The Secretive Billionaires Building a Tech Utopia + Casey’s External Brain + HatGPT 59:35

This podcast is supported by ServiceNow. Here's the truth about AI. AI is only as powerful as the platform it's built into. ServiceNow puts AI to work for people across your business, removing friction and frustration for your employees, supercharging productivity for your developers,

Providing intelligent tools for your service agents to make customers happier. All built into a single platform you can use right now. That's why the world works with ServiceNow. Visit servicenow.com slash AI for people. Casey, we like to break news on this show. Yeah, let's break some news today. I have some news. Okay. Which is that

I recently learned that it appears that the self-driving car companies in San Francisco are cracking down on people having sex inside the cars. No, really? How? So the other night, I had some friends in town from New York and we decided to go on some self-driving car rides. This is what one does when you visit San Francisco. This is the new riding the cable car in San Francisco is you ride the self-driving car. So I get out.

my Waymo app and we order a Waymo. Very smooth ride, no complaints. In the middle of the ride, so this is at night, we're driving down the street in the Castro and the light in the back seat comes on. Okay. And a voice comes over the speaker saying, we're conducting a routine cabin check.

And what does that mean? Who knows? But this has to be a response to the people having sex inside the autonomous cars, where they're basically like reminding you, we're watching you and we can remotely turn the light on in the backseat. Now, Kevin, what were you doing with your friends when they started the cabin check?

Well, weren't we doing it? No, we were just riding there, but it was very dark, and theoretically, there could have been some hanky-panky going on if that had been, you know. Well, you know, once again, here comes the nanny state to police individuals' behavior in their private spaces. Yeah, this is going to go to the Supreme Court. Hmm.

I'm Kevin Roos. I'm a tech columnist at The New York Times. I'm Casey Newton from Platformer. And you're listening to Hard Fork. This week, the secretive group of billionaires building a tech utopia outside of San Francisco. Then, the rise of AI note-taking apps. Can they actually make us smarter? And finally, it's time to pass the hat. We're playing another round of Hat Chupi Tea. ♪

Casey, we love a mystery on this show. Love a mystery. And we have a good one today. And unlike some shows, we're going to uncover the entire mystery in one episode. It's true. No eight-part podcast required to crack this cold case. And by the way, in this mystery, we're actually going to tell you who done it. It's true. So this is something that I became aware of over the past few years. There were stories coming out of the local news outlets around San Francisco and Sacramento about this mysterious group called Flannery Associates.

that was buying up a huge chunk of Solano County. Have you spent much time in Solano County? You know, I have been to Vallejo on occasion, but that was about the most of it. You know, Solano County, the Bay Area has got a lot of beautiful places to explore, but I would say Solano County may be a little bit less developed.

than some of the rest of the Bay. That's true. Fun fact, my dog is actually, one of my dogs is from Solano County. Is that right? Yeah, we picked Lenny up at a shelter in Vacaville. Lenny's a great dog. So this story is very interesting to your family because this part of your family comes from this area. Exactly. So I started hearing these stories and there seemed to be this pattern emerging where this group, Flannery Associates, they were just kind of

overpaying for big parcels of land in Solano County. And the residents of the nearby towns were kind of mystified. Who is Flannery Associates? Where are they getting all this money? You know, there started to become these sort of like questions and even conspiracy theories about why this group of mysterious investors was buying up all this land. And now how much were they overpaying? Because take it from somebody who recently bought a house, my understanding is basically everyone is overpaying for everything these days. Yeah.

Well, people were getting pretty good prices for their land. So clearly this was a group of people, whoever they were, that had a bunch of money and a lot of motivation to acquire land in Solano County. So I saw these stories. I thought, hmm, someone should probably figure out who those people are. And were you the person who investigated it? Absolutely not. I had one of those moments where I was like, someone should do that.

Not going to be me. You put a message in Slack and you said, what have you reporters out there? Right. But luckily, my colleagues at the New York Times did pick up that mystery and they have actually cracked the case. So my colleagues, Connor Doherty and Aaron Griffith, revealed that the people behind Flannery Associates, the people behind this massive purchase of land, are in fact...

a number of the best known names in tech. They're famous people. At least tech famous, I would say. They're tech famous, yeah. So this group includes people like Michael Moritz, who was a longtime VC at Sequoia Capital. He's also a billionaire. Reid Hoffman, Mark Andreessen and Chris Dixon, who are both partners at Andreessen Horowitz.

So this group is spending hundreds of millions of dollars to buy and sell the Silicon Valley.

buy up land in order to create what is essentially going to be a techno utopia. Wow. Well, you know, techno utopia was actually a rave I went to in 1999 and it was a lot of fun, but I'd be curious to see what the billionaire's version of it is. So,

We still don't know everything that they are planning, but the story starts off with an email written by Michael Moritz, one of the participants in this project, that describes what the article calls, quote, an urban blank slate where everything from design to construction methods and new forms of governance could be rethought. So this isn't just like a bunch of wealthy tech people like buying up land as kind of a speculative investment. They really are going to try to make some kind of

tech-focused city out here in what is essentially the middle of nowhere in Solano County. Right. Well, now, what do we know about the tech aspect of it? Because my understanding from these guys' involvement in local politics is that they're all people who think it should be a lot

easier to build houses. It should be a lot faster to build houses. And that maybe that's where they're putting the bulk of their energy. Do we know anything else about the actual tech angle of what Tech Utopia is going to look like? There's still a lot of details we don't know about this. The investors behind it are not saying much. This was a scoop that, you know, my colleagues had, you know, this was not a planned rollout. And so I think everyone's still trying to sort of get a sense of what to say. But

They did this all in secret. In part, the reporting suggests, because if these farmers knew that these tech billionaires were trying to buy their land, they would have charged them way more than even the generous sums they were offering. And by the way, if you're a farmer in Solana County and you've been holding out thinking, let me just find out a little bit more about these guys before I decide if I'm going to sell. Oh boy, did you just hit the little jackpot? Yeah.

