Offline with Jon Favreau - Healing Our Broken Brains

Episode Date: May 9, 2026

Cal Newport, computer scientist and author of Deep Work, joins Offline to explain why we need a revolution in cognitive fitness, and how AI is going to get in the way. Cal and Jon explore how smartphones and AI are destroying our ability to concentrate, how our attention spans are a third of what they were just twenty years ago, and how we can practice “deep work” in our constantly interrupted digital environment.

Artificial intelligence is moving very, very fast, and it's raising new questions just about every day about what it is, what it isn't? And all of a sudden done, what is the end game? I'm Chris Hayes, and as part of my podcast, Why Is This Happening? I'm speaking with leading experts each week to help ground that conversation. We're right now in a situation where it's very difficult to understand what is real and what's not real. Why is this happening? The AI Endgame, a special mini-series from MS Now.

Start listing today, wherever you get your podcast. Get Road Ready with CarQuest. Auto parts. Save now on premium brand name products. Castro Magnetech Synthetic Oil 5-liter jugs, 297. Rainex Summer Breeze washer fluid, 3.78 liter jug 397. Air 1 diesel exhaust fluid 9.46 liters, 1397. Seafone motor treatment 473 milliliters 897. Hit the road with CarQuest for the parts you need. Visit a Carquest Auto Part Store or Carquest.Cast.Ca for more information and full flyer details. Only at Participating locations. Everything we care about as moderns comes from the connection of those two tasks.

You know, reading, strengthens your brain writing helps you put it into effort. So I'm very wary. I'm always very wary of any technology that comes in and says, let's reduce those efforts. And so when I think about AI, I say, here's the division to make. If this is helping you with something that's boring, that's probably good. But if what you're doing is getting rid of hard, but you're trying to avoid that feeling of intellectual strain that comes whenever we use our brain into ways that, again, I think is key to the modern human experience. So I don't want to read and try to understand this,

or I don't want to face a blank page because it's hard. It feels strain. Trying to automate away hardness, that's where I get worried. I'm John Favreau, and you just heard from today's guest, computer scientist, and author of Deep Work Rules for Focus Success in a Distracted World, Dr. Cal Newport. A couple of weeks ago, a piece in the New York Times written by Cal caught my eye. It was titled, There's a Good Reason You Can't Concentrate, and in it, Cal argues that America is facing a crisis. Quote, the negative impact of digital technology on our ability to think. Obviously, a familiar theme on this show. Cal maps all the familiar contours of the story that we've talked about, the way smartphones and now AI are destroying our ability to concentrate,

how our attention spans are a third of what they were just 20 years ago, and what it all means for our cognitive fitness, as he calls it. But good news. Cal is optimistic about what it'll actually take. to heal our attention spans. He notes that Americans' understanding of diet and exercises role in our health developed rapidly over a short period of time in the middle of the 20th century, and that a similar understanding of what it takes to ensure our cognitive health could develop rapidly too. I invited Cal-on to talk about that piece, what a cognitive fitness revolution could look like, and how we can practice deep work in our constantly interrupted digital

environment. It was a great conversation, and we'll get to it in a minute. But before we do, The CricketCon presale begins next week. Starting Tuesday, May 12th, Friends of the Pod subscribers can purchase tickets to CricketCon and receive a special subscriber-only discount. General tickets go on sale the following week, Tuesday, May 19th. But if you want to guarantee your tickets for this year's bigger, better CricketCon, head on over to cricket.com slash friends

and become a friend of the pod subscriber today. All right, let's get to it. Here's Cal Newport. Cal, welcome to offline. Well, thanks for having me. So I came across your New York Times. piece, there's a good reason you can't concentrate, and it became an open tab in my browser that I didn't read it first because I can't concentrate. But then my five-year-old's head of school wrote his

weekly school-wide newsletter about your essay, saying he thinks a lot about how to prepare our kids for the symbiotic relationship between humans and technology while preserving for them, and he quotes you here, our heritage as contemplative human beings. And one reaction I had, is that if educators are thinking about this for kids in elementary school, my child's school is pre-K through six, something has clearly shifted. So I want to start there. What does it tell you that this conversation about technology's impact on how we think isn't just happening at the policy level or even the high school level anymore? It's happening in kindergarten. Well, it's obviously something I've watched happen in real time because I've been talking about this

issue for over a decade now. And things do feel different. So I think it is meaningful that you're seeing, for example, at the elementary school level, the head of school talking about this. A few things have changed in our culture in the last four or five years that I think have tipped people past this point of, you know, we're just adjusting to a new technology and starting to worry that there is a real concern. I think it had to do with the shift that social media made. I think it had to do with the hyper-engagement we got out of our phones. I think it has to do with partisanship. I think it has a lot to do more recent. with AI, which for a lot of people feels like a product no one asked for, and yet it is

dominating so much their lives and sowing fear. All these threads are weaving together, and it puts us at a point now that we haven't been at before where we're thinking, wait a second, this might be an actual crisis that we can respond to and that we need to respond to. It's a meaningful moment right now. So you argue that we're in a moment analogous to the mid-20th century shift on diet and exercise, except now it's in our brain. brains, the data is alarming. You cite Gloria Mark's research on attention spans collapsing to a third of what they were in 2004, the Financial Times headline about whether humans have past peak brain power. Walk us through the strongest argument for the case you're making.

