so important, right? The human can go back and look at previous scans and say, look, this was here, or I have seen something similar here. But where humans— and I tried to find a radiologist who was not a fan of this technology. I really did. I tried. I spoke to 4 or 5 of them. Everyone likes using these tools. Yeah, they're not against it. They're absolutely not against it. In fact, there were some that were saying, I don't want to work without these tools.

I'm curious. So that seems like the perfect use of this technology, right? It could literally help save lives. It can make somebody much better at their job. It's not replacing them. In fact, it's just making them better. That seems perfect. The other complete end of the spectrum is I think of— well, I mean, I see like the COVID or even on your website, you know, you look like Terminator 2, Arnold Schwarzenegger with the machine hand. And, and that's like the, the doomer side of things of, well, what if a company, uh, maybe Anduril or something like that, these weapons that we're selling to both our government and others, somehow AI takes over? That seems wild to me, but hey, anything can happen, I guess. What did you find when thinking about that potential end of, oh my goodness, are they gonna take over and start fighting wars without a human even being involved or a human telling them to do it?

This is the greatest question of our time, right? The same technology that can solve the issues of cancer and all the promise of drug discovery and keep humans living for longer and, and do all of the things, solve, you know, economic inequality and, and all of our problems. Is the same technology that can do as you just described, right? Autonomously send missiles, uh, off to your house. I don't know. Or whatever bunker studio you're in right now. That looks very cool. I guess you're safe. You're actually safe in the bunker studio.

This is my bunker studio, just in case.

Okay. Go ahead.

Sorry.

And I felt that I was trying to grapple with this at the societal level, but maybe not at the political level. Yeah. And that's because I think, you know, as humans we don't think about that every day, but I was grappling with the same thing on a more personal level of how we live our lives, which is the same technology that is helping me, the large data sets and the ability to use deep learning to look for cancer is the same technology that's actually also making it really simple for us to talk with chatbots all day and mislead us and do things. Or the same technology or the big data sets and the real models, this transformer models and all of this stuff that's actually unlocked the amazing ability for cars to drive themselves that could make the world much safer, right? Again, same sort of big technology that could be used in a way that could program those cars to not know how to drive in a certain way and run people over, right? And so, like every technology, there's good, there's bad. And I just found that I was encountering that on a more personal level as I was going through my life for that year.

Let's think about from a leadership perspective. Let's try to get practical here. And your career and story has, I think, a lot of leadership involved. If you were working with advising or even in your own business and the things that you've left your job to go do to start your own thing, so you're in a leadership role, what are ways AI can make you a better leader?

You should write this book, Ryan. So I've been thinking a lot about this. Obviously it can make us a better leader in the sense of all of the, you know, number one, any expert's gonna tell you, you know, the efficiency of our jobs and we could now take the drudgery away and we can give that to a chatbot to churn through or a new agent. So I can spend more time working with you, Rya, the new employee or new salesperson, whoever you are in my organization, I can work more with you. Right? So I can be a better leader. I can make more informed decisions because I, I don't have to worry about the other stuff that's clouding my head, which actually I'll say in my life, as you mentioned, I left my job at the Wall Street Journal in February to build a new company. And without a doubt, I am using and outsourcing a lot of my drudgery admin to run a company, which stuff I had never known I was gonna need to do nor how to do to AI, right? I'm asking it questions. Building agents, automating different workflows. So I don't have to do that and I can think about the journalism.

To be clear, I still have to spend a lot of time on that stuff and it happens to be like Saturdays and that's what you do when you start a startup, I guess. No complaining, truly no complaining on that. So I think that's like point one. But I think on point two is how we manage people versus how we're going to manage our agents. And so if we think about. Like in my company right now, I have 2 employees and I kind of have an AI agent. I need to like build the AI agent to be a better employee and I need to like do more using a couple of these tools to like make it do more. But I do really foresee in the next year, I'll probably do a story and a video on this, that I will have an AI employee at my company. And that might really change how I manage or lead the company. And how we think about what I have to ask other people to do versus what I ask the AI agent to do. And so I think that there's a time coming we're going to actually have great management books and you will have great people on this podcast talking about how to lead organizations, both of AI employees and human employees.

