Ted Audio Collective Hello everyone.
You're listening to After Hours.
I'm Felix. I'm Mir.
I'm Sarah. Hey! Sarah's back.
Good to have you, Sarah.
It's so great to be back.
Thank you guys for having me back on.
Of course. Mir, I actually had a question for you with all this excitement around technology and AI.
I know you've often confessed to being a light of sort.
Are you softening up?
Yeah, you know, that's in part why I wanted to talk a little bit about technology today, because I'm trying to figure it out, Felix.
Oh, okay. So, I actually thought we could begin by talking a little bit about ChatGPT.
We had a fun exchange about it at the end of the year when it first came out.
We had that poem that it wrote for us.
Yeah, you know, the last three months have been kind of dominated by it.
And we haven't really tackled it fully.
So, I thought I'd love to get your thoughts about that, in part, Felix, to figure out the answer to your question.
Good. Okay. What did you bring, Sarah?
Do you have a topic for us?
I do. It's also related to technology.
So, unlike me here, perhaps, I have been sort of a proud early adopter of a lot of technologies.
And I actually find myself, perhaps, now moving in the opposite direction and asking, where can I prune back?
Where can I unplug?
So, I would love to talk with you about digital detoxing.
So, this episode is really about technology and self -discovery.
There you go. I like it.
So, me here, AI and ChatGPT in particular, what's in your mind?
Yeah, it's just remarkable what's happened in the last three months.
So, since ChatGPT was released, which is effectively an AI chatbot by a company called OpenAI, it has just become remarkably dominant as a consumer technology in a way that we have not seen since the advent of at least the iPhone or perhaps before.
So, it has become pervasively used and maybe even more just talked about as this remarkable moment in technology.
To give you a couple of markers for that, first, of course, it impacted the world of search.
We had both Microsoft with Bing release a version that was powered by ChatGPT and then Google really stumbled with their version called Bard, wiping off tens of billions of dollars in market capitalization because of a small difficulty they had.
And then, of course, you have had it rear its head in weird ways, which is there was a famous New York Times columnist who had a very odd interaction with ChatGPT, which was vaguely hallucinatory in some ways where the AI chatbot tried to convince him that he didn't love his wife.
And then you have had pronouncements that range from someone like Jensen Huang, who runs NVIDIA, a very important technology company in the world today, saying that this is the most important technological change since the advent of the internet and the iPhone.
And so I'm just curious about how transformational this is and where you think the transformations will be.
So, Sarah, what do you make of all this?
I go back and forth on it, but fundamentally, I guess the question for me is, how quickly will I lose my job?
I'm sort of laughing, but also just to hide the pain.
I wrote a column and it took me a couple of days to do the research and then to write it up.
And then I thought, I just want to see how much faster ChatGPT could do and how good it would be.
And so I prompted it to write a draft in the style of Bloomberg opinion about the topic, which was understaffing.
And it wrote a very prosaic 800 word column immediately about why understaffing is bad.
And it even used some of the same examples that I had used in my column.
It wasn't very entertaining.
It was a little bit dull, but I thought if this is what this thing does at basically the moment of birth, what's it going to be doing five years from now or 10 years from now?
I feel like we are seeing the dawn of something really new.
And any solace you take right now from the fact that it's not always very good at what it does is sort of a false hope because it will rapidly improve is my prediction.
Right. And just to be clear, Sarah, some of those problems will likely get conquered over time.
It can be inaccurate, for example, it seems to stumble on math in a weird way.
So there are definitely certain things about it that are limited.
Well, and I asked it to be funny.
Like the first draft was so boring.
I was like, could you do that again?
But this time make it a little funnier.
And it was so completely unfunny that I was like, OK, as long as human beings can get paid for being a little bit funny, maybe I'll hold on to some income.
One of the things that is really interesting to me is whenever a new piece of technology comes out that feels really new immediately, we jump to this point of how similar is it to us?
So humans are just really self obsessed.
The first question we want to know, oh, push the boundaries, try to see where it can go.
How much is it really like a human person?
And that's in some sense, particularly funny because irrespective of what the answer is, if you think it's very far away from being human, we're alarmed because it's full of misinformation and it's misleading and it's terrible.
And then if it's pretty much like human, we freak out as well because it's oh, my God, what is happening to my job?
What will I do in the future?
Because it's so human already.
And it strikes me as maybe something that is first and foremost just fascinated by the capability of the technology, but not really thinking through the core uses of what it will be able to do.
And we're all thinking of these fancy uses of it.