Well, I think there's a few things going on that all are a problem, right? I actually begin this attack on our ability to think in the workplace. So if you want to start this story, you start this story in the late 1990s, early 2000s, with an innocent seeming innovation, which was email. I wrote a whole book about this a few years ago. It was ushered in, if you read the original reporting on it, email was just supposed to replace fax machines and voicemail. And it was clearly a better replacement. But with it came a whole new way of working, where we were engaged in these back and forth ongoing ad hoc asynchronous conversations at all time, which has been we had to continually divide our attention between these ongoing unskittled conversations and what we

were trying to do. This is where we begin to see that slide in our attention span. All right, then we got mobile computing. Now this ability to be engaged in this attention switching, context switching followed us wherever we went. It was on our phones. Our email was on our phones. The chat was on our phones. Then social media went through the big engagement transition where they transitioned from where they were in the first decade of the 2000s, which was about how do we build audience by being as useful as possible, and it shifted to how do we get as much time and data out of the audience we already have. And now suddenly people who bought phones for one reason found themselves looking at it

even more than they had ever expected. I think all these threads came together to put us in the place we are now. And the place we see it is you can look at the data, but I think the reason why this is an issue with real prominence at the moment is we don't need data. unlike other past health crises where you might have to tease out correlations from datasets of cancer clusters around smoking or what have you. Everyone's just feeling this themselves. They'll say, I used to be able to read books.

I can't. I used to be able to sit down and enjoy this movie. Now I can't make my way through. My kid is staring at this glowing piece of screen and gets incredibly upset if it's taken away from them. I can no longer organize my thoughts in a way like I had before. So we don't even need studies, I think, to capture what most of the country is just feeling in their own life. I mean, I completely agree with that. Most of the show is

about that. I've been doing it for several years now. Because the argument makes so much sense to me, and I'm so inclined to believe it, I do want to push on it a bit for people who might be skeptical. A lot of the data underneath peak brain power is contested still. You know, student math scores have taken a hit, but some of that could be pandemic learning loss. Certainly some of it is. Adult literally declines have several possible explanations, measurement changes, economic changes. Some people still question the methodology. How confident are you that the cognitive decline you're describing is real and that smartphones and social media are doing the causal work that you say they're doing and it's not just a correlation? I mean, researchers refer to this

as the 2012 turn because you see this again and again with whatever you want to study. There's almost always if it's related to our ability to actually concentrate or to produce hard cognitive output, when you see effects of a turn, whether it's IQ scores going down, whether it's Gloria Mark's measurement of attention span, whether we're looking at performance on these international test about ability to do abstract reasoning. We see a turn that happens right at that first decade of 2010s. That turn, when you get ubiquity of smartphones, when you get ubiquity of high engagement, social media, as opposed to high audience social media that switched to mobile, that switched towards engagement.

So that's sort of the clue, right? That is the clue that says, huh, something's going on here. And then when you dive into the data, you find all sorts of specific causal studies, right, where we'll take, you know, one I was just talking about on my show, you know, just last week. A good study, one of my colleagues at Georgetown was involved in it. They did a randomized control trial. They said, what we're going to do is we're going to take 500 participants. and this group is going to use their phone as normal.

This group, we are going to block mobile internet on the phone. And because they're having softer doing the blocking, we can check for compliance, which is a real problem with past studies. Within two weeks, you see very non-trivial jumps in measures of attention. In two weeks, jumps and measure attention. Real jumps in positive affect from experience sampling. It's a randomized experience sampling where you randomly, at random times, interrupt the person and say, how are you feeling right now?

the experience sampling method being the gold standard for this, and notable decreases in anxiety and depression symptoms in that group as compared to the control. We can replicate this again and again. You'll studies where, okay, let's put someone in a context where they're now have a phone nearby. We want to see how they're doing on this type of test. Or there's been some kind of diabolical studies

where they will say, oh, your phone is interfering with our equipment during this experiment. We just need to move it over there. and then a confederate calls the phone from another room, and they can see what happens to your performance on the test, what happens to your cortisol, what happens to your brain waves. So I think they have a whole pile of evidence that backs up

what I think is intuitive for a lot of people, that if you're spending two to six hours a day in fragments staring at a screen that's trying to do pure signal engagement, this is going to affect the way your brain operates. So I want to read you something you said in 20, 19 in GQ, you said that I'm convinced that within a five-year window, the culture is going to shift on young people in smartphones. You're going to look at allowing a 13-year-old to have a smartphone the same way that you would look at allowing your 13-year-old to smoke a cigarette. I think that's going to be a cultural shift that's probably going to happen maybe even before regulation is needed. So a couple of things from that. I think you can argue that there's definitely been a cultural shift. The school phone bans are spreading. Australia passed a ban on social media for kids. Parents are clear. more anxious about kids and screens than they were. You know, I think you should take your share of credit for that shift as well. But a lot of the metrics in your piece have also gotten

measurably worse during that same period. Tension spans keep dropping as we just talked about. So the shift hasn't just been slower, I think, that you thought back in 2019. It also seems like it hasn't been enough to hold back the problems that have been unleashed by this technology. Why do you think that is? I mean, I think the the shift is just really starting to get rolling, right? So things got worse after that. And I really think it's the last year or so that we're beginning to get that cultural shift that I was predicting back then in 2019. So I think the school phone bans have been really important psychologically because what happened is we demonstrated to ourselves, oh, we actually have agency. There's a sense of sort of passivity and

inevitability when we deal with technology, at least the 21st century, this idea of technology is washing over us, that the tech leaders are these reluctant stewards of these inevitabilities. And there's nothing really we can do except for sort of adjust to make sure we keep up with the future. And that gave us a sense of agency. Well, we don't have to just take what is being dealt out to us. Maybe we don't want phones in the schools. Subjectively, people felt like this was a positive effect. We have data on this too. That's, you know, it's complicated data to gather, but just talk to teachers, talk to students, like, yeah, this is actually better. I'm glad I don't have to deal with this in school. I think that's the beginning of this agency shift towards,