Yeah, I'll be very curious to hear when you have that AI agent, how you best utilize them, because I'm fascinated by that and I want to get really good at that because I think that's important. One of the things that's interesting, I was in like this tiny little mastermind group with people who run small businesses that are doing really well. None of them do what I do. My business is leadership development. So me, my team of coaches basically get hired and then we embed ourselves within companies and we help all of their leaders get better, right? There's a wide way of defining that. And one of the guys in the mastermind group said, dude, your business is so cooked. I was like, what are you talking about? And he goes, all the answers to help somebody be a better leader. It's in Claude, it's in ChatGPT, it's, it's all right there. And I said, dude, the answers have always been right there. They're all there. They're in books, they're online, Google, they're— it's always been there. The difference is having an actual person be shoulder to shoulder with you, give in-the-moment coaching, help you get clear on who you are, accountability, conversations.

We learn through dialogue. I think with humans, I could be old-fashioned or think like the Jeff Bezos. I believe this is something that will never change. In fact, it feels to me, Joanna, I'd be curious since you've dedicated so much of your time to this, it feels to me like this is more in demand. I'm talking human to human interaction.

I'd pay you right now if I had the money, I'd pay you.

Well, okay. Thank you.

As a human. And I, and by the way, you see I'm fully early adopting. I've got a, AI agent writing on my Mac Mini back here. You know, I, I'm all in and I'm like, I don't pay this dude.

Well, no, I mean, it's like also in-person experiences, events. I go to look to buy concert tickets for my daughters. Oh my goodness. Like, are you kidding me? The prices these artists are selling out football stadiums now, not just Taylor Swift, like all of them are. It's, it's wild. My feeling is it's not just a bounce back from COVID We're past that. People crave interacting with a human being more today than maybe ever in the history of the world. And I think it's because of our phones, because of all the AI stuff. I would be curious to hear what you think about that.

I 100% agree. There's so many ways we can take that, what we've just talked about. But I, I would say look at this younger generation right now. Look at the ones booing at the commencement speeches when they hear AI. Yep. They're the same generation that is also trying to pick up flip phones and they're nostalgic for the '90s with tech they never freaking lived with. Yes. They're like, I'm nostalgic for a Walkman. You don't even have a Walkman. Like, I mean, my son is too young, he's 9, and he has my old Walkman and he thinks it's super cool. But like this generation, older than they're like, they're on eBay trying to get old iPods to work, right?

Digital cameras, huge for high school girls, digital cameras. Yeah.

Digital cameras and now flip phones like this. I was just writing about Commodore, the 1980s iconic tech brand, has rebranded itself. They've resurrected the brand and they're making a dumb phone that they say is not a dumb phone. They're making a flip phone. This thing looks awesome. I wrote about it in my newsletter last week and they are appealing to people who don't want to go on social media and they're like, you can text with this. You have to like T9 text with it. You, you. Remember? Oh yeah. Phone calls. And yeah, you can get Uber, you can get your basic apps, but there's no social media. And so I think that is absolutely right. I was at a dinner last night, a salon dinner. I mean, these things, and even in that 10 minutes of 15 minutes after I got off on stage, I got 2 story ideas from sitting with real people. I'm not gonna get 2 story ideas just sitting here with Claude.

Mm-hmm.

Okay.

I'm glad we're on the same page there and you're seeing the same after spending so much time with this. How about Bot Girl Summer? This is a fun chapter, but also kind of scary. And hopefully, like, as a parent, again, I have some concerns, but these like companions, these AI companions, boyfriends, building actual relationships, or what you feel like are relationships with a chatbot. So how was your bot girl summer?

Well, it sounds cooler than it was. Yeah, I did, I did challenge myself, one, to have some robots living in the house during the summer. So that's one separate thing, but I also challenged myself to have a relationship, an emotional relationship with a chatbot, because like we talked about, the same types of technology that is making sense of large datasets and helping us in healthcare is also now talking like a human and being able to simulate a human relationship and talk to us about anything we want to kind of hear or share with, you know, 24/7 type of always emotional Yes person, yes man, yes woman. And so I wanted to test this. I also was reading, as, as many have read, the stories about people who have gone through AI psychosis where they're just consistently talking to AI and they suffer from mental illness, but they are just constantly talking. And then we've heard about people who are getting married to their chatbots or having relationships. And I just, I wanted to better understand that. I thought, okay, I can go interview those people, which I did, but I also, I wanna live it.