But really the core of what it's really good as is doing dull work, that same customer letter, the response to some complaint, reminding someone that a bill is to asking your students to sign up for a particular task.
All of these not so interesting things.
I would predict that that's actually the core use.
And this is big because if we don't have to do dull things, we can do more interesting things.
I confess this whole question has really flummoxed me of how to make sense of this.
The first thing that's helped me is to understand a little bit more about what it really does.
So I think what's important to understand is it's a little bit of what's called a large language model.
What it is really good at is basically stringing together words and then figuring out what the best next word is.
And it just combs through its universe that it's been trained on to figure out what the best next word is.
And it effectively puts together sentences in that way.
I think that's a useful way to understand what it does.
It really is just doing that.
And as a consequence of that, it won't necessarily be terribly creative.
It will be somewhat derivative.
But to Sarah's point, I think it drives the cost of a lot of content creation down to zero.
And that is kind of profound.
Now, I don't think it comes anywhere close to a Sarah Green Carmichael column, but the vast industry of PR, the vast industry of a lot of financial journalism, sports journalism, where you're just taking scores and putting them into sentences.
Legal writing? Maybe some legal writing, maybe some interesting applications in teaching.
And certainly there are interesting questions for the education industry generally.
But beyond that, I'm not sure how transformational it is.
So all of those things are important.
But are they fundamental changes to productivity, for example, in large chunks of the economy?
What are the real use cases?
So I get driving content creation costs down to zero is powerful.
But tell me how an industrial producer of widgets is going to change what they do and how it's going to change productivity in the large sense of the word.
And that's where I get a little bit stuck.
My fear is that by driving content creation costs to zero, it will spur the creation of vast new oceans of terrible content.
Novels that are sort of wooden PR pitches that are somehow even worse than the ones that we already wade through, emails that are longer than they need to be and more numerous.
And so it's not like we have a shortage of content now.
And now there's so many shows you can't keep up with them all.
And there's so many emails you can't even read them all.
And so we're going to need an A .I.
that can actually also ingest for us.
That will come through our whole email inbox and say this is the one interesting pitch from a PR person that you should pay attention to today.
So it'll be like robots talking to robots because as it is now, I feel really the limits of human comprehension on the amount of information to process every day.
So the idea that there would be the creation of more information for my brain to handle is sort of intimidating.
To me, it sounds mostly like a shift in the division of labor between machines and humans.
So say, for instance, in your story, Sarah, if for some reason human humor is just really hard to replicate, then your task is just a different task.
You get these dull columns written by Jet GPT on something that you want to write a column about.
And your skill that is really valuable and that is interesting is to make dull columns into engaging, funny columns that people will really like to read.
One of the things that these models do is as you interact with them, the most recent interaction has greater weight than sort of the corpus of knowledge that exists in the first place.
So, for instance, this is, I think, the reason why we get all of these funny or not so funny accounts from journalists about, oh, it was really angry or it was really depressed or it fell in love with my wife.
Well, because it mirrors the use of language that you have in these conversations.
And the Microsoft fix for now is just not a very long conversation.
But that will always be a problem.
So if you ask it to do something complicated that it doesn't naturally get right, then the increased interaction will give you more of what you already know, who you already are, how you already speak.
And so undoing that kind of a bias, that might actually be what humans will have to do.
But we start from just a very different point.
I think that's exactly right.
I do think one area where I could see it driving really fundamental change is actually education.
So content creation and journalism and all that stuff, I think a lot of journalists are interested in.
But in terms of the economy is not necessarily that important, frankly.
It's important to our society, but it's not necessarily important to the economy.
You could imagine, I think very easily a world where you have personalized tutors and you have a person who is basically interacting with an avatar and having a conversation based on something they need to learn more about, much like you would do with a tutor.
And that to me strikes me as pretty transformational.
So that whole class of activities and then education more broadly, where we really try to help people learn to think and learn to write, that all strikes me as completely transformational and interesting.
Yeah, I think what you're saying me here is right.
And I see all these articles from educators freaking out about, you know, how are we going to clamp down on cheating if these essays could be written by chat GPT.
And I actually think what we might see is more oral exams, for example.
And I think in some ways our society is already pivoting away from writing and much more towards speaking as like a way of influencing if you look at the importance of video now.
And so I think that as writing becomes more and more of a commodity, something that machines can do, I actually think we might see a return to sort of, hundreds of years ago, what maybe education looked like when it was much more about how well you did at speaking and orating and debating.