oh, we can take some more control. The other thing that it's just beginning to happen now is the relationship to social media with young people is shifting. It's shifting because they no longer think about these companies as avant-garde or high-tech or cool or counterculture or something to be a part of. They see the entanglement between these companies and these leaders in politics. They aren't particularly impressed by Mark Zuckerberg or Elon Musk. There's a sense of these are giant corporations, they're trying to take advantage of you. So I think that that sort of positive counterculture valence around these tools has really dissipated for young people. And you can lose to Zikeyes very quickly. I mean, this is what happened with teen smoking in the 90s where

there is a Zikeye shift, the sort of cultural function of smoking among the population shifted as that we began to think about after the tobacco lawsuits of these sort of faceless tobacco, you know, upper middle age men in suits trying to take advantage of all the kids. The teens were like, okay, the role of smoking in my life socially, psychologically has changed. So I think we're just at the early stages of this shift where these technologies no longer seem like these inevitabilities. And I really have to emphasize how big of a shift this is. I mean, back in 2017, I believe, I wrote a New York Times op-ed where I was pushing back on social media in a pretty actually tame way. And it created an uproar. The Times actually commissioned a response op-ed that two weeks later they put

in the paper that just went point by point to my op-ed to say why it was dangerous to tell young people they didn't need to be on social media to succeed in their career. And so I really want to emphasize it wasn't that long ago in which it was just accepted that we should all be on our phones and all be using these services and they were all serving the greater good. That's not the way anyone thinks about it now. So the shift is happening, but it's still in the early enough stages that we're not seeing the, we won't see results for another, I don't know, make another prediction, another three years before we begin to really reap those rewards. Offline is brought you by Quince.

The warm weather is finally here, so it's time to change up your wardrobe rotation. Quince has you covered for all the high-quality pieces to upgrade your closet at shockingly great prices. We're talking 100% European linen shorts and shirts starting at $34 and clean 100% Pima Cotton T's with a softness that you simply need to feel for yourself. Their pants also hit that same balance, relaxed and comfortable, but still polished enough to wear pretty much anywhere. Everything's priced 50 to 80% less than what you'd find. similar brands. Quince works directly with ethical factories and cuts out the middlemen, so you're getting premium materials without the markup. Love Quince. I just got one of their linen shirts recently because I need a nice lightweight shirt for spring, summer. They got some

great shorts there. They got some great comfy clothes. Great stuff. I love Quince. Go check it out. I get some new stuff because it's a new season. Refresh your everyday with luxury you'll actually use. Head to quince.com slash offline for free shipping on your order and 365 day returns. available in Canada, too. That's Q-U-I-N-C-E.com slash offline for free shipping and 365-day returns. Quince.com slash offline. Offline is brought to you by Thrive Market. Here's some good stuff I found on Thrive Market. You get some organic kettle-cooked potato chips. Mexican shortbread cookies that are vegan and gluten-free. There's a wide variety of nuts and seeds for snacking or sprinkling on your yogurt or your oats. Huge assortment of protein bars, options for whatever

your dietary needs are. Thrive Market is a membership-based online grocery store for just $5 a month, you get access to a curated selection of organic and non-GMO brands, weekly sales, free gifts, and so much more. One annual fee, priceless piece of mind. Thrive Market restricts 1,000-plus ingredients, no stressing over labels or Googling every additive. The in-app barcode scanner allows you to find a healthier swap for almost anything in your pantry right now from high-quality brands you probably couldn't find at your local supermarket. The $5 a month membership pays for itself.

Most members make it back in their first two orders. No pre-order delivery fees. service charges or tip math, plus their 90 plus dietary filters mean you're only seeing what works for you and your household, no more decision fatigue. Ready to do your own spring reset. Join Thrive Market with my link, thrivemarket.com slash offline for $20 off your first three orders, plus you'll get a free $60 gift. Back in 2019, you know, the villain in your story was Instagram and notifications.

Now we've got AI that we're dealing with on top of all this. You distinguish between AI that saves real time, you know, sifting through documents, fixing formatting, all that kind of stuff, and AI that lets you or causes you to sort of avoid sustained concentration. And you drew a similar line in your Chronicle of Higher Ed interview a couple months ago, basically augmentation, good, automating thought, bad. I've been wondering as, you know, AI becomes more advanced and ubiquitous, like how distinct that line is because I've used AI for research to inform writing that I do myself,

but there's still like a big range between traditional fact-finding research that's like a step up from Google and, you know, writing yourself without any kind of AI at all. And I wonder, you know, like if AI helps you find the right structure for an argument and then you iterate on that argument yourself, is that automating thought? Is that augmenting thought? like how do you think about the different ways and the range of ways that you can use AI? I mean, I have a bit of an extreme take on it, which I'll acknowledge off of the bat. And maybe this is just steeped in nostalgia. But I put reading and writing in an incredibly privileged category when it comes to possible human behavior.