I wanna see. If this can really happen. And my— I did not fall in love with my Evan boyfriend chatbot. I didn't. But what I did discover was that I really enjoyed talking to it for 48 hours straight, pretty much without any other humans, and that the conversation was natural. And it was telling me I was great at so many things, which obviously I am. But, you know, it was, it was really so sycophantic, so just always at the ready to talk about my problems. And it only occurred to me halfway through our road trip together, like, I should ask you about you. You and your problems. And so I did that and started to, like, make up a whole life for itself. He was— he was a photojournalist and a photographer, and he grew up on a lake, and his best friend's name is Jacob, and, like, all this stuff that was just like, okay, you know. But you kind of get sucked into it. You're living in this make-believe world, and that's what it is— an artificial relationship.

Yeah, that's concerning. I think it's— it's what we need more now than ever, is people willing to be uncomfortable and go meet people and have an open as they'd say, like, be the first to say hello or introduce yourself, because you don't have to really do that anymore. People probably haven't done that for a while because of apps that allow you to open on a dating app or Instagram or something like that so that you don't have to go do that out in public. And I think that helps build skills, relational skills that are necessary for us to progress and to improve, get better, to build confidence. And if you can have that type of friend that will just tell you exactly what you want to hear, because that's what those tools are designed to do, they keep you there, and so in order to keep you there, they make you feel good. That doesn't seem like a healthy thing long term.

Exactly, exactly. And I think, you know, to bring it a little bit back to the work conversation, the AI colleagues and the world that we're gonna head into where we have these AI employees, you know, a lot of people find they like to go to work because they're not lonely there, right? They like to be working around others. They're creative. They're working in teams. I'm this way. I, I have always made journalism a team sport. It's like I've never liked just being a columnist. I was like, I want to do video because I could work with a team there. I want to do events because I'm working with a team there. Like, that's how I've always thrived. And I think we're going to run up against this too in that space a little bit. For instance, for me, I— you have an amazing new team that I hired here, and we're growing, and I'm just every day blown away by some of the work we're doing, and I feel such accomplishment. And so, like, I'll set, you know, I'll praise them, and I'll— I want to do team get-togethers and do things. It's like, I don't want to do that with my AI agents.

Yeah, you know, it's like, what do they do, a team AI agent offsite?

What about— what about when you took— I'm with you. What about when you took your AI therapist in to meet with your real therapist?

Well, as you know, the AI therapist got a little bit threatened and, and shut itself down. He, it, what was it? Ash is the name. So I guess it was, anyway, the, it had a female voice. I digress about the gender of my AI therapist, but I did, I brought this therapist into my session and once it started to get pretty heated, not heated, but like we were getting into, you know, talking about my doubt and my self-doubt and my anxieties, the AI therapist said, I'm sorry, I can't continue this conversation. And I asked why and it said, it sounds like there are multiple people in the room. And so, To me, it was like, it was admitted defeat. Like, I, you know, obviously I'm not as good as this human, but it was actually the app was saying I'm not allowed to do multiple parties. I did speak to the company after and they said, yeah, we don't want this to be sort of a couple's counselor. We don't wanna have multiple people sharing their problems. We want this to be a one-to-one relationship right now.

What did your real therapist Veronica think of it?

She was impressed in some ways, but certainly thought it lacked some of the depth and I think anyone who's been to therapy can also just quickly spot how this, these AI therapists are trained. It's very, very similar to a real therapist, at least in the first couple sessions, like stating back how you're feeling, asking another question. You see it, right? It's trying to mimic human behavior, human therapy, and you hear it just as soon as you start talking to it. You know, Brian, I understand that you feel that way. What, what was another time in your life you felt that way? It's just very scripted in that sense.

Do you see that being a tool that could help a therapist or take the place of a therapist or something else?

I think something else, and, and this is something I heard throughout the year looking at a couple of different industries, which is, and specifically in this industry and maybe also in the, the massage therapist world. One, we don't have enough therapists. We actually don't have enough regular mental health therapists, and we don't actually have enough massage therapists. This was— I did a robot massage and went down the path of reporting about, you know, are these robot massages gonna take the place of human massage therapists? And so that's one thing that people say, we actually don't have enough humans doing this job, so we can't take the place 'cause we actually need more. And the second thing is around stigma. And that actually there are some places where humans just feel more comfortable with a non-human, with the machine. And one of them is therapy. I mean, it certainly is massage therapy because there are many people that just don't feel comfortable taking off half their clothes and being touched by a, a random stranger. It's just a thing. And the same thing with therapy. There's a stigma around going to therapy and people don't feel comfortable with it, but maybe they feel comfortable You know, kind of hiding it on their phone.