Trying to limit the use of chat GPT in educational settings strikes me as completely wrong.
That's like saying we're not allowing people to use calculators.
We educate in an environment where we pretend calculators don't exist.
That's insane. What we want is that people get really good at using these tools.
And as a result, maybe the starting level of, say, a first grader will be very unlike what the starting level is right now, but that's okay.
And then we can spend more time on be critical about the results.
How would you recognize the mistake and all these kinds of things?
Yeah, I totally agree with both of you.
I would just disagree maybe on one thing, Sarah, which is to me, writing is thinking and thinking is writing.
They are inseparable in my mind.
When you learn to write, you're learning to think.
And so I hear you on the oral tradition and how important it is.
But I think there's no denying that if we are no longer teaching children how to write, then it's hard to know if we're teaching them how to think.
I am so glad you said that because I completely agree.
I don't even know what I think until I figure out how to write it.
And then on the flip side, in terms of taking in information, it is much easier, I think, to be a critical reader and find the flaws in someone's argument if you are reading.
Then just listening to some people talk, it's very difficult to detect BS when you have a really smooth person just chattering at you.
But if you see a written down, you know, I think it's much easier to detect.
So I'm so glad you mentioned that point.
Let's talk a little bit about search because this has the possibility of really upending the search market, as we've already seen.
So Bing, which you had probably forgotten about, is back.
And Google, in its initial efforts, appears to have stumbled.
And Nadella has made comments about really wanting the search market, which is, of course, 90 plus percent of Google's business.
And really maybe taking that search market and devaluing it in a way.
Because ChatGPT right now is not associated with advertising.
So it would be interesting to think about how you would monetize search that's driven by ChatGPT.
Do you think of all this as fundamentally transformational to Google and the search market?
My sense is that these claims about the end of Google dominance are completely overdrawn.
And I think that's true for two reasons.
The first is, let's not forget that the core of OpenAI's technology is Google technology.
It's built on advances that Google has made.
And yes, so, power wasn't the success right out of the gate.
But will they get it as right as OpenAI can get it right?
Yes, of course. So we will have similar functions.
And frankly, even today, you remember when you search on Google, you often now have a little paragraph that gives you an answer.
Right, right. Which is not interactive in a ChatGPT sense, but it's also generated by AEI taken from web pages and so on and so on.
What I completely don't understand is how this is supposed to revolutionize the ad market.
Even when I look at the examples on Bing's website, you could for instance ask it, what's a fabulous three -star menu?
Or what's a special trip that I could take for my anniversary?
Well, guess what kinds of ads are going to be placed right along these suggestions of that trip?
So in a way, I think the sometimes lazy thinking is that the ads themselves are substitutes for search.
But no, actually it's consecutive.
Most of the time you search for something and then we serve you ads that invite you to take a particular second step, mostly having them to do with commercial intentions.
And I don't really understand all these claims that it's going to undo or somehow transform the market for advertising.
Well, let me try, Felix.
What if Microsoft, which has remarkable revenue streams from Azure and from lots of other businesses, decides they don't want to make money in search?
What if they don't care?
Then doesn't that really damage Google's search business?
Because Microsoft won't sell ads against chat GPT output.
So let's start by not forgetting Bing's global market share is 3%.
So lots of things have to happen until the rest of humanity finds Bing in the first place.
But even if they do, I think if their stance is we're not going to serve ads, which actually even now they don't do.
So if you look at the model chat GPT conversation that they have on their website, it's about the food and then sure enough you see links to all the sites where you can book restaurants and so on and so on.
So already that's not really consistent how they're using it today.
And then in the organic search, of course, you have all the usual links to more commercial websites as well.
But I think the most dangerous thing is not if they don't sell ads.
The most dangerous thing if they provide advertising opportunities and they give it away for free.
How likely is that?
Microsoft right now is $11 billion in advertising revenue from Bing.
If you could say, wow, now Bing is a little better than it used to be and it could be $100 billion.
Are you really going to say that's $80 billion that we're not going to touch?
I don't know. Fair enough.
Sarah, what do you make of all that?
I guess I think that if it lights a fire under Google to make its products better, then that fundamentally is a good thing.
I feel that there has been some drift at Google as they have this sort of cash cow in search, but actually I feel like their search is often not very good.
So I think if the existence of a stronger Bing makes Google actually take this seriously, then hopefully consumers will benefit because I do think that there has not been that much competition in search for a while and I think that actually consumers and end users are the ones who have suffered the most.