That price just went up. True. So we don't know exactly what their plans are, but we do know from some of their pitches that they've made to investors and from the sources that my colleagues drew upon that this is both a financial investment with an expected return attached to it and kind of an ideological plan.

play, right? So right now, one constraint on the tech industry in the Bay Area is that it is very expensive to live here, right? There is not a lot of housing. Housing prices are very high. And so if you want to start a company, say, it's very hard to do that in the Bay Area because you're

just paying people a ton of money. Everything costs way more. And you yourself may want to live here in order to be able to sort of raise money from local investors. Yeah. And the traditional solution to this, building more houses, it can be next to impossible. And the rate of housing construction in San Francisco is just abysmal. And while a lot of the attention focuses on San Francisco, which is fine, the truth is most of the cities around San Francisco have very similarly restrictive zoning laws.

Right. It's not a place where it is easy to build things. And in some of these materials, the investors did say that, you know, if this effort succeeds, if they do manage to build this new city, that it could relieve some of the pressures on rising home prices, homelessness, congestion, etc. It's like crazy to think that in San Francisco, it really is easier to build something like Uber, which...

turns the entire global taxi industry upside down than it is to build like a condo project. You know, it's like Uber got off the ground a lot faster than any condo project in San Francisco. Totally. So this pitch is not just for a sort of suburban development. According to my colleague's story, these investors have an idea of making a bustling metropolis spanning tens of thousands of acres of

that, according to them, would be as walkable as Paris or New York City's West Village. I mean, it sounds good to me. Right. So obviously, this is still a very early stage project. They have not broken ground on any new construction yet. A lot of what they have bought is sort of what was described as kind of a checkerboard of plots. They're sort of going farm to farm and offering people money for their farms. So, Casey, what do you make of this?

Well, look, there's a lot, as you point out, that we still don't know. And I suppose I could learn things about how this project is developing that make me very concerned. But at the end of the day, we do have a housing crisis in this country, in the Bay Area in particular. It's very acute. And while we have tried many, many things to try to improve that problem, it still isn't much easier to build a house today as it was 10 years ago. So of all

of the things that these billionaires could be choosing to spend their time and money on, building a bunch of houses in an undeveloped area within a relatively short drive of San Francisco seems like it could be a really good use of their time to me. Yeah, it also just feels like not that risky a bet, right? Because...

In the best case scenario where all of your plans to buy up this farmland and rezone it and build houses and businesses and commercial districts and a new walkable city, if all that works, then you own a city. Then you own all of the infrastructure around the city and then you can sort of profit from that.

But even in the worst case scenario where these guys buy up all this land, they pay these inflated prices, they try to maybe successfully rezone some of it, but they end up not being able to sort of build a whole new city from scratch.

they still own a ton of land in Northern California, which historically has been a pretty good investment. And just with the amount of land that they now own, they could throw one of the greatest summer barbecues of all time. Right. And we should say that there's also like, there is a long history of tech people sort of getting interested in maybe building a government or starting a city. Like this is not the first time that something like this has been proposed. I'm thinking of, of course, the seasteading movement. Love this movement. Yeah.

I mean, it was just an amazing story where Peter Thiel, the venture capitalist, funded this thing called the Seasteading Institute. With Milton Friedman's grandson. With Milton Friedman's grandson, the conservative economist. And their plan was that they were so sick of kind of the restrictive anti-growth, anti-tech attitudes of kind of

the United States government and also other governments that they were going to go out into international waters and they were going to build floating libertarian tech cities on these like platforms. And I did a bunch of research on this a long time ago. Um, it,

it never really materialized. They tried some like early experiments where they would like try to live there for a few months and it just didn't go that well. So that project did not work out, but they did raise millions of dollars and get a lot of attention. Well, look, talk to any pirate and they'll tell you it's very hard to live on the ocean. That's true. The high seas are not a hospitable place. And in retrospect, I think we can all sort of maybe imagine that that would have failed. But what I will

say, I will say one sort of like pro-Seasteader thing, which is that the Seasteaders observed that there is not a lot of room in American democracy for innovation, right? The systems of government that we use today are basically the systems of government that we used 100 years ago, and the overall trend is just to sort of make everything more restrictive and a

a lot of those restrictions are good and I support a lot of those things. But when it comes to the specific question of building houses for people to live in, I think the restrictions have mostly been bad. Right. And so some ways that tech people have proposed alleviating the housing crisis are like, why don't we move Silicon Valley to Texas or Florida or some state where it's much easier to build? That really hasn't panned out. And so it seems like these investors are kind of saying, well, there is this place where we can get relatively cheap land. It's pretty close to

San Francisco, you know, it's an hour, hour and a half drive from San Francisco. Maybe you could still get people who, you know, are going to super commute into the city or something, but you could really actually turn it into like its own town. And yet maybe in this world where you do have a city that's

In the middle of Solano County, like maybe it can be more pro-development. Maybe it can be more pro-growth. Maybe the zoning restrictions can be easier. Maybe you kind of avoid some of the pitfalls that other cities in the Bay Area have stumbled on. Yeah, and that seems like it's worth exploring. At the same time, one of my curiosities here is ultimately how much room to experiment with government infrastructure.

do they actually have? They are still going to be part of a county. That county is part of a state. That state is part of the United States. And so there may be a bunch of pressures on them that prevent them from being maybe as experimental as they want to be. Totally. And those pressures, there are a lot of them, and they're going to be very intense. So there was a follow-up story that came out this week about all of the kind of political fights that lie ahead for this project. Exactly.

Because in California, nothing is more controversial than proposing building a house. 100%. And, you know, throw in some tech billionaires and things are going to get very dicey. Like, if you live in California at Thanksgiving this year, just experiment by saying, we should build more houses in this neighborhood. And just watch your entire family fall apart.

So one of the first things they're going to have to do, according to this follow-up story, is to hold an election because Solano County has a longstanding slow growth ordinance, which is basically a law to try to prevent suburban sprawl and things like that. So they're going to have to put this on the ballot and win, and that's going to require them to kind of make nice with some of the locals because one other thing that has come out of this is that some of the locals are pretty uncertain. They're sort of saying, like, why?

Were we consulted? Why did you do all this in secret? Why didn't you seek approval before doing this? So they're going to have to persuade some locals to overturn that slow growth ordinance. Then they have to face all these other rules, environmental rules, lawsuits, tussles with the state's Air Resources Board, the Water Board, the Public Utilities Commission, the Department of Transportation.

So I think these tech people, I believe that they are probably well-meaning. I think they do want to solve the housing crisis. I do not know that they understand what a world of pain they are in for. It is quite possible. You know, it's like, you know how everybody who goes to remodel their kitchen winds up telling you that it was one of the worst experiences of their entire life because they didn't understand how many things could go wrong, how much it could run over budget, how much longer it would take. This is essentially the biggest kitchen remodel in the history of California.