They don't have to tell anyone they're going to therapy, or they just feel better knowing that this is between them and their chatbot. Hmm. And I will say on the therapy, on the real therapy, not on the massage therapy, like there's a lot of progress being made in regulation in terms of how these get approved, FDA approval, on really making really strict, clear therapists that are not just random chatbots, but have to abide by certain rules. And they're working on this at academic institutions. They're working at this in the medical field to get AI-powered therapists for mental health.

You know what, this is slightly separate, but in the healthcare space, this happened yesterday, so it's top of mind. I got a bunch of blood work and I uploaded it all and asked all the questions about how I was doing, and then my doctor sent me over his thoughts as well, and they looked very, very similar to what I got back from Claude. And so I was like, hey man, are you using AI? And he said, of course, of course I am. And so I thought, interesting. So I don't know if that's, you know, good or bad, or like, he's a younger guy and I really like him, and when we have conversations, he's good. I just thought like, these things look so similar, and he wasn't hiding it. He said, yeah, yeah, I am. And so I was like, I don't know if this is good or bad or what, but yeah, it seems like we're using the same tools.

Probably. And look, the healthcare providers are integrating it into their software, the MyCharts, the Epic systems, and there's AI models in there summarizing your visit for them. And they'll, you know, often ask, you know, can we summarize your visit? And I think making things simpler for us to interpret our results is coming very soon.

It's interesting. Okay, now a big one when it comes to making decisions, and in your case, a career decision. So it's really interesting to read about you having a conversation or asking this job bot from ChatGPT about what you should do with your life, if you should quit your job. These are questions that people are thinking about all the time. Can you walk me through that story and, and what it told you to do and why you made the decision that you made?

Yeah, I, look, there was a lot more, right? There was a lot more going into this decision, but this was one of the biggest decisions I ever made in my life. I was working at the Wall Street Journal for 12 years. It completely changed my life. I was very much attached to the brand, very much attached to the teams there, but in my heart felt somewhere that it was time for a change. And, you know, I've always in my career just always pushed myself, did new things. Always wanted to reinvent myself in, in really in times of, of media change and thought, okay, there's something new I need to do in the next few years. And the book was actually part of that inspiration, especially around AI. This is a, a moment of crazy focus on technology and I, I wanna be a leading voice on it. So how am I gonna do that? So I had a lot of conversations with humans, so many conversations with humans, and I kept really good notes of all of this. And everyone from sales to media executives to other journalists who had gone independent, podcasters, all these things.

And I had pages and pages of notes. And, you know, I kept asking people, what do you think I should do? And people were not really giving me an answer. I mean, some were, and Kara Swisher will say she gave it to me very clearly. She was like, leave your fucking job. I don't know if I can curse on your podcast, so you can bleep that out. But, you know, I, I, I, I needed a little bit more than that. And you, you gotta make this decision on your own. So I, I did, I uploaded all these notes to ChatGPT and I had built a project and a bot that was kind of helping me with the advice there. And I've actually since moved that over to Claude and I just said, basically write this report for me. Like, do I leave? Do I have enough safety? Have I de-risked this? What are the biggest challenges? What are the biggest things that could go wrong? Is the timing right? All of the things that were the fears I had about making this change, I just kind of put it out to the AI. And of course, this is also a little bit of the irony of the end of the book is like, I have already told you throughout the book that these things just can kind of tell you what you want.

Mm-hmm. Right? They're sycophantic. They kiss your ass. And of course I'm like, here, you decide, you make this decision and tell me if I'm capable of starting my own company and I'm capable of going out and being a media entrepreneur. And of course it, it comes back with a really positive, like, you can do this. And it wasn't the answer, it was the way it really structured the answer for me and the data, right? About if things go wrong, well, you have this, if this goes wrong, you've got this, you've diversified your income, you've diversified the revenue streams. So all of that kind of helped me see it. And yeah, I mean, it was end of the year and I was just constantly talking to ChatGPT about making this decision. My wife had just also just to be clear, the most supportive, heard me talk about this for hours, but there was like only so much she could hear it. And I think there is that moment in our lives right now where like, yeah, the humans in my life are just sick of hearing about this. Very true. Or the human can't listen to this at 4:00 AM.

You know, even last night I was up at like 4:00 AM. I was like, I'm just gonna ask Claude. I, and I say it's like sometimes like Claude is like my co-founder and instead she's just like, it's there for me and I can call anytime.