What do you think, Mihir, do you buy the claim that this is about demonetizing search?
I'm actually kind of open to it.
I usually think the search market is going to get much more interesting than it was before and it's not just Bing.
These technologies can be deployed in all kinds of ways by all kinds of potentially quasi -search -like bots.
So I won't go to Google.
I will go to a specialized food AI bot, which is super terrific and I will use it there.
And so you could see this kind of lock on global search that Google has, not just get eaten away by Bing, but just get fragmented and splintered.
There's going to be a lot of bots that are going to be very good at doing very specialized things.
That, to me, I think is threatening to Google because they have effectively had one stop for global search.
And maybe that gets fragmented and broken up because you're able to get a really good bot on particular domains and you go there instead.
So that's not about competition from Bing in the global search market.
It's about the fragmentation of global search.
That, to me, is interesting and worth just understanding what will happen to this.
I'm curious maybe just to wrap this up, how do you compare these AI innovations with previous technological changes?
So they're being compared to the advent of the internet.
They're being compared to the launch of the iPhone.
It's being compared to really massive technological changes.
Is this something that transformational or is it relatively marginal?
I think it's transformational.
I think that it will reshape a lot of our behavior in ways that we simply can't predict now.
I think when the iPhone came out in 2007, we didn't know that there would be like an Instagram and then a TikTok.
We didn't know that there would be a mental health crisis among teenage girls who were spending too much time on those apps.
There was so much we couldn't have predicted.
But I do think it's transformational.
Felix, what do you think?
That sounds exactly right, Sarah, in the sense that many daily activities, the way we search for information, that maybe the way we consume or produce information, I think it's going to change.
Where I'm more skeptical is about its impact on productivity.
Even for computers, it took us three decades to see the impact of PCs on the economy -wide productivity.
In part, it has a little bit to do with the shift in the division of labor.
If it just so happens that we do things a little differently, and now my main job is to add humor to columns, for me as an individual, that's a very different experience.
My job is very different.
What I do is very different.
Ultimately, is this going to be so much more productive in the sense that it takes far fewer inputs to produce something that people will love to read or that I can do 15 at a time and I wasn't able to do that before?
I'm more skeptical.
I think, in particular, in the early phases, there's almost no chance of this being really productivity -enhancing because it's just so messy, but even longer term.
I think we will do things differently, as you point out, but we won't be more productive.
That sounds right to me.
I think that's the key thing, Felix, to me.
By the way, that concern about the link between technology and productivity is much more widespread than just chat UBT.
In the last 10 or 12 years, remarkably, despite our obsession with technology, productivity hasn't been manifesting that.
Unlike the revolution from the late 1990s and early 2000s, where you really had massive productivity gains from technological changes, we just are not seeing it.
This feels like more of the same.
I think that's something we don't talk enough about, Felix, which is we don't talk enough about this broken link between technological change and productivity improvements.
That's why it is both completely transformational and amazing, and yet somehow seems oddly marginal.
Yeah. Because in a way, it's impacting consumption more than it's impacting production.
Yeah. It's impacting our consumption habits.
We're all getting so much better at consuming, but I'm not sure it's really manifest in ways of productivity measurements that are actually what drives wealth creation.
So that's, I think, my way to kind of understand both pieces of it, at least after talking to you both about it.
Mm -hmm. Okay. So this entire conversation, of course, was the output of a chat GPT.
Yes, of course. How else would you explain all the terrible predictions, all the misinformation that was conveyed?
That can only be an AI conversation.
What's the problem with that?
All right. I think we already need a digital detox, Sarah.
What do you make of this trend towards digital detoxes?
So as far as I can tell, the sort of recent upsurge in talking about this came after Marc Benioff, the CEO of Salesforce, was reported to have taken a 10 -day trip to Polynesia to digitally detox.
And this, I think, has sparked a broader conversation about to what extent normal people who are not CEOs with assistance and billions of dollars can afford to step away from their devices.
And I think it's also just come at a time when people are feeling overloaded by new technologies and thinking, wouldn't it be nice if I could do that kind of thing?
So I'm curious to know, just to start, have either of you ever tried to take a digital detox for any period of time and how did it go?
I've never really taken a digital detox over, say, several weeks or anything like this.
But what I have done is that I have a particular set of activities where I never take my phone.
So I don't take my phone to restaurants.
I don't take my phone to the gym.
For leisure kinds of things.
I exclusively read books in paper formats.
And I'm trying, in part, to balance a little bit because I do find it has an effect on me.