I think the brain circuits that we're able to forge through this highly unnatural but remarkable act of taking these artifacts and symbols on paper and converting them into actual fully formed thoughts within our brain, that sort of act of intellectual telepathy that we do when we read is incredibly important to the human experience. Basically, everything we think about that we value in the world of thought, politics, philosophy, and theology all comes out of forming that brain. And writing is where you take those circuits, you reverse them and produce original ideas in the world yourself. Everything we care about as moderns comes from the connection of those two tasks. You know, reading, strengthens your brain writing helps you put it into effort. So I'm very wary. I'm always very wary of any technology

that comes in and says, let's reduce those efforts. And so when I think about AI, I say, here's the division to make. If this is helping you with something that's boring, that's probably good. If it's, look, this is boring, I have to go through and this data and clean it up. Oh, this is boring, I have to try to look at all of these constraints that my colleagues have to figure out when to have a meeting. Getting rid of boring is good. But if what you're doing is getting rid of hard, so you're trying to avoid that feeling of intellectual strain that comes whenever we use our brain into ways that, again, I think is key to the modern human experience. So I don't want to read and try to understand this, or I don't want to face a blank page because it's hard. It feels

strain. Trying to automate away hardness, that's where I get worried, right? And so I think we should we should grapple with blank pages. I think the, let's use AI to help sharpen your argument. I think often what that ends up being is I don't like the startup cost of having to start with, I don't have the idea and it's a blank page and that's scary and it hurts the brain to think because it, you know, this is an unnatural act when we yoke a paleolithic brain to do this decidedly non-post-paleolithic type of activity producing abstract symbolic thought. But I think that strain is good.

That strain is to burn into lungs or the burn on the muscles when you go for a run or lift the weight. That is how we stay mentally strong. But I say I'm considered out on the extreme because of that. I put much stronger gates around those two fundamental activities than probably most people. Yeah, I've had this sort of funny, not evolution here, but I think that, you know, when I first started testing out chat GPT, I thought it was horrible for a million reasons. And one of them is just how sycophantic it is, right? And I think it's still like that. And then I started using Claude. And then I started, I think I put some instructions in about how much I hate sycophantic AI with that. And you see people doing this now where you come up with an argument yourself, you write an argument yourself, and then you have AI argue against your own draft,

stress test the argument for you, give you sort of the, here's the counter arguments to that. And I wonder what think about that, whether you think that is, because, you know, the way we are, if, if we're, sometimes we, it's hard to stress test our own argument because we're so biased that we wrote the argument in the first place and having another source sort of push you on that, then does do something where you're starting to think again and you're sort of exercising those muscles, but I don't know, maybe you disagree. Look, I don't hate that, right? But it's not, I would be wary of it myself, right?

I mean, so I, I'm uncomfortable with the idea of, of having a machine express itself through fluent language, right? Because I feel like the issue here is almost like emotional fraud. When our brain encounters interactive fluent English, it's going to simulate another brain on the other side of it. So we're creating a fiction that we're interacting with another entity when we're really not. And I think that does weird things to us.

We see this at the extremes with the rises of AI-induced psychosis, which completely makes sense because our brain is going to simulate another brain. And I think the future of AI, by the way, is not going to be interactive chatting. That's sort of an accident of the early stages of gen AI. And I don't think it's a very healthy stage. I don't think we should be having human-like conversation simulating another human. It's not an appropriate way to interact with a machine. I think there's better ways to tap the intelligence of it. If you've never heard another argument to your thought, maybe like, wait, maybe I'm wrong on this. Let me check it with against AI. That's not the worst thing. It's not that much different than

Googling articles on the other side of something like this. But there is a something a little bit different about, for example, I have this take on this, but I know other people have a different take, and I want to read an example of a counter take to this, a book, an article, listen to a podcast as someone who disagrees with me, because there you're actually engaging with another real human mind. And I do privilege that among above a simulated human mind. So it's not the worst use of a chatbot, but I don't even like the whole concept of chatbots. I don't think we should be having human-style conversations with a machine. We should be getting what we need out of machine without having to simulate a fake human brain behind it.

That's the thing that gives me a little bit of an ick. It gives me an ick to, I guess what I'm trying to figure out is, and I think about the social media experience, is people are going to be using this. They're using it everywhere. It's becoming ubiquitous, just like social media was. And you can tell them it's bad for you and that when you are chatting with a chatbot that seems very human that seems like it's helping you think better, that that's bad. But we know how seductive these technologies are, how addictive social media is. And I sort of worry that AI could be even more seductive and addictive because I think, I mean, you know, social media turned out to be there's a lot of trash and there's a lot of shit that makes you feel bad and there's a lot of really

dumb opinions and there's a lot of opinions that become like the group collective wisdom of social media that does not seem that smart at all. And if AI to a lot of people who are maybe tired of doing that with social media seems like, oh, well, this is smarter and more interesting than they're going to be more taken by it. And I wonder what you say to people who are thinking to themselves like, well, you know, I'm just, I'm just learning more. It's making me think. What is the argument that could convince people that this is bad for them? Yeah, I think we have to raise a lot of red flags when it comes in particular, again, to this chatbot relationship with AI. And the thing I would say is, okay, what made social media addictive is, oh, when they figured out at some point how to use

these good machine learning algorithms to essentially build an approximation of your reward centers and figure out exactly what audiovisual combination is going to most press your buttons. And we could do this through a bunch of experiments. TikTok basically mastered this. With chatbots, what you need to be worried about is we care a lot about other humans and other human brains, right? And so we can tell ourselves as much as we want, there is a collection of GPUs on the other side of this wire that are just multiplying matrices together. But deeper down in our brain, we are simulating another human brain on the other end of that conversation. And there's nothing more engaging. There's nothing more engaging to us.

And I think the ability to hijack our brains potentially is going to end up being much worse with this. Again, if we look at these emerging cases of AI psychosis, we're getting people who are ending up in the desert waiting for the aliens to come. because we really take seriously having a brain that we think is a real human who cares about us, right? Who thinks that we're, you know, interesting, that gives me really interesting information, backs up my thoughts, makes me feel like I'm special. This is why I call it emotional fraud and I really worry about it. But there's also a difference between, and I don't know what to do about this, but I just said something I'm noticing. With social media, it's a different experience than we had we're having right now with AI. Because when social media came along, it was solving a real problem.

