Episode description:

 

Episode Brief

Misinformation, rumor, psy-ops and propaganda–whatever you want to call the four horsemen of today’s media apocalypse—have been with us as long as the media itself. But you have to admit that the arrival of digital technology, led by social media, has given all of those forces outsized power. We still haven’t quite come to terms with how tech has shattered things like a shared reality, democracy, civil discourse.

That’s why today’s guest plays a key role in the journalism landscape. Julia Angwin majored in math at the University of Chicago before launching a remarkable career in investigative journalism.  She’s a contributing opinion writer for the New York Times on topics of tech and society, a winner and two-time finalist for the Pulitzer Prize for Explanatory reporting. She’s also an entrepreneur, the founder of the Markup, an innovative data-first online newsroom and just this year, she founded Proof News, which builds on the computational techniques of the Markup to hold tech companies accountable.

Julia and Eric discuss how she uses the tools of technology to inform journalism; about why reporting is like finding mathematical proofs; how she hopes transparency at Proof will build trust in its journalism; about the role of independent creators in the news environment; and how to hold big tech accountable.

 

Transcript

Eric Schurenberg (00:01.358)
Julia, welcome to In Reality.

Julia (00:03.946)
It’s great to be in reality.

Eric Schurenberg (00:05.934)
I am so glad to have you on the show. I have been a fan forever. But I have to ask you, you know, your tech savvy is such a part of your career. It’s such a part of your journalistic brand. So what is a nice tech savant like you doing in a beleaguered profession like journalism? You could have gotten an honest job as a programmer.

Julia (00:31.178)
I mean, you should talk to my mom about this. She has very similar feelings. Yeah, I grew up in Palo Alto, heart of the Silicon Valley, learned to code in fifth grade, studied math and computer science at UChicago. And I think my parents were a little shocked and saddened when I chose journalism, but I fell in love with it on the campus newspaper, which I had actually joined because I knew how to use the page maker software to lay out the pages.

But then I sort of fell in love with writing and I really just thought, I’ll do it for a little while and then I’ll go back to the real world. But honestly, you can’t have more fun than being a journalist.

Eric Schurenberg (01:14.67)
It is truly a great profession. And it’s a little under the gun right now, but it is. It does give you a sense of, you meet the most amazing people, you do work and you get a sense of satisfaction when things work right. Now, you, I mean, we’ve talked about journalism, but you’re not just a journalist. You’re also a founder of a couple of nonprofit newsrooms, The Markup, and just this year, Proof News. I guess that makes you a repeat offender.

Julia (01:21.77)
Yeah, no kidding.

Julia (01:44.042)
Yes.

Eric Schurenberg (01:44.142)
Both titles lean heavily on your tech background and both aim to hold tech companies to account. So tell us about Proof, your new baby, and how it differs from Mark.

Julia (01:59.178)
Yeah. I mean, I think, when I founded the markup, I was really trying to use the techniques that I had been using throughout my career of, you know, collecting original data, doing data analysis, and essentially using the tools of technology to cover technology, right? Cause it gave me a special kind of insight into the, into the big tech companies. And proof is about bringing that a little broader than just covering tech.

I think the industry of tech is everything now. And so it’s kind of like, are we really just focusing on big tech? You know, today we published a story rating these different voice cloning services. We tried eight of them and found that basically the only consent they really ask for is just like a checkbox. Like, did you get the consent of the person before you clone their voice? And

You know, that’s not a very big technical barrier, but those companies are also not really big tech. They’re little startups here and there. So I wanted to take the techniques that I have applied towards big tech and apply them in a broader way. You know, we’re doing a big investigation on climate that’ll be coming out later this year. So there’s, I think there’s some issues, honestly, just feels like a very urgent time in the world. And tech is still a really key piece of what I’ll be covering, but I think I wanted to bring the techniques that I use to a broader array of topics.

Eric Schurenberg (03:31.118)
okay. And so is that the difference from the markup is that proof is not confined to tech.

Julia (03:37.034)
The other difference is that I’m really experimenting with video. So this hasn’t launched yet. It’ll be launching next month, but we are, one of the things that I think has happened since in the past five years is that the way that news is distributed has gotten broken. Primarily, you know, news was being distributed through Facebook and Twitter and both of those platforms are now actively hostile to news, not just

passively hostile. They have declared, we don’t want this on our platform. And so it’s really hard, I think every news outlet will tell you, to get the word out. And at the same time, we have the creation of AI that creates weird written text that looks plausible, might be news. And so I’ve been thinking about how to build trust with an audience in this kind of fractured environment.

And so what I’m going to be doing is experimenting with video platforms, YouTube and, and Tik Tok. So we’re launching that strategy next month. And it’s basically, it was an interesting insight I had, which was I’ve done data journalism my whole life. And so I have all these numbers and graphs and stuff, but somehow I have confronted the reader with a wall of words and not any visuals. And so I thought, you know, honestly, I should try to have the video format really forces you to think visually about your data and how to visualize it. And so I’m giving this an experimental try.

Eric Schurenberg (05:09.006)
God, that’s admirable. You know, I couldn’t help but notice that both the markup and proof news are not for profits. Do you believe that that is the, at the moment at least, the only way to survive in media and keep your integrity?

Julia (05:25.93)
I mean, the way it looks to me is that you can declare yourself a nonprofit or you could not, but either way you’re not making a profit in journalism these days. So I’d really rather just have the tax benefits. I mean, look at the Washington Post, right? Just lost $77 million. So I just, I think there is a profit model if you can do what the New York Times does, which is have a loyal audience and also bundle entertainment, right? That seems to be the one model…

…that is still working and it’s really just working for them. It didn’t work right for Buzzfeed, right? They bundled some pretty great entertainment and that didn’t work. So I just think realistically, if you’re gonna start a news outlet today, just confronting that reality makes sense. And I also think that when it comes to the video platforms like TikTok and YouTube, one reason that serious journalism is not really playing heavily there

is because the economics don’t make sense, right? Video is expensive to produce and the revenue share that the platforms offer, which is actually pretty generous, it’s a 50 % revenue share of the advertising, it doesn’t really support original reporting. So the types of places that succeed there are much more entertainment, reality TV type stuff. And I think that’s a disservice to the audience. I think