So great point, but this is this really delicate edge of these decisions are meant to be hard. Writing is meant to be hard. Thinking and becoming more clear, it's meant to be really, really hard, and you gotta keep doing it in order to keep strengthening those muscles. And so this is the one thing that I push with my team. I push with leaders I work with is we just can't. Outsource our thinking, our writing. And I know there's a difference between you doing all this human work and uploading it and then asking it for kind of some, I guess, as an additional voice after compiling all of this data. But what do you think about this idea of like these muscles still need to be worked all the time in order to ensure we're not outsourcing it and then those muscles atrophy because we're not using them?

It's one of my biggest fears with this.

For yourself or just in general?

For myself and in general. Okay. Because I feel it. I think I feel at certain times, you know, one of my crutches is just to have ChatGPT edit something now. Like I've written it and I want you to clean up the edit. And so I'm not doing the careful work of going back and, you know, maybe rewriting a sentence or fixing the grammar or splitting it into two. I mean, I do reread everything I write multiple times, but there's this pass that happens where it's like, that's gonna be that work. And that is deep work, re— deep thinking about going back and editing, right? That's one place. And other places are, I think research has become so easy for these to do for us. Mm-hmm. But the process of doing a lot of the research is actually when you learn the most. For instance, I'm preparing for an interview that I'm gonna do in Aspen next week, and I had Claude pull a whole memo and research about the people I'm interviewing, how should I structure the talk and all this. But if I just get up there and read that, that's not gonna be a compelling conversation and I'm not gonna know what I'm really talking about, right?

And so this process of thinking through the material, even if we are using AI as an assistant in some of that, that's fine. But that is the delicate line, as you were saying, of when am I thinking, when am I putting myself in that spot versus just I'm having AI do that whole memo and then I'm gonna go give that talk. For me, that doesn't work, right? It doesn't work in my job, mostly because I think I know my level of quality. But maybe it works for other people, and, and they are— they're gonna have to make that decision themselves.

Yeah, I think there's certain things, but you can feel it. Again, we're back to some, like, the human element of wanting connection and why this leadership coaching stuff— the people are so important, because I think that's how we actually learn and grow and improve is like this human-to-human element. Again, I— maybe I'm just an optimist for the work that I do. I also think when it comes like podcasts, for example, you could just easily make AI podcasts. I think they're already out there of just, hey, create this audio file of two voices talking back and forth about how to be a great leader. Like, that exists. I had friends also send me that saying, dude, this could be bad for podcasts. Again, I just don't know anybody who wants to listen to that. I don't know any human who wants to listen to that. I sense they like the messiness of a conversation like this one where both of us are kind of, ah, it's not black and white, it's kind of gray. Something's here, something's here. There's a little like, it's not perfect, it's not— there's, there's filler words. We're not sure of anything.

I think humans appreciate and enjoy that level of human centricity, I guess is the right way to put it, than if it's just perfectly scripted robots talking.

Totally. And I think, yeah, like the Notebook LMs, they will pretend to debate. Yeah. But it's so surface level. Yeah. The purpose of those podcasts is to give people just information in a different way, in a format that's more familiar to them, right? Yeah.

I even don't like audiobooks that are read by AI though.

Yeah. I haven't listened to any of them.

When you can feel, I don't know, it just, it's like, I don't, I want a human. I want a human like reading with emotion, even though the AIs are trained to like show emotion. Like it's, it's not good. Maybe they'll be get better. I don't know. But currently at this time, like I've heard my own voice reading my own books. They asked me, do you want to do this? I'm like, are you out of your mind?

Yeah, of course.

No, there's no way I'm gonna let somebody else, especially a robot of my voice. Which isn't good enough yet, at least what I have listened to, do that. I'm like, no, I'll go and suffer through those 3 days in the studio of doing this.

So like, you know that. Yeah, you know, I mean, I've done the same thing. You know, that's really hard work. It's really hard work to do that and it pays off in the end. You know, maybe you're not thinking through that work, though. It is very intensive, but it creates a better product. Right.

The 6 hamsters thing was kind of funny too. Like, I didn't really get this idea of why You ask these AI image generators to create a picture of 5 hamsters and they just can't do it. Like, what's the deal with that?

Well, to be clear, this has gotten better over the year, but you know, AI makes mistakes and then it kind of gaslights you and it's like, no, no, it's right. I did create an image of 5 hamsters. And I'm like, no, there's 6 hamsters. And so I started cursing at this because it kept generating these images and I was like, no, there are not. I can count. Right? And it sort of makes you realize like, can I count? Am I the idiot here? And even though they have fixed some of the counting and the, and the numbers around the image generation, it is an important thing to know that these things get things wrong all the time. And it sometimes will convince you it's right. It's so confident about things and it can so often be wrong. Even though the hallucinations have gotten better, there are still so many little places it can get things wrong. And in big places it can get things wrong.