People really liked it, right? I mean, it was clear, you know, early Facebook. It was I want to know what my friends are up to and I want to tell them what I'm up to and I can see their relationship status. And that's interesting. Early Twitter. Oh, wow. People who are famous, right, I can see what they can put out thoughts and I can see it. I mean, typically it would be if I didn't see a comedian on a late show, I wouldn't know anything they're up to.

Now they're talking about things. It's funny. This is interesting. I can connect to interesting. people. It was really solving a real problem. And then it got worse in sort of this water getting hotter in the pot and the lobster not noticing type away. And then we looked at, you know, 10 years later like, oh, wait a second. I'm now scrolling continuously on short form video. So it changed us. AI, they're not even trying to do that. It's the weirdest thing. They're basically saying,

you have to use this. We're not even going to really tell you why it's useful. Oh, and by the way, us, the leaders of this company are going to keep telling you how it's going to destroy everything you hold dear. So we're going to take all your job. jobs, we are going to kill all of the arts, and we are going, you know, we're going to be, potentially also destroy humanity in the future. But go, you have to use this. So it's, I don't know what to make. I've never seen this before in the history of technology.

The leaders of the technology companies themselves continually telling people not why they should use their tool, but how their tool is essentially going to destroy them and everything they hold dear. And so I'm wondering if that is a stronger foundation against which to build resistance. Like, look, if you're not even going to try to convince me that if you're just, this is interesting, if you're just going to tell me how much damage this is going to do to the world, why do I have to figure out how to, you know, to integrate this chat bot into my life? So I'm hoping there's a, we're going to be quicker to minimalism on these tools now than maybe

we were in the past. We're a little bit more cynical and they're not even trying anymore to convince the user that it's important. It's as if their only audience is the venture capitalist. Look, we're inventing the future. That's what matters. Oh, by the way, for the rest of you, yeah, it's going to make all your lives miserable. But hey, what could you do? It's the oddest thing. I've really never seen something like this before in technology. Yeah.

Well, there's a lot of use cases that I think the tech world is like, well, we know the use cases. We don't need to tell all you plebs about the use cases yet. And you see this, it was like a couple months ago. And you see people be like, well, I sat down and started using this for coding. And suddenly I was up for four hours. And you wouldn't believe what it is. And I'm always like looking at those comments. And I'm like, what do you mean?

What did you use it for? I don't understand. I think that for most people, for me, it was like, okay, this is a faster, better Google at times, right? Because, and Google was like getting, you know, it's getting a little better now with Gemini, I guess. But it was getting really, really bad, Google. And there was an shittification. And then I think when the chatbots came, you're like, okay, well, I don't want to have like a conversation with this chatbot. but if it can pull information together faster and collate it faster than Google can,

then like that's somewhat useful. Beyond that, I do think you go down a pretty dangerous road pretty fast. I'm with you on that, by the way. And also, Google's just going to win that battle, right? They're like, okay, we have the time, we have the resources, we have the hardware capability. Like basically chat TPT and Google, just Google now, Jim and I will get better. And that will be our new Google experience. Yeah.

And sure, that'll, that I suppose that will be. Okay, but you're right. I think the engineers are super excited about coding harnesses. The problem is most people aren't engineers. Right. And they keep extrapolating from, I can do this like really technical, nerdy thing in these sort of nerdy ways as much cooler and better. And I can now build my like JavaScript interfaces with JSON connections much quicker than before. Why aren't the rest of you excited? And the rest of the people like, I'm happy for you, but honestly, I don't care. I'm glad I don't, there's only so much I want to think about computer programming. I need to go back to. my regular life. So the flip side of this is, you know, you've talked about an AI chief of staff model that could absorb the email and slack noise. Like, is there any version of AI you think that could actually make focus better? Are you optimistic that we might get one? I hope so. My concern is the current trajectory that the main frontier companies are on is probably not the trajectory that's going to get us there. So if you look at the frontier AI companies right now, They've gone all in on the digital brain for any AI system should be a large language model.

That's good for some things. It turns out to work okay for computer coding because, I mean, LLMs can produce computer code very well, but to build these harnesses around them to actually figure out what code to produce and do multiple steps, they actually had to do a lot of sort of old-fashioned computer coding. If you look inside the harnesses for like Claude Code because that code was leaked, you see there's a huge amount of just old-fashioned pattern matching and if-then statements and hand-coded logic that's just years of experience of developers figuring out how other developers code. But when it comes to other types of jobs, right, when it comes to why can't AI answer my email,

which is a New Yorker piece I wrote last year, like why can't answer my email? Why can't it handle my calendar for me? Using an LLM as a brain is proving to actually be a bit of a broken strategy, right? Because if you ask an LLM, what should I do in response to this email? What is my best option here? What plan should I follow to act on this request that just came in? An LLM is actually not the right brain to use there. It will write a plan that sounds like a reasonable plan.

But because it's created by auto-regressive token generation, not with a world model, not with guardrails, not with a planning engine that can actually test different plans and see which one match certain constraints, the plan it comes up with is just as likely to be disastrous as it is to be the right plan. So I'm a big believer that really what we're going to need, and we'll see this in the next decade, is more narrow AI systems that are built for particular use cases that are going to be modular in their architecture. You're going to have multiple different types of models in there.