If we're a group of friends in a restaurant and none of us have the phone with us, it's a different conversation.
No one feels compelled to look at, oh, yeah, that ancient Greek god of something.
What was he exactly called?
And then all the phones come out.
And of course, it's completely useless and nonsensical because who cares what that particular god was called?
So the quality of life is different without digital devices.
That's interesting.
Mihair, do you also take this mindful approach to leaving your phone behind?
Felix, I think it was Aries you were thinking about.
I just looked it up.
I mean, my general view of this, similar a little bit to Felix's is any chance anybody has in any way to find respite from their devices is to be applauded.
I don't judge anybody on any of this stuff.
Is it going to have long -term effects?
Probably not. But I'm at this point where I'm very open to any mechanism anybody has to manage their devices in more helpful ways.
And if a detox works for you just a week without it, I say power to you.
So I'm becoming very open -minded about this.
And I think it is tempting to kind of ridicule people who take these things either as being privileged or as being ineffective.
But the reality is I'm open to anything at this point.
What do you make of it all, Sarah?
Ten years ago -ish, there was a real mania over the idea of the text Shabbat that once a week you would take just a day away from your devices.
And I feel like this allure for this more intense detox that lasts for a week is sort of a reflection of how much more enmeshed these devices have gotten into our lives.
I can't imagine going away for ten days without my phone to get me on the plane, on the rental car, to find the trailhead that I'm hiking.
I use it for all these things.
So I am much more aligned with Felix's perspective of maybe there's certain times or places where the phone just doesn't go in the course of a normal day.
I have set, for example, time limits on social media apps on my phone so that I don't spend more than ten minutes a day on the social media apps because that's sort of the amount that I'm comfortable with.
I think some of those smaller moves are like a more realistic and sustainable way to approach it because you're right.
I think there are studies of people who've tried to do this.
It doesn't overall improve mood with the exception of people who have depressive symptoms sometimes do report that their symptoms are not as bad if they spend less time on their devices.
But overall, the pressure's really on each of us now to try to manage our own technology use because no one's coming to do it for us.
And these forces are out of our control.
I think in a way, Sarah, what you said is making me think also about one of the problems with the detox formula.
So one of the things that can happen to me mentally or maybe other people is, well, I'm going to be doing a detox.
And so that allows me to do certain things in steady state that I wouldn't normally do.
So by creating these very extreme periods of our lives where we don't do something at all, it somehow can excuse the performance of daily life.
And so that I think is why you're right and Felix is right.
And I just reversed myself a little bit from what I said earlier, which is there is this weird way in which if these longer detox periods give you license to misbehave more during the non -detox period, that's problematic.
And it's got to dominate to do regularized patterns as opposed to binging of detox, non -detox kinds of behavior.
So I think that is really true.
And maybe building on this, Mihir, I think it's also important to think about the underlying issue that you want to solve.
Right. So often we say, oh, I'm spending so much time on TikTok or I'm spending so much time on email.
And I think if the underlying issue is digital distraction, then even the time that you spend on your phone or the time that you spend on email is not really indicative of the severity of the issue to begin with.
I'm reminded a little bit of an experiment that they did a little while ago where students had to solve some set of mathematical problems, but nothing too sophisticated.
And they had different groups.
One group had their phone outside the classroom.
One group had the phone in their bag, but right next to them, one group had the phone right at the top of the desk.
And so you literally see these conditions matter dramatically for performance.
Yeah. Even if the phone is in your bag, the idea that, oh, maybe I did get a text message or maybe I should look at my email, then I think that also makes you rethink a little bit about the basic detox idea that it's a really severe cut from how you relate to your devices.
What I think is right and good about that is that it takes complete separation.
And I think part of the reason that I find that idea of a total separation interesting is because in my own personal life, I do sometimes feel that a hard reset of habits can produce a lasting change.
It's like the dry January effect.
People who do dry January do drink less throughout the year.
And I feel like my own life, if I go on one of these spend nothing challenges, you do notice that you start thinking differently about, do I really need to buy this sweater or whatever?
And I think that with digital devices, it's hard to have that total separation because they are so enmeshed in our lives.
Like, how would people even call you?
But when I am away, for example, in another country for vacation, if I can just put the thing in airplane mode and leave it in my bag more of the time, I do find that that produces some good effects in my brain.
Unfortunately, they only last as long as I'm on the trip, and then it's right back into it when I get back home.
The other element that is really interesting to me is the flip off.
I don't want to be digitally distracted.