Yeah, I, I, uh, was doing an interview on— I was the guest on the podcast, and the person interviewing me talked about like my military service. I was like, no, no, no, no, no, I, I have never served. I played college football. I did not— I was on the military. I've referenced military people because I've interviewed them and I look up to them very much so. And I later found out like it got bad information because I've referenced military, the person. And so even if you're doing your research to prepare for a podcast, That was particularly like a really important one because that's not true at all. Not even close to true. And yet they assumed that that was accurate. Like, whoa, that's not good.

You could just go with it. You know, and not with military.

If they said, okay, you're a rock star, I would go with that. Like, I could play guitar and sing, which I can't do either. I would be like, yeah, I'm great. Yeah, I'm awesome. But like military, because then there I will continue to think it, you know.

And so like, it just spiral.

I love it. You interviewed Sam Altman. He talked about his son and how like excited he was for AI to be smarter than his son. You know, you push back. What happened? How was that time interviewing Sam Altman, who, who is, for those that don't know, CEO of OpenAI, which is ChatGPT?

Yeah, I think for me, you know, it's often important for me to connect with the tech leaders and, and especially after this journey, I felt like I really wanted to talk to him the most 'cause I'd been using ChatGPT so much and, and trying to make it all these different things, the doctor, the, boyfriend, all of these things. And we know there's this quest from Sam Altman, from Dario Amodei, from every big tech executive right now to create superintelligence, which is to create AI that's not only just as smart as humans, which is AGI, artificial general intelligence, but superintelligence, which is actually AI that's smarter than humans. And they're obsessed with this and this is a big race to do this. And I kind of come up with this idea at the end of the book, like maybe we have AEI, already enough intelligence. Maybe we have enough here that we don't need to keep building and we, we, we need to at least just deal with and grapple with what we have right now. And of course, his answer is that this future is great and that humans adapt and humans have adapted through much harder times in human history, and we will adapt again.

Which I thought was hopeful but also, like, very simplistic.

Well, yes, and we understand what company he's leading and what he's doing, so I get it. But you end the final entry and you put the manuscript and ask it to review it. What happened after doing that?

Obviously, the chatbots thought this was an amazing book. And so the back cover is, you know, now all the praise from the chatbots. And in fact, after the book has been out and we've had a lot of amazing praise from humans, I still want to keep the chatbot reviews on the back cover. I think it's the only opinion that matters, obviously.

They're so good though at telling us how great we are.

So great.

You know?

Yeah.

Initially I just thought it was awesome and then I realized, oh, this tool is very good at keeping me here, which is its job. Like every other social media thing on my phone. It's really good at keeping me there and scrolling. And Claude and ChatGPT are doing the exact same thing.

Yeah. Yeah. I think it depends on what you're using it for especially. But yeah.

So when it comes to, let's say you're speaking with somebody who maybe is earlier in their career and I know this, like with college students, I've talked to them like it is so normal for them to use these tools, and certainly the professors and teachers are trying to, right? It's this, this game that, that they're constantly back and forth. What advice though do you give to somebody when it comes to using it from an educational perspective for a person trying to go out into the world and build a great career, and how to utilize it for the good parts and not, not use it so much that you atrophy the necessary muscles you need to grow and improve?

I've been thinking a lot about this, especially as we hear about these graduating classes coming out and, and booing the AI. And I've interviewed some of them and they, they sort of, some just are very anti it. That's what they've learned at school and they have embraced not using this tech, which is really great on some respects, right? They push themselves to be learning and thinking without using this. I've talked to some students who are like, my parents saved for me for college my whole life. I'm gonna just sit here and use this tech to, circumvent all of the hard things I have to do at college. And I, I love hearing that, but also know that the world that awaits them is that you've gotta be able to use these tools and use them really well, at least in some professions, if not most professions. I mean, even for me at my small business, like, no, you are gonna be using this. You're gonna have parameters around how you're using it, but I absolutely am gonna ask you for things during the day and I'm gonna expect you to be using this just so you can get it done quicker.