Some of them will be machine learn neural network style models like language models. Some will be old fashioned human program models. Like if you look at stockfish that plays chess at the expert level, there's a lot of just hand-coded human AI in there. We're going to simulate trees of different moves and evaluate which one is going to go better. it's not as sexy as the pitch that if we can just get the GPT7, it can power everything ever. But that pitch is really not coming to fruition. So I think we're going to have a much more diverse array of AI systems in the future. I sometimes call this distributed AGI where there's no one brain that does everything,

but we just look up one day and say, you know what? Almost everything I can think of, there's some system for that activity that does it really well. That's a future I'm actually somewhat excited about. I mean, I think it really could improve a lot of things. And it sidesteps a lot of the concerns people have about just building one giant frontier model that tries to power everything. Some of the work product you were talking about that seems like it's great, but it's just crap. You know, the Harvard Business Review piece you say calls it Work Slop, which is just AI generated work that sort of masquerades is good work. Another study, I think, coined the term Brainfry.

And this is just like AI is making. a lot of white collar work just sort of feel worse and not better. Why do you think some companies are still mandating it? What is what is the gap between the productivity story that executives are selling and talking about and the experience that a lot of employees who are using this are actually having? Well, I think it makes executives feel smart. I think it makes executives feel like they're forward thinking. I think it makes executives feel like they're on the waves of the future. I can't tell you how many conversations I've had with very high-level people at organizations that are telling me stories about either what they're doing with AI or what they

imagine they're going to do with AI that completely misunderstands how language models work or what they can even do. They've written technological fairy tales that they're now trying to live out on the backs of their actual employees. I don't think there's deep engagement in the reality of these technical tools. Maybe because you can interact with them in natural English, it feels like you have your arms around them. We have very charismatic leaders at these tech companies talking about them. But I think that's a lot of the push that's going on. And the reason why we get things like work slop and brain fry, it's not AI didn't cause this problem. It's actually a problem I've been writing about for a while now. Like in my most recent book, I used the term pseudo-productivity,

which I think is the defining feature of modern knowledge work. And my argument in that book is that when we transitioned from primarily having an economy on industrial manufacturing to actually growing out a knowledge sector in the mid-20th century, the big issue we faced is we didn't have clear productivity measures. When I ran a factory, I had clear productivity measures. It would be items produced per paid worker hour. Knowledge work was messy, right? People would do many different things. It was unclear who was working on what, and often the individual knowledge workers had more knowledge about what they were doing than the middle managers that were above them. So because of that, we had a lot of autonomy in knowledge work, but we didn't know how to measure productivity. So we fell back

on this heuristic as pseudo-productivity where we said, we'll just use visible activity as a proxy for a useful effort. The more I see you doing, the more I'll assume that you're doing useful things for the company. And this reality, I've been writing about this for a decade now, has caused all sorts of problems. This was the reality that when we brought email into the office, we ended up where we are today where the average person checks as an inbox or a chat channel once every three minutes on average, right? Because there's no deeper productivity metric that we can measure that says, actually, we're producing less good stuff if you're checking email all the time. Without, you know, in the absence of that, it's going to be the business. So AI just feeds into the

pseudo-productivity reality. If we, if it's a hard to have a number that says, here's how many widgets you produce, an activity is what looks good. Well, of course, I'll just use AI to accelerate the activity. I can write more emails. I can send out more PowerPoint slides. They're sloppy. They don't help anybody. They don't make more money for us. But it makes me look more active. And Suter productivity says that makes me better. Offline is brought to you by Mint Mobile. I don't know about you, but I like keeping my money where I can see it. Unfortunately, traditional big wireless carriers also seem to like keeping your money, too.

If you're fed up with crazy high wireless bills, bogus fees, and free perks that actually cost more in the long run, then switch to Mint Mobile. Stop overpaying for wireless just because that's how it's always been. Mint exists purely to fix that. Mint Mobile is here to rescue you with premium wireless plans starting at $15 a month. All plans come with high-speed data and unlimited talk and text. delivered on the nation's largest 5G network. Bring your own phone and number.

Activate with ESIM in minutes and start saving immediately. No long-term contracts, no hassle. Ditch overpriced wireless and get three months of premium wireless service from MintMobile for $15 a month. You can be saving so much with MintMobile. You're spending too much on your phone bill. You can see the bills every month and you're like, what am I doing? Go check out MintMobile. It's so much better.

If you like your money, MintMobile is for you. Shopplans at MintMobile.com slash offline. That's mintmobile.com slash offline. Upfront payment of $45 for three-month, five-gigabyte plan required, equivalent to $15 a month. New customer offer for first three months only, then full-price plan options available, taxes and fees extra. See MintMobile for details.

Get road ready with Carquest Auto Parts. Save now on premium brand-name products. Castro Magnetech synthetic oil, 5-liter jugs, 297. Rain-X summer breeze washer fluid, 3.78 liter jug 397. Air 1 diesel exhaust fluid, 9.46 liters, 13-9. 97. Seafone motor treatment, 473 milliliters, 897. Hit the road with CarQuest for the parts you need. Visit a Carquest auto parts store or Carquest.comst.com for more information in full flyer details. Only at participating locations.

You know, we talked about your piece sort of frames the cognitive fitness challenge as the next diet and exercise revolution. That fitness resolution has certainly improved people's health, led to longer lifespans. it's also led to some quack science, questionable supplements, grifty influencers. We have a vaccine skeptic in charge of health and human services now. What do you think the dark side of focus culture could look like and how do we avoid it? It's already here. So there is already a whole substrate of grifters and quack scientists and influencers that sort of circulate around ideas of focus or technology, or let me sell you this tool to solve

this tool, or you need fancier technology to make your technology more tame, or they'll take an idea like deep work, which is a very basic idea that actually the best way to use the brain if you're producing hard thoughts is don't context shift. You should keep it focused on one thing. And you'll get these long Twitter threads about how you can do 15 hours of deep work straight if you're just in the right sort of drama. So, yes, I think it's inevitable that if there's going to be any sort of culture-wide pushback on something, any sort of revolution, it's going to take a lot of actual literacy among the people engaging with this revolution because you're going to get a lot of the quacks.