What am I hoping to achieve with the help of digital devices or not?
We live at a time when many people feel lonely.
They have difficulty connecting with others.
And of course, technology plays a big role in enabling exactly that, so that I can call people.
I can be on video calls where I feel more of a connection to the other person.
And so I'm always thinking, if we complain about technology, it's sort of a surface level conversation.
So thinking about the kinds of things that you really want, the kinds of things that you would love to cut out from your life.
And then sometimes technology is helpful, and sometimes technology is useless or distracting.
But I don't think it's very helpful to start with technology.
Think about the underlying thing that you really want.
I think that is so helpful, Felix.
I had not really thought about this, but so many of our conversations about technology are really displaced feelings about other things in our lives.
Yeah, that's a great way of saying it, yes.
And I think a lot of it has to do with loneliness and happiness.
And we know so many of these metrics are in decline, especially for younger people.
There's like this epiphenomena of like, oh, we're really spending too much time on technology.
But the underlying phenomenon is just about our happiness and our loneliness and all the things that really, really matter.
And I think that can be distracting because you end up talking a lot about technology and not thinking about connection, which is really what we should be thinking about.
I do think that there is a way in which if we even agree that what we all want is connection, that's all that we're really here on Earth to do.
But I do think that the way that we've set up these devices means that we have to be much more intentional and expend a lot more energy to use them in the way that is good for us as opposed to the way that is easiest.
I have this great idea.
How about if we get together on a weekly basis and bring a couple of ideas and talk to each other about it and kind of hang out?
No, I think that's terrible.
And then we could broadcast it.
Oh, my God, that's even worse.
Who would want to listen to us?
There you go. Okay, recommendations.
Felix, what did you bring today?
I have something that's related to our first conversation about AI.
Yeah. Of course, at this moment in time, it's all about generative AI.
That's the big topic.
But I think there's so many other interesting things that are happening.
And part of what I have been paid a little bit of attention to is text -to -video conversion.
So there are many services now.
The one that I have the most experience with is a company called, I don't even really know how to say it, Synthesia or something like this.
And essentially what it is is you type in a text, a message that you like.
I recently tried something for after hours, whereas I'm pretending to be an enthusiastic listener who really loved the conversation on after hours.
And then it takes them, I want to say maybe a minute or so, when they send back a video with a professional -looking actor that says your text in a completely natural way.
And on top of everything, the background is also sort of a seam that you alluded to.
For instance, for this optimistic listener call, all of a sudden Steve Carell popped up in the background and he had an Oscar trophy.
If you have any sort of video message you would like to send to someone, just think someone's birthday, someone's promotion.
The videos can be really long, they can be up to 30 minutes long.
So if you have, say, a training manual that you need to put on video, it's really amazing.
That's fantastic. All right, Sarah, what have you got?
I have something totally different, which is a crime show.
And I know you guys have recommended crime shows on the podcast before.
This one's called The Investigation.
And it is set in Copenhagen and it is in Danish.
What I really liked about the show was that even though it was a bit heavy at times, you never meet the perpetrator, you never meet the killer.
The focus is entirely on the investigators and the family that has been bereaved.
And I thought that was just such an interesting choice by the director.
And it really made me see some of the other cop shows that I like to watch in a totally different light.
So it's called The Investigation, and it was really good.
Fabulous. Wow, any Nordic crime thriller, I'm all for that.
What do you have for us, me here?
So I have one AI related and then one a little bit more of the detox related.
So the AI related is the best article I've come across about all this is something by Ted Chang in The New Yorker.
I've recommended Ted Chang as a writer before on the podcast.
He's this kind of crazy sci -fi writer, but also incredible technologist.
And he really characterizes AI and chat GPT in the best way possible as kind of a blurry JPEG of the whole internet.
And the way he talks about it is just brilliant.
So I recommend that article.
And then for your digital detox, I have previously recommended and would like to come back to Legos yet again.
So we did the biggest one that we've ever done, which is the Titanic.
And I got to tell you what a piece of engineering mastery that is so beautiful and completely consuming and makes you forget about your device for like hours on end.
So if you're in the market for something that will take you like a week to do, I recommend any of these large Lego architectural projects.
But in particular, the latest one, which is the Titanic is spectacular.
And you look inside it and you can see the engine rooms.
You can see the turbines.
You build the turbines.
It's just spectacular.
So that's my detox moment for you.
And this is it for today.
Thank you, everyone, for listening.
This was After Hours from the TED Audio Collective.