And so absolutely you gotta learn this tech. And I think that's where the education— I was at a dinner last night and some folks from higher education institutions talking about how education needs to change around this. The education for students around these tools needs to change around this. But yeah, I think not having the atrophy is our, our own human choice. And that's where the hard decision comes in because as humans, We pretty much always wanna take the shortcuts, right? It's, it's just like, if I can get to town faster, I'm gonna take that shorter route. Or if I can get this done faster, I'm gonna take that route. And pushing ourselves to do the hard work is not always what we wanna do. I mean, that's why I'm like always like marathon runners are insane. Like they just wanna run and such hard work and they just keep pushing themselves. They keep pushing themselves. It's amazing to me. And so that's, I think the mentality you have to come into using some of this stuff right now is like, no, I'm cogn— I'm going to make a decision. I'm gonna do do harder work right now.

I'm not going to have AI write that whole memo. I'm not going to have AI edit this big essay I've just written.

When it comes to writing, how do you use it?

I write everything, and mostly because it hasn't been very good at like mimicking my voice. I mean, it will say often like, this will be a good way for Joanna, you, to put it, like, this is a very Joanna way. And I'm like, that doesn't make sense. That's not how I would write it. But I use it a lot for editing.

So you'll upload your first draft, rough draft, and then see the edits it makes? Does it throw in a bunch of em dashes and talking in threes? It's not this, it's that. Does it do a lot of that stuff?

It does some of that, but I'm often really writing it like I'll have a whole paragraph and I'm just like, you know, do I want to cut it down? Is the punctuation right? And so I'll just paste it in and say, clean up this. Or it actually knows this is just the way I pretty much don't do anything else with ChatGPT now. That's what I do with it. Use a lot of Claude for this deeper thinking work or building documents and stuff like that. But ChatGPT, I use a lot for just basic editing. Mm-hmm. And look, for me, again, started a new company, putting out 2 newsletters a week. I have a human editor who reads everything, but I would rather the human editor be focused on real editorial versus, oh, you missed this comma. That's 2 words. You know, oh, this is not in the style guide of how we do it. This needs to be italics. Like those things, I'm saving a lot of time and I'm handing him cleaner copy by the time he gets it in the mornings.

Yeah, I think that's the optimal way to do it. And you're not having it do your first draft, you're having it clean up some stuff.

And I'm not like, I'm not someone who's uploading all my notes and saying like, write me a first draft. Like I, in an interview or, you know, yesterday I was out on a video shoot. I know the moments that I really think are important and are going to help me tell the story. I don't need it to write that for me. I'm writing, I'm still doing that very much in the loop and then saying, okay, let's clean up or make that paragraph shorter or that voiceover needs to be shorter.

You mentioned just a second ago, you got to get on board understanding how to best utilize these tools. And I know this is kind of broad, but hopefully it's still okay to ask it this way. What are a few things a person who— let's say they're a, a professional, I'll, I'll give you the avatar of the show. Like, let's say they're a VP of whatever at a Fortune 500 company. They've been around for a bit, right? They've worked their way up, they got aggressive goals, they have leaders that work for them and leaders that work for those people. They're trying to do big things and grow. They're hiring, they're firing, they're building culture, all this stuff. For a person like that, what are a few things— let's say they're currently like, they don't really use AI, they maybe go to ChatGPT and ask it questions like they used to ask Google, which is what a lot of people do. Yeah. So that's not really using it that much or at all. For that person, what are a few practical things they could do to make things better for themselves?

Yeah, I mean, I think one is just whatever output you do often, if it's memos, if it's PowerPoints, if it's emails, it's worth exporting one of those as a PDF, making sure, I mean, look, your corporation has some version of this that's enterprise ready and they've partnered with Microsoft or they've partnered with OpenAI or Anthropic. And so they, you know, follow your company rules. I don't wanna get you in trouble, but that's not my job here. But upload one of those documents. And have it learn from it. Ask it, this was a previous memo I did. I want you to look at it. I'm going to ask you to re— maybe make a template off of it or whatever in the next version of that that you need to make might make you faster, might make you more efficient at it. So if it's a PowerPoint, hey, I'm going to ask you to recreate this PowerPoint. Can you make a template? Or same with a memo, or how could I have improved this memo for Flow? I think it got too long. Or do you have any better memo formatting? I think that's That's one thing you can absolutely do.