But I'm okay with that, right? Like the health and fitness revolution, we have a long way to go. But by all measures made a fantastically big difference, both in terms of what happened with people's individual health, but also the infrastructure around health that got built up around that revolution happening. We forget the degree to which before this happened, if you were a man in the United States, you just sort of assumed you probably would die in your 60s. And if you were lucky, you want it. And that's just kind of the way it went. Like, maybe I'll probably have a heart attack, 50% chance it'll kill me. Like, that's just the way

we went through life. And we don't think that way anymore, right? So there's been, there has been massive improvements and a lot of health and fitness grift that has come along with it. But I say that's probably just a price to pay. Once you start carrying about something important, you're going to get a lot of nonsense as well. So there's individual behavior changes, you know, keep your phone in the kitchen, read more. There's systemic changes like, you know, the social media bands, the phones and school bands. It seems like any cognitive fitness revolution would require both individual and systemic change. But I'm wondering what your view is on which is more important. if you had to put your weight on either individual change or systemic change as sort of the primary

lever, what would it be? I mean, up to this point, I've been putting my weight on the individual change because that was the only lever I could actually pull. In 2019, when I was promoting that book, Digital Minimalism, this would be the number one pushback I would get, especially if I was doing something like NPR or something like this, is they say, yeah, but what about the systemic changes at the organizational level or at the government regulatory level? And I was like, yes and yes, but there's no traction happening there right now. This is something we can actually. Let's focus at least on something we can do,

and the individuals can make a massive difference, right? So I've long put my energy into that. More recently, I've had some hope, though, that now we're able to conceive of changes at other levels as well. And again, I break it out between like the organizational levels, so we're thinking school districts and the government level.

Now we're thinking about legislation or regulation. Both of those are on the table now in a way that they haven't been as vigorously in the past. I think the school phone ban was one of the first wide-scale organizational changes we've seen, and it's really having a big difference. So that gives me hope. I think the Australia's social media ban is interesting, right? I mean, there's mixed results about what's happening over there,

but that was major regulatory activity, right, that was aimed at exactly this problem. I think we have a growing amount of research that's talking about what happens in the brain, especially in the centers we would associate with addictive behaviors when we're using some of these technologies. We have a past history, especially when it comes to kids, with how we deal about, deal with this through a regulatory lens, what suddenly is seeming relevant again. And, you know, as a Georgetown professor, I spend my fair amount of time on the hill, talking to senators, et cetera, like what's going on, briefing on things I happen to know about. And this is all on everybody's radar.

And so this is a shift for me. So I think we need all three. I've been pushing hard on the individual because that makes a big difference and we can do it right now. but for the first time since I've been on this issue, I'm very interested in what's happening at these higher levels as well. I think with the current administration, it's obviously going to be very difficult to pass any sort of meaningful legislation or, and I don't count on them for any kind of regulation. But I do think as we head towards 2028, this is an issue where, you know, I could see,

I mean, I know that the Democrats will be talking about AI and focusing on AI, but I could see Republican candidates, maybe it's doing the same as well, depending on who it is. It's a confusing issue politically. I, you know, I worked on a paper with a student here at Georgetown where he was tracking the state-by-state legislation on phones. And it's a very mixed partisan map. It's confusing bedfellows. It's really hard to get your arms around. You'll often end up in the same place, but with different types of issues that you're focusing on, whether we're talking about a red state governor or a blue state governor. It's one of them or politically complicated issues that I've encountered.

So I am very interested to see. I think the state level is interesting right now. You're right. I think the federal level, Silicon Valley just has an incredibly strong lobby. It's just very difficult at the federal level. It's such a part of the economy. We're not Europe. It is really hard at the federal level.

But I'm very interested in the state level, the experimentation that might go on. And I do think it's going to be a big 2028 issue for sure. So a lot of our audience is probably listening to this and thinking, this is great, but my boss expects me to answer Slack on weekends. My mom's on Facebook. My kid's school sends notifications through the app, the pictures. I keep in touch with my friends on Instagram. Like the constant companion model of the phone isn't really a choice anymore. It's sort of the substrate of modern social and professional life.

For someone who can't just opt out, what does a realistic version of a cognitive fitness plan look like? Well, I mean, first of all, I want to separate away the first thing you mentioned, which is my boss's expectations on Slack from the second thing, which is like I'm keeping up with my friends on my phone. These are two completely different sources of two different forces that are pushing us towards our phone. If we look at like what's happening with my phone and my life outside of work, it's actually never been easier to have much better digital hygiene because an important shift happened, right? When the major social media platform starting around 2023 began to chase the TikTok model of completely algorithmic curation for engagement, they gave up their number one competitive advantage, which is the social networks. And what happened is we saw more and more of our actual digital social sociality move off of these social platforms and onto things like chat. So it'll be in group messages. It'll be in WhatsApp. That makes it very easy to take Instagram off your phone, to take Facebook off your phone, right? Take TikTok off your phone because it no longer has that hook into an important part of your social life. It's become just pure entertainment. So I would first