It's true. And I should say that this is not a new thing for California either. There have been lots of planned communities in California. I grew up near Irvine, which is one of the biggest. Totally. So this is not the first time that wealthy industrialists have gotten ideas about building cities. During the Industrial Revolution and immediately after that, there were a number of wealthy industrialists who had sort of religious or utopian visions, but some of them were just like, look, I'm

build factories and they're very efficient and well-run and maybe I could do a good job with a city too. This thing happens if you're very successful at business, which is that you wake up one day and you say, what if I reshaped all of society in my own image? And there may be some of that going on here. It's true. So this story blew up on social media when it came out last week.

And a lot of the reaction that I saw from especially people who are critical of the tech industry was that this is just hubris, right? This is a bunch of wealthy tech investors who think they can just kind of barge in, you know, mysteriously to this sleepy rural community and basically turn it into a glimmering tech metropolis. And that these people are about to get a rude lesson in how things actually work. What do you think about that?

Yeah, I mean, those folks could absolutely be right. You know, this is a very difficult project. The politics are complicated. And even though they want to go really fast, they're living in a world where most of the things that matter here take a really long time. So, you know, if I were to wake up in 10 years and you pulled me aside and you said, oh, hey, by the way, the tech utopia didn't work out, I would say, yeah, I guess that's probably what you would expect.

At the same time, these are investors. They're used to their bets not working out individually. Over the aggregate, though, some of them pay off and some of them pay off huge. Marc Andreessen, instrumental in the creation of the first graphical web browser, right? The Collison brothers invented an incredible payments platform that is now a really huge business.

In those cases, the odds were against them, but it paid off. And I think if you're an entrepreneur and something has worked out once, you're much more likely to try something again. So could this be a case of hubris? Absolutely. But again, I just keep coming back to the question of like, well, if we have to live in a deeply unequal society where some people have way more power and influence than others, how would we like to see them use it? And something I would probably put on that list is build more houses in California. I'm very curious about how

radical their vision actually will turn out to be. You know, I can imagine a version of this where it's just like, look, we want to get some dense housing up because that'll alleviate cost of living around the Bay Area and it'll give people alternatives to like paying through the nose to live in San Francisco. I can imagine also a world in which this becomes sort of like an experimental laboratory for lots of different tech

gadgets and products and innovations where maybe you do have streets with built-in sensors in them or drones flying everywhere. Maybe this is going to turn out to be a playground for

people who may not be able to experiment legally in other parts of California to go test out their new things. Yeah, I think the self-driving cars are probably going to have a pretty easy time of getting approved for the road in Tech Utopia. Right, not a lot of cones being put on their hoods in Techno Utopia. And by the way, you will be able to have sex in the cars in Tech Utopia, okay? That's just a principle they hold dear. Live and let live. What happens in Tech Utopia stays in Tech Utopia. ♪

When we come back, I'm going to give Kevin some notes. Oh, boy. This podcast is supported by ServiceNow. Here's the truth about AI. AI is only as powerful as the platform it's built into. ServiceNow puts AI to work for people across your business, removing friction and frustration for your employees, supercharging productivity for your developers.

I'm Julian Barnes. I'm an intelligence reporter at The New York Times. I try to find out what the U.S. government is keeping secret.

Governments keep secrets for all kinds of reasons. They might be embarrassed by the information. They might think the public can't understand it. But we at The New York Times think that democracy works best when the public is informed.

It takes a lot of time to find people willing to talk about those secrets. Many people with information have a certain agenda or have a certain angle, and that's why it requires talking to a lot of people to make sure that we're not misled and that we give a complete story to our readers. If The New York Times was not reporting these stories, some of them might never come to light. If you want to support this kind of work, you can do that by subscribing to The New York Times.

Casey, you had an interesting piece last week about something that I have been waiting to talk to you about on this show since we started the show. People who don't know you may not know this about you, but you are an organizing freak. You love productivity hacks and life hacks and software and note-taking apps. You are the world's most organized journalist, and I've never understood that about you because it just seems so foreign to me.

That's nice of you to say. And I would actually wager there are many other journalists who are at least as organized as me and probably more organized than me. But it is true that many years ago, I fell into the cult of the productivity tool software community, and I've been trying to rescue myself from it ever since. Yeah, you're always telling me about some new app you've discovered that automatically sorts and organizes every contact you've ever made or something you'd use to keep track of all your stories. And

I want to talk about this topic on the show today for a few reasons. The first of which is that it is just an area where we are very far apart. I think I am probably the least organized journalist in America, and so I desperately need your help. And second, a bunch of companies, including Google, are starting to roll out AI tools for essentially personal productivity, note-taking apps, calendar integrations, meeting summarizers, things like this.

And of course, one of the big theories about AI is that it is a productivity enhancer, right? That we are going to use these tools to make ourselves and our organizations much more productive, to get more done in less time.

And I think this question of like, will AI actually make us more productive and better at our jobs is actually maybe the central question of the next five to 10 years as far as the economy and work is considered. And so I think we should just get into it today because you had a really set of interesting thoughts about it. Well, thank you.

Yeah, so I wrote a piece about why note-taking tools are not making us smarter. And for a long time, I don't think that that is what note-taking tools were promising us. The first note-taking tool I ever used seriously was Evernote, which launched in the mid-2000s, and they were really responsible for releasing a lot of the stuff we now take for granted, like the idea that you could take a note on your laptop and

be able to look at it on your phone later. That was an Evernote thing. They came up with a lot of other creative stuff, and there was a lot of value in just making notes that you were taking ubiquitously available to you. Right. This is like a filing cabinet for your sort of digital life. That's right. And so I used Evernote pretty happily for a long while, and then...

Let's say three years ago, this thing comes to my awareness called Rome research. And Rome starts promoting the idea of network thought. And of course, the minute I heard that, I was in love. I mean, my thoughts, the Casey Newton bat signal. Come on. I mean, I'm so excited about this.

And they came up with a couple things that seemed pretty innovative to me. The first was every time you open up the app, it just creates a note for that day. Basically, a journal. Why is that interesting? Well,

Well, it does, I think, become interesting to have a journal where you can look back over the past month, get a sense of what you have been taking notes about. Maybe you're starting to develop ideas over time, something you took notes about three days ago while you read something else and it's related and you kind of start to build on that a little bit. So that journal was the first idea that I thought made things interesting. The second thing they added was something called bidirectional links.

regular links like you find on the web only go one way. You click it, you go to another website. A bidirectional link can kind of go back and forth between pages. So, you know, maybe if I have a hard fork page and I have a page for you, like those two things can talk to each other. Do you have a page for me? Yeah. And it's just like, he's getting worse. We need to, no, no, of course. But, but look, I mean, a lot of people

are already doing this sort of thing in something like Salesforce, what we boringly call customer relationship management software, right? But if you work in sales, you want to have a list of all of your contacts. And it might be helpful to know, like, what's Kevin's wife's name, right? And like, where did Kevin go to college? And just sort of things that you can casually bring up in conversation and make you feel like I care about you when I'm trying to sell you something. Yes, I remember like learning...

maybe 10 years ago that one of my colleagues had like what seemed like an insanely complex system for like remembering details about people. Like he would go to an event and get someone's business card and he would immediately write down like five details about that person, like brown hair, you know, two kids, likes lacrosse or something like that. And he would get home and he would enter it into a spreadsheet and he had like

thousands of entries on this spreadsheet of everyone I'd ever met. And can I guess that this person is very successful? Yes. Yeah, I mean, this is just like, it's obviously that sounds incredibly labor intensive and tedious, but you know that the person who's doing that is going to have some success in his life. 100%. So Casey, just-

What is your productivity stack right now? Like, what are you using? Like, just what is sort of baseline Casey, how he organizes his thoughts? Okay. So I use one of these tools and we can talk about how the tools I use evolve, but I'm not using Rome anymore, but I'm using something that is like Rome. Every day I wake up and I write a little bit in my journal and some of it's just personal journal. No, it is a digital software journal.

and I'm writing about what happened last night, just a little personal stuff, and then I have a section for my reading. So I flag articles every day that I want to read. I go through them, and then as I read, I'll highlight little snippets that are interesting to me. I'll paste them into the journal. I'll maybe take a note or two. And because of the way that I tag them, I can sort of bounce back and forth between different stories, different ideas, and over time, ideas for columns or stuff I want to do some reporting about emerges. Got it.

Can I tell you about my system? Yeah. Okay. So I have two kinds of notes that I take. Okay. If I have an idea that is short, I email it to myself. Okay.

I used to just do drafts. I used to leave it in drafts. And so if I wanted to figure out what I was thinking about a week ago or that idea I had for a column, I would have to go through all these unmarked drafts in my Gmail and look for the one. But then I realized that I should probably just actually send these to myself because that would make it easier to look at. Well, I think email is a horrible place to really put anything. So if I have a short idea, I email it to myself. If I have a long idea, I take a voice memo.

Why? Because, I don't know, it just seems easier than typing it all out. But the problem is, the horrible thing about this is it's not searchable. And because I record the vast majority of my voice memos at home, they're all titled like, you know, my address,

47 or my address 128. And so if I want to find something in a note that I took like a month ago, I have to like listen to like, you know, 10 different voice memos to figure out which one it was in. I'm having the experience of like one of the hosts of a reality show where you like go into a hoarder's house and you just see like piles of newspapers on the floor from 30 years ago. Like that is a horrifying way to try to remember anything. Yeah. It's not good. And I,

I have tried all, like you have been writing about productivity software for like basically the entire time I've known you. And everything that you have recommended, I have tried for between 12 and 24 hours. And then it just seems too complicated. I can't get into it. And it just does not beat the old system that I have, which is faulty in so many ways, but at least it's a system. So-

Tell me what you're thinking about AI, because one thing that I took from your piece is that this problem that I have, this horrible organization system that I have, AI could actually help me fix that.

potentially. So I should say the conclusion that I came to was that these note-taking apps, we cannot trust them to make us smarter because getting smarter is not about taking notes. It's not about copying and pasting text off the web. It is about engaging with an idea. It is about thinking. That is something that takes place in your mind. It is not

something that takes place in note-taking software. And so I think if you want to think better, you're going to have to figure out a way to do that in your brain, right? And I very much believe that. At the same time, I am also a degenerate productivity software nerd. So this AI stuff has me really interested. So what is interesting about AI? Well,

I'll tell you about another part of my little stack that I use to live my life. Everything I read, in addition to taking notes on it, and I should say, I only take notes on, you know, let's say 30% of what I'm reading. But there are other links that I just put in my newsletter. I will save those into a database. And I've been doing that ever since I started Platformer.

And that means that for thousands of articles, I have a database, and in many cases with the full text of those articles, and that represents a huge body of knowledge about all of the things that are important to my publication.

So where AI comes in is what happens when I can put all of that into the context window of a large language model and I can just start chatting with it? What happens when I can talk to my database and I can say, give me a timeline of events at Twitter over the past year, right? Or what are some of the big trends in online speech laws around the world? And what happens when the database can just say, well, here are the last five

15 links from that, and they're perfectly cited. And all of a sudden, I have a perfect research assistant. So I save things in a pretty organized way. You save them in a very disorganized way. But there's no reason why the chatbot shouldn't be able to make just as much sense out of your garbage system as it does out of my perfectly manicured one.

That's very exciting. And also must make you very mad that, you know, you spend all this time curating your information and here comes the chaos Muppet Kevin with his AI tool and he's going to be just as productive as you. My number one productivity tool principle is whatever works for you. If it works for you, I will not have one thing to say about it. The hardest thing to do is come up with something that works for you. Based on what you've described is not working for you and that's why I have opinions.

Well, I have hope here, but I also like I want to talk about this issue of AI assisted note taking and record keeping and kind of personal filing, because this is not theoretical anymore. Right. Google has reexamined.

recently started testing something called Notebook LM, which is essentially what you're describing. It's a kind of personal note-taking system. And I understand you've actually tried it. I have, and I have only barely scratched the surface, just got access to it a couple days ago, and it has some pretty real limitations. So it's essentially an AI-powered notebook. The basic idea is that you can add...

up to five of what they call sources, which are text documents of right now it's up to 10,000 words. You can sort of let the notebook slurp it all up and then you can just start chatting with it and you can start asking questions with it. It is effectively the conversational interface to the material that you have just uploaded. I got a demonstration of it and someone was showing me that they had essentially 25 years worth of notes that they had been taking and this person didn't have all the limits that I did.

but they were just able to ask it questions and get really amazing answers and sometimes really like provocative answers that, you know, maybe this person might not have considered. I did not have that experience. I uploaded one PDF of a paper that's coming out this week that I wanted to read and maybe summarize and chat with a little bit. And the paper was 20,000 words and the note-taking software couldn't digest it. So... Well, that's just too long for a paper. So just...

Help me understand this. So this is a chatbot essentially where you can upload a set of data that you have produced that's maybe nowhere on the internet, that's your notes, your diary, your last 10 years of, I don't know, expense reports or something, and have the notebook program kind of make sense of that.

Using AI. Yeah, that's right. And the hope is that if you do it this way, that the notebook will be much less likely to make mistakes, right? Because it is, and it's really quite good at citing the exact spot in the text that you have uploaded where it is drawing its conclusions from. And as we've talked about before, one of the main reasons I don't use chatbots for very much is because they're terrible at citing their sources. So I can't use any of the information because I don't know if it's credible or not. Yeah. So...

There are obviously some risks of using AI for this kind of personal note-taking. One obvious one is just a lot of people just don't want to upload very sensitive personal information into these chatbots because they don't know how that data is going to be used. I know a lot of companies are thinking through this. There's also this problem of hallucination or confabulation. These chatbots sort of make things up. Have you found in any of your experiences with AI note-taking stuff so far that hallucinations are a major problem? Yeah.

Well, again, I've only barely been able to use this notebook LM thing in the very short time I've used it. I have not seen those kind of hallucinations. So that makes me optimistic. But I think the question comes when you're able to look across a much broader body of text, right? Like if I am able to eventually upload my entire Notion database into something like this and chat with it, you know, then we're just going to have to see. Right. I mean, to me, the big question about these apps is,

is like, will they actually facilitate the kind of thought processes that spark new ideas? I mean, what you're describing, an AI model that is trained on all of your old notes that can kind of go back and serve as kind of your external brain, for lack of a better word. It seems like

that's useful for some tasks, like remind me what I was writing about six or seven years ago, but it's not necessarily going to help you come up with that breakthrough new idea. Sure. I think all of that is very real, but if you'd like, I can tell you how I used it this week to put together a column. Please do. So after I write this piece last week about how note-taking apps

won't make you smarter, there was a very funny and predictable response, which is that about 25 different degenerates wrote to me and said, okay, but have you tried this note-taking app? And the even funnier and more predictable thing is I looked at all of them and I picked one and I downloaded it and I've basically given my entire life over to it. Oh, no. So it's called Capacities. It's from this small team in Europe. I know basically nothing about them, but it's a very...

slick, pretty note-taking app, and I thought it did a lot of things right, and so I've been messing around with it. And is it an AI note-taking app? Well, so there is a free version, but if you pay them, you get access to these AI features. And I thought they did it in a pretty clever way. So I was taking one of my daily notes,

And I had seen this really great story in the Washington Post by Kat Zuckowski about how OpenAI says you cannot use ChatGPT to create targeted political campaigns. So you can't say something like, write a message to get people to vote for Joe Biden who are like women in their 30s in Pennsylvania, right? You're not supposed to be able to do that according to OpenAI's policies. In fact, you can go in and you can do that. And so Kat wrote about how for two months now, this has just been the case. I thought, oh, that's really interesting.

So I start taking notes about this, and then I open up this AI feature, and I ask, like, you know, what are some considerations around, like, building a policy around AI and political advertising? And it immediately spits out some, like, you know, pretty reasonable considerations, and I thought, okay, that's interesting. And then I said, show me some positive applications for AI in politics, and it was like, well, you could use it to run, like, a giant scaled-up get-out-the-vote effort. And I was like...

Well, that's actually pretty interesting. And so from there, I was able to import the AI generated text into my original note. And effectively, I had a back and forth with this kind of prototype research assistant. So just as I was thinking through,

what are all of the ways that I should be thinking about AI in politics, the AI was able to spit something out. And I thought, well, that might be worth mentioning. Now, the funny thing is that when I wrote my column after I sent it out, I realized I'd forgot to include the detail about getting out the vote, which just goes to show you that the AI isn't going to do anything for you. But to me, it was this first moment of, okay, I'm getting a sense of what it's going to be like to have an AI assistant with me, someone in the room who can just...

throw out things that maybe I wouldn't have thought about myself. That's interesting. So you have, I would say, an unusual job that requires you to synthesize and process large amounts of information on a couple sort of running topics that you like to keep very close track of over time. Yours is a job in which that kind of knowledge accretion actually really does play a big role.

How do you think most people are going to use AI note-taking apps? What does the teacher get out of this? What does the accountant get out of this? Sure. So if you're a teacher, I imagine you might take notes about your curriculum. You might take notes for a lesson plan that AI might be asking you, "Hey, do you want to enhance your curriculum with this or do you want to bring in this lesson plan that seems like it's really popular?" That seems like that's going to be really helpful.

If you are an accountant, maybe it's a little bit more reminder-based. Maybe it's saying like, "Hey, usually around this time of the week or the month you're working on this thing, I noticed that you haven't done that yet. Do you want to get on that?" Or, "I've reviewed these spreadsheets that you created and it looks like this is an error." That helps you along.

It's true that for most people, note-taking is just totally transient. You write it down, and it has very little value beyond a few milliseconds. Right, pick up bread at the grocery store. Yeah, and that's totally fine. But I think there are a lot of people who are trying to get smarter about something over time. In fact, when I wrote this piece about note-taking, I got a note from a professor who was just like,

It's funny to me how similar our jobs are, you know, because he's like, you know, if you're a professor, you do sort of have a beat. You have something that you're devoting your life to learning more about. And a lot of the people who write and think the most about note-taking are people who work in academia. Totally. Or people who write books. Like, I've been gazing at this world of book writing software for many years because people keep trying to get me to, like, use Scrivener or use these other ones that I, like, have tried them. It's too much organization. They want you to, like, have...

It just felt like overkill. And I'll tell you what I think my beef with these productivity tools has always been is that they just feel like fake work to me. It's like you are spending all of your time organizing the filing cabinet of your mind, and what you really need to be doing is getting things out of that filing cabinet and putting them into the world. And so for me...

I always felt like if I spent like an hour on Evernote organizing all my notes on a given subject, like it still didn't do the work of like writing about that subject. And so maybe that's where AI comes in. Maybe it takes your notes and say, hey, what about a newsletter on European social media regulation this week? You know, I first got into this whole cult in like the early 2000s. It is a cult, by the way. It is. It is. In the sense that it has charismatic leaders and they're all trying to sell you something. Yeah.

They make you shave your head. They make you shave your head and all sorts of things. But, you know, it was the mid-2000s. I was working in my second newspaper job, and this website Lifehacker came along. And it was honestly a great website. I learned a lot of stuff from Lifehacker over the years and introduced me to a lot of, like, you know, now famous productivity concepts like Inbox Zero. You know, the idea that you should have no email left in your inbox, which I actually think is hugely problematic and you shouldn't ignore it. But I would spend an hour or more a day just reading Lifehacker.

And I was like, yeah, I got to get more productive, got to get more productive. And what I realized like after a year or two of browsing was like, I didn't need to get more productive. My job was just too easy. You know, it was like, I was working in a job where the expectations for me were pretty low. I would write my story a day ish in a pretty short amount of time. And then I would just have time left over. It's like, well, now what? You know, so I needed to move onto something that would sort of, uh,

like use more of my brain. And that wound up being the solution to being more productive. I didn't need different tools. I needed a different job. I will say the note-taking system that I have found

actually beneficial for me, but that has some other problems with it is actually writing notes on paper. There's some research that shows that actually writing notes in longhand, it helps with retention. So for a couple of years, I like used to carry around these little field notes, notebooks, and I would like carry them with me and I would write down stuff and actually worked. The problem is I would lose them. And so then like, it would be like, you know, two months of notes just like gone because I like left it at a restaurant or something or it fell out of my pocket. But

I also don't like these phone-based productivity and note-taking apps because for me, the phone is almost antagonistic to productivity. If I pull out my phone to take a note in Evernote or save something to Pocket, then I'm getting distracted. Then I'm going on TikTok. Then I'm looking at my email. Then I'm responding to texts. I almost want like—you know what a great productivity tool was? What, what, what?

The Palm Pilot. Okay, say more. A dedicated device. All it does is takes notes and keeps track of your calendar. And you can upload it to your computer at the end of the day, so you're not risking losing an entire month's worth of notes if you lose it somewhere. But it was there, it was in your pocket, and it was a separate device from your phone. Yeah, I think this is really interesting. I observed recently that

that when it comes time for me to write my column in the afternoon, for the morning, I'm sitting in front of like a 38-inch widescreen monitor, and I can analyze, you know, 15 data streams at once, and I'm feeling very proud of myself as I'm looking at this, you know, magnificent battle station in front of me. But when it comes time to write the column, I find often just unplugging my laptop, going and sitting down on a couch and typing, and I realized...

Basically this month, the reason I'm doing that is that the smaller screen is less distracting. When I'm only looking at one thing at a time, it's just easier to be able to think in a complete paragraph. So I agree with you that if you want to talk about productivity, putting constraints and limitations around what you're doing is super important. If you were going to try to improve my

shambolic note-taking and life organization system. Can I get like a free consultation now? Absolutely. And honestly, well, what I would say to anyone is experiment with taking a daily note. Just when you're sitting down at your work computer at the start of the day and you have your coffee, open a note. It can be in the basic notes app that comes with your laptop. I encourage you to use a bulleted list because there is something about a bullet point that

that really takes the pressure off you, you think, I'm just going to write one sentence. And just write the first five or six sentences that come to your mind. It will probably be about something personal. It will be about something that happened last night. It'll be something that you're anxious about. And then after you kind of spill that out of your mind in two or three minutes, you'll think, you know, I've been thinking about that thing that that person said that's relevant to my job.

And so I'm just going to write a sentence. And then I just read this thing. And that actually sparked this thought. And again, just write like one sentence a piece. Do this. And then at the end of the week, just go back and look at the notes. See if you maybe have one idea about something. This will actually work for you. Like I guarantee that this will be somewhat interesting for you. And it may inspire you to keep doing it. So what I'm hearing from you essentially is that saving my short notes in my Gmail outbox is...

And my long notes in my voice memo box is not your productivity guru recommendation. No. And again, if you're one of the people are still emailing things to themselves. I do not understand this at all. If it is working for you, God bless, please don't email me. But I will say if you're emailing things to yourself and you feel like you're not as productive as you can be, believe me, those two things are related. Wow. Okay. So one app could fix me. Here's what you're saying. Here's the problem. And, and,

Email inbox is just a list of other people's priorities for you, okay? It's just what other people woke up and wanted you to do today. That is not a good place to put stuff that you personally want to work on. You want to create a safe space for yourself to do stuff that you care about. And it's not Gmail, honey. Yeah, well...

You're really turning me around on this. I'm going to try your system. So I'm going to try this new, what's this new one you found? So I've been using this. It's Capacities. It's at capacities.io. And again, I know very little about it. All I can tell you is I downloaded it. I thought it was beautiful. It comes from like Austria or Germany or something. So there's like a real sort of like engineering sheen to it. So I'll have to click on 47 pop-ups before I use it? No.

Do you want to allow cookies? Actually, you know what's up? Allow all cookies? Something you'll love about Capacity is that you can just save audio notes to it and they'll just sort of categorize those. Now we're talking. So that could really be up your alley. So what I actually need, here's my desperate feature request for the startups of the world. I need...

what you're describing, but for social interactions. I need a note-taking app for social interactions. I need something where like, it'll tell me like, oh, you talked to this person six months ago and you already told them that joke. Or like a little recorder that I could just wear everywhere I go that would catalog, transcribe, and analyze every conversation I have and tell me what I could be doing better.

So look, an automated way of doing that would be great. But in the meantime, again, if you just sort of do a daily journal and you say, I hung out with Casey yesterday and we talked about these things and he gave me so many good ideas and like, I should probably like, you know, buy him Starbucks or something. You know, you can just do that. And then if it's backlinked in the right way, eventually you'll just be able to click my name and you'll be able to see our last like dozen interactions and you can almost like browse the history of our friendship. I'm not ready for backlinking in my social life yet. I'm still trying to get on the notes app.

That would be progress for me. I remember it was a big deal when Jack Dorsey revealed, when he was the CEO of both Twitter and Square, that he used the Notes app for all of his thinking, and all the productivity people had a real fit. Yeah, and that was a case where I would say the productivity people were right, because look at what happened to that guy. Good boy. When we come back, another round of Hat GPT.

This podcast is supported by KPMG. Your task as a visionary leader is simple. Harness the power of AI. Shape the future of business. Oh, and do it before anyone else does without leaving people behind or running into unforeseen risks. Simple, right? KPMG's got you. Helping you lead a people-powered transformation that accelerates AI's value with confidence. How's that for a vision? Learn more at www.kpmg.us.ai.

Casey, it's time to pass the hat. We are playing Hat GPT. Hat GPT is, of course, the game where we draw tech headlines from the week out of a hat and then generate some plausible sounding words about it until one of us tells the other to stop generating. All right. And today...

I have a very special present for you. Oh, let's see this. A custom hat GPT hat. That looks gorgeous. A navy blue hat with white print that says hat GPT. Do you want to try it on? I would love to try it on. Although, you know, I have to say, I've never thought I look great in a hat. Oh, that looks good. You look good? I like that on you. You look good? It fits well. Oh, okay.

I'm so excited. I gotta say, it's not very roomy. It's hard to fit a lot of slips of paper in here. For our purposes, I think it works just fine. All right. You want to go first? Sure. Let's see this. Pulling this out. And our first headline for today, the SEC takes its first action against an NFT project as an unregistered security. This comes from The Verge. This week, the SEC announced its first

first enforcement action against NFTs. Of course, those are non-fungible tokens. The SEC said the NFTs issued by a company called Impact Theory were unlicensed securities. The company agreed to a cease and desist order, paid $6.1 million in penalties, and agreed to destroy all of the NFTs in question still in its control.

Impact Theory will also eliminate any royalties that it might have received from sales of those NFTs on secondary markets. Previously, Impact Theory had said it was trying to build the next Disney. Disney, of course, a famous company that was built by selling NFTs. So, Kevin, what do you think of this story? So I am obviously like fascinated with the NFT story in general. You know, I had this very bizarre experience of selling my own NFT out of a story that I wrote for $560,000.

The proceeds from that were donated to charity. But I continue to be just obsessed with the story of NFTs and the many dumb and shady things that they have been used to do. And I think it's important to note, like, the SEC is not saying here that all NFTs are securities or should be registered in the same ways that stocks would have to be. There are many kinds of NFTs that are just more like artwork. And so there's been a big question about, like, where does it tip over into being a security? In this case...

impact theory. They were selling these NFTs. Have you actually seen any of them? No, I haven't. I would love to see an illegal NFT. Okay. This project was called Founders Key.

You can see some of the NFTs here on the website OpenSea. Okay. I would describe these as sort of, it looks like the back of a computer monitor. That looks like a gamer battle station to me. Yeah. So these NFTs, they are not like the Bored Ape Yacht Club or the Pudgy Penguins or these sort of NFT cartoon characters. They are what some people in the industry would describe as utility NFTs, which basically means that if you buy one, you get some real world perk.

And in this case, it seems like they were promising people the right to sort of own part of a project that may one day wind up being a major entertainment franchise. And so in this case, Impact Theory does appear to have made some sort of promises or at least stoked some hopes among investors that they were getting in on the ground floor of something that could make them a lot of money, which to the SEC and frankly to me sounds a lot like a security. Yeah.

I think that the lesson here is that if someone ever tells you that you can get in on the ground floor of something and you're not looking at an elevator, it might be an unregistered security. All right. Stop generating. Okay.

AI brings the robot wingman to aerial combat. This is a story from my colleague Eric Lipton at the New York Times about a new type of drone called the Valkyrie, which is being tested by the U.S. Air Force. And according to Eric Lipton, quote, its mission is to marry artificial intelligence and its sensors to identify and evaluate enemy threats and then after getting human sign-off to move in for the kill.

Apparently, the Air Force is planning to build between 1,000 and 2,000 of these drones for as little as $3 million apiece. Casey, what do you think of this? Oh, is that all? Is that all for only $3 million? Well, you know, this is the military. Like, a stapler costs $4,700. $4,700.

So that's true. You know, I got excited when you first started reading that headline because I thought a robot wingman like that could help me at the bar. You know, there was a guy that I wanted to talk to, but I was maybe feeling a little bit shy. But no, it turns out that these are murder drones. Yeah, these are murder drones. So what do you think about murder? Well, my general philosophy about this is I think it's bad to create murder drones. I think when you sort of automate killing and slaughter, it seems to me like you could do a lot of harm. And

to my knowledge, we don't have a lot of good international rules about using drones for murder. Am I wrong to say that? No, and actually, there are already many drones in use by the militaries of the world that are unmanned vehicles. But,

But what's new about this is that they would not be all remotely piloted. They would have some kind of AI helping them identify targets. And then according to the story, there is still a human, right? These are not being, you know, deployed and just told to like kill whatever they find. It's like there still has to be someone pressing the button, but the

The AI is going to help them do that. And interestingly, this is a type of technology that the tech industry has sort of refused to build for many years. There were sort of protests at Google years ago about whether they would work with the military to build AI into some of this technology. So the Air Force is moving ahead. They have presumably found vendors who are willing to work on this stuff with them, and they are –

testing it now. Do you remember a couple months ago we went down to the Google Robotics Lab? Yeah. And they showed us these like what looked like giant vacuum cleaners with hands that were sort of, you know, rolling around on the floor and they're like, let us show you the state of the art and it was legitimately very impressive but the state of the art was that you would say to the robot like, go pick up the plastic dinosaur and then it picked up the plastic dinosaur and it was like, okay,

well, that was pretty cool. You're telling me that in the same world where that is a state of the art, there's a $3 million murder drone that could just use AI to fly around killing? Like, is it actually that easy to create a murder drone?

I guess so. We'll see. But I do think that they got to sort out the like hallucination thing. Like you don't want to be telling your murder drone, like go bomb the munitions factory and it like misunderstands you and like takes out a bakery instead. Yeah, I got it. I mean, this stuff is all very dark to me. You start like combining this stuff with facial recognition technology and like police departments. And a lot of this stuff just feels like let's just stop this right now. All right. Stop generating.

YouTube will waive a creator's content violation warning if they attend a class. This is from Emma Roth at The Verge. YouTube wants to give content creators who violate its guidelines a second chance. Creators will now get the option to take an educational training course when they receive a community guidelines warning. And in exchange, YouTube will lift that warning from their account.

account. This is basically a second chance policy for YouTube creators who break the platform's rules. What do you think of this? I love this. This is traffic school for creators. If you were going a little too fast, you didn't put on your turn signal, and all of a sudden you've got a YouTube content strike, you're going to have to go sit in front of a laptop and

and look at some incredibly tedious materials about community guidelines, and you will get your strike removed, and you can re-enter the digital economy. I love this. It's rehabilitation for YouTube criminals. Exactly! I hope they just bring it. I hope the instructional video, I hope it's like Driver's Ed, where they make you look at

videos of fiery crashes. I hope they just make you look at videos of YouTubers getting canceled and issuing tearful apologies. Exactly. You have to, part of the training is you have to watch at least 10 tearful apology videos, 10 notes app apologies, and maybe write a short essay reflecting on what it would be for you to be canceled. That's what I'd like to see. I hope they have like a scared straight program where like Logan Paul comes in and he's like, trust me, kids, it's not worth it. Don't get those strikes. Keep it clean.

Look, joking aside, I do think this is great. There are thousands of people who are building their entire businesses on these platforms. And the enforcement of those community guidelines can actually be quite capricious. You can think you're doing everything right and you're not being an edgelord and something comes out of nowhere, maybe a rule change and you didn't know it. And all of a sudden your business is at risk. So giving people

particularly those like well-meaning, good faith creators, a way to just kind of get this thing scrubbed from the record, I think is an awesome idea. Do you think if they miss traffic school, they should be required to maybe like clean up trash on the side of the highway or something? I mean, I think, you know, it would be a nice gesture. Clean it up, YouTubers, or you're going to traffic school. Oh, wait, it's my turn to read. Yeah, it is. Okay, you got the last one here. All right. Last one.

Google meets new AI will be able to go to meetings for you. From Jay Peters at The Verge, one of the biggest new AI-enabled features is the ability for Google's Duet AI to take notes in real time. Click Take Notes From Me, and the app will capture a summary and action items as the meeting is going on. During the call, you'll be able to talk privately with a Google chatbot to go over details you might have missed. And when the meeting is over, you can save the summary to Docs and come back to it after the fact. And then another feature

new feature. Let's do it. Attend in quotes, a meeting on your behalf on a meeting invite. You can just click a button that says attend for me and Google can auto generate some text about what you might want to discuss. Those notes will be viewable to attendees during the meeting so that they can discuss them. Wait, this is amazing. So this is a new feature that was just announced this week. And this is like, I think

a thing that can either summarize your meeting notes or actually go to the meeting in your place. That's right. And it could actually do something even worse, which is imagine that you're in a meeting for work and one of your colleagues isn't there and the little Google Meet window pops up and says, hey, Casey couldn't be here, but he wanted you to read this auto-generated text that I came up with about what he might have wanted to discuss if he cared to attend this meeting, which he absolutely does not. Can you think

of anything more annoying than that. Have you ever been in a meeting with someone or like scheduled a meeting with someone important and like you show up and it's not actually them, it's like their assistant? Like this is going to be like this, but times 100. Like the poor intern who has to sit there in the meeting

As the one human, when 40 people have sent their AI assistants to the meeting for them and you are just the person who has to sit there and listen to all these robots talk to each other, like what a world. Let me tell you, I want to talk to the first person who is fired for continually sending the AI to attend meetings in their place. If that happens to you, please get in touch.

And I will also just say, you know, you sometimes read about people who are concerned that when AI tools are introduced into their workplace, they feel like they are just training their own replacements. You've read stories like this, right? Let me just say, if you find yourself continually sending an AI to a meeting for you, that is what you're doing. You are training your replacement. Totally. You either show up or you don't. I don't think...

In almost any case I can think of, you want to send the AI to the meeting for you. Totally. I did love this one detail in this story, which is that apparently if everyone scheduled to attend a meeting sends their AI assistance instead of going themselves, it does actually stop the meeting. Like you don't just have these like phantom AI meetings where the robots are just, you know, going through agenda items with each other, which I think is a nice little touch, but it does raise the possibility that there's just going to be like a, a,

one human for every 20 people in a meeting. Now, let me say something constructive, though, because I do think there is a role for AI in these meetings. Because sometimes you can't attend a meeting, and you do kind of want a readout, but I don't want the sort of standard readout. I want like the bitchy readout, you know, where you're like, well, you're never going to guess who is 15 minutes late again. Oh, and guess who talked over this person? Did you see John pick his nose? That's right. There was a 15-minute incident of mansplaining, actually, by this one over here.

I think a bitchy AI who goes to meetings for you and then sort of gives you all the tea afterward could be very interesting. Google, let's get on this. Let's get on it. All right, stop generating. That is HatGPT. That's HatGPT.

This podcast is supported by KPMG. Your task as a visionary leader is simple. Harness the power of AI. Shape the future of business. Oh, and do it before anyone else does without leaving people behind or running into unforeseen risks. Simple, right? KPMG's got you. Helping you lead a people-powered transformation that accelerates AI's value with confidence. How's that for a vision? Learn more at www.kpmg.us.ai.

Hey, before we go, we're planning another round of hard questions for an upcoming show. As a reminder, hard questions is the segment where we answer your dilemmas, your ethical quandaries, anything tech-related that you've been struggling with and you could use advice on. You can email them to us or send us a voice memo at hardforkatnytimes.com.

Hard Fork is produced by Rachel Cohn and Davis Land. We're edited by Jen Poyan. This episode was fact-checked by Caitlin Love. Today's show was engineered by Alyssa Moxley. Original music by Dan Powell, Alicia Baitube, Marian Lozano, and Rowan Nemisto. Special thanks to Paula Schumann, Hui-Wing Tam, Nal Gulogli, Kate Lepresti, and Jeffrey Miranda. You can email us at hardfork at nytimes.com.

If you suggest a different note-taking app for me to try, your email will be deleted without me reading it. This podcast is supported by Meta. At Meta, we've already connected families, friends, and more over half the world. To connect the rest, we need new inputs to make new things happen.

And the new input we need is you. Visit metacareers.com slash NYT, M-E-T-A-C-A-R-E-E-R-S dot com slash NYT to help build the future of connection. From immersive VR technologies to AI initiatives fueling a collaborative future, what we innovate today will shape tomorrow. Different builds different.