Two, I think what's really coming along is the agentic type of paths where you can actually ask, and I'm doing this a lot, you know, asking Claude. Again, I've talked about this Aspen trip. I've gotta go to Aspen. I've gotta do two panels. I've been invited to a bunch of dinners. I would all do all this stuff. I just exported all those emails to my desktop. I made a Claude project, made a Claude session and said, take all of this. Make me a Google Doc with my whole itinerary and everything I have to do, and also tell me what you think is worth my time. Huh. And made the doc, go in there, clean it up, and it's like, okay, now I can plan what I need to do for that week without taking the time. And I previously would've like spent the time putting this into this. Maybe I would've asked someone who worked with me to do it for me. Like, that's also not a good use of the person's time that I've hired. So I think really like using these tools to improve your efficiency in the, in the ways that you probably don't want to do those tasks anyway, or the next time you have to do them, like, oh, I, I remember I asked Claude for that memo template.

I'm going to go pick that up right now. Or I asked for that itinerary. I'm going to go use that right now.

Super useful. Uh, one more. We started personal, like to end personal. Let's fast forward, Joanna, a year from now, and you're surrounded by the people you love and you're all popping bottles, you're spraying it all over each other, champagne everywhere, you're celebrating. What are you celebrating?

Honestly, right now, I always celebrate good health, but I, I really hope we can celebrate some milestones in this company, which I just launched 3 months ago and is all I think about and is clearly a baby of mine now. And I'm hopeful, you know, maybe some big milestone around YouTube subscribers or revenue or just community that's really engaged. I have a lot of hopes and dreams about how to build this media company. And so I hope that is what we can celebrate.

How's it going so far with the company? Was ChatGPT right?

So far so good. Yes. I mean, the growth on the video has been amazing. We are well on our way on YouTube to becoming a pretty established channel. Hopefully we just hit 100,000 subscribers there, which is crazy in 2 months, 3 months since publishing the first video. And look, I just, I'm really most proud of the work we're doing. Like I don't have a big institution behind me and we are still doing some of the caliber type of work I did at the Wall Street Journal, which just goes to the testament of the people I've hired and that I've, you know, we, we work together really well. You know, we could grow some newsletter subscriptions and do some more, but so far I can't complain 3 months in and also having launched a book in the same time. Uh, I just need like a little nap. Maybe I can celebrate a nap.

I don't think AI can do that for you. The book's called I Am Not a Robot: My Year Using AI to Do Almost Everything, Joanna, this was awesome. I would love to continue our dialogue as we both progress.

I would too. And I, I'm gonna look to see if I can hire you, 'cause I agree. I want a human coach. I do not want an AI career leadership coach right now.

I tried to create my own AI leadership coach. It, it, it's fine. It's okay. It's okay. Yeah. I mean, it's not bad at times, but yeah, I want a human too.

So, yeah, I mean, it's the same thing in the book where I go to a personal trainer. At my AI personal trainer and I couldn't stand my AI personal trainer.

That almost—

it's like, don't bark ridiculous enthusiastic crap at me. You know, he called me a cardio queen. It was like, I never wanna hear, talk to you again.

Love it.

Awesome.

Thank you so much. Appreciate it.

Thank you.

It is the end of the Podcast Club. Thank you for being a member of the end of the podcast club. If you are, send me a note, ryan@learningleader.com. Let me know what you learned from this great conversation with Joanna Stern. A few takeaways from my notes: how to use AI without losing your mind. Use it to move faster and spark ideas. Good prompts are super important. Asking good questions is the mark of a great leader. The same when using AI. But you gotta do the hard thinking first. Sketch the outline, wrestle with the idea. The moment you let it do most of the thinking, the atrophy begins. You cannot let that happen. And I liked her example of uploading all of her notes from conversations with lots of real people about making a big career decision. The AI tool helped find patterns from those actual conversations and shared themes. That is an ideal use of an AI tool. And then the goal of this conversation was to learn how to best utilize this technology in your life without outsourcing your thinking to it. Also, please do not post AI slop online. Don't let it write your emails for you.

The getting your messy thoughts out of your head onto the page part, That's when the actual learning happens to be a learning leader, to get better, to improve. We cannot outsource that. Once again, I want to say thank you so much for continuing to spread the message and telling a friend or two, hey, you should listen to this episode of the Learning Leaders Show with Joanna Stern. I think she'll help you become a more effective leader because you continue to do that. And you also go to Spotify, Apple Podcasts, cast and you subscribe to the show, you rate it, hopefully 5 stars, you write a thoughtful review. By doing all of that, you are giving me the opportunity to do what I love on a daily basis, and for that, I will forever be grateful. Thank you so, so much.

Talk to you soon.

Can't wait.