say, make that phone less appealing. I think that's really important. Two, you have to get regular extended amount of time where you don't have that phone within reach. Because when it's within reach, you're firing up all those short-term motivation circuits that have learned, this is a very clean and consistent source of reward, which means it's going to vote for the activity of picking up that phone constantly, and you will be constantly trying to apply willpower to do almost anything else. So one of my listeners calls this landlining now, but having your phone plugged in when you're at home. And it doesn't stay on your person when you're having dinner, when you're watching a movie, you're reading a book. It seems like a minor thing. But it's a massive relief for your brain

because with the phone not being with you, that short-term motivation circuit quiets. And you feel suddenly like you just took the limitless pill. Like, oh my God, I can do all the. What's going on? Why can I focus? Why do I feel like my ability to think and concentrate as highers? Because that's not with you. The work issues are more complicated. The reason why you have to answer your Slack messages or check email all the time is because we have implicitly adopted a collaboration style based on unscheduled ongoing back and forth messaging. So if you're not checking, the stuff is going to fall behind. That's a whole other issue. And that's solved by putting in place alternative collaboration styles that are not based on unscheduled messaging. That's a

hard problem. I've been working on that problem for a long time. It is a really hard problem to shift. So I will put that aside. But your phone and your life outside of work, you can make it less interesting and you can keep it out of your physical proximity for longer parts of your day. None of this is about a sort of extreme abstention, but it really can make a big cognitive difference. I wonder whether you think there's a class element to this. You know, a lot of private schools, schools in wealthier districts, ban phones, limit devices. People working two or three jobs or seven days a week might not necessarily have the extra time to work on their cognitive fitness or pay attention to their children's.

do you worry that we're living in a world where the ability to concentrate is stratified by income or education, or has that already happened? It's going to become more stratified, right? And it's not great. We saw something similar, obviously, happened with physical health. If you had the means financially or the time affluence to be able to exercise, to be able to think a lot about what you eat, to actually have the time to sort of be careful about your diet, you get much better outcomes.

And we got economic stratification of health outcomes really clearly. We're probably going to see similar stratifications if cognitive fitness becomes something that we take seriously. I think we're going to, again, see stratifications there as well, which are going to exacerbate inequality. I mean, basically, we're in a place now. It's not like this is better. We're in a place now where just everyone is so cognitively unhealthy, it's all the same. So it's only once we really start caring about something that you begin to actually get that differentiation. I think it's a problem.

And I think that is going to be a problem. And I think especially from a policy perspective, it's something we need to think about. That's why things like school phone bans are important. That's why having the educational system prioritize cognitive fitness, right? Like one of the things you're here to do is not just to do well on content-based tests, but we want to help prepare your mind to be comfortable with contemplation, with putting the target of your attention on an intentional target and holding it there and not having that feel really bad to be able to take ideas and manipulate them in your head, to push on them to understand them, that we want your brain to be strong.

Just like if you were coming to Marine boot camp, we would say, we want you coming out of here physically strong so that you can actually hike with a pack for 20 miles and you would be okay. The more we can do that in these sort of more universal institutions like public education system, I think the better.

And so we're probably going to get stratification. The inequality is going to be something that's glaring and we're very uncomfortable about. And hopefully that will push, now more of the sort of systemic fixes as well to try to close that gap a little bit. But I don't see how we're going to avoid that. Coming back to where we started, you know, my boys are five and two. Clearly their school is already thinking about this.

What do you think kids need right now in terms of a formative experience for their cognitive development that they're not currently getting? They need to be protected from unrestricted access to the internet, right? Until a relatively late age, you really should not. have unrestricted access to the internet. That really is like we keep the liquor cabinet open and there's the cigarette jar go at it, right? It's the developing brain really cannot react well to that type of hyperstimulus that you get when you just have unrestricted access to the internet. So that has to be something we need to think about with kids. We also have to prioritize

activities that do cause some mental strain. I think reading is the fundamental cognitive calisthenic, right? The whole post-Pylaolithic experience is defined around the brain that was formed by reading. Before we had literacy, it was a different brain. A brain that literally was not capable of the type of abstract symbolic thinking that leads to all of the sort of symbolic institutions, again, whether it's politics, philosophy, or theology, required reading before our brains could even do that. So that has to be an important activity. Yeah, it's boring, but it'll become more interesting and it's making your brain stronger. So, I mean, I think this is, these are the two things. Just like right now we're realizing, okay, kids probably shouldn't just have two liters of soda and pure junk food for all their meals.

We see like this causes problems. Unrestricted internet access at, you know, pre-high school age, we see that as a problem. No engagement with reading, which again, I privilege. I privilege that particular activity, reading and writing. We have to privilege those of like this is what's required to prepare a brain that matches the context of the sort of world that we live in now. So reading and no unrestricted access to the internet, whether it's through an iPad or a phone or unrestricted Chromebook that they can just keep in their room. To me, these are like the two pillars that we need for youth right now. Cal Newport, thank you so much for joining offline. And thanks for focusing on this. A whole bunch of excellent books you should read from Cal.

Slow productivity is the most recent. World Without Email, Digital Minimalism. And of course, deep work. Thanks for coming on offline. I appreciate it. I enjoy me. Thanks for having me. Take care. Offline is a Cricket Media production.

It's written and hosted by me, John Favre. It's produced by Emma Ilich-Frank. Austin Fisher is our senior producer, and Anisha Banerjee is our associate producer. Audio support from Charlotte Lambs. Adrian Hill is our head of news and politics. Matt DeGroat is our VP of production. Jordan Katz and Kenny Siegel take care of our music.

Thanks to De Lawn Villanueva, Eric Chute and our digital team who film and share our episodes as videos every week. Our production staff is proudly unionized with the Writers Guild of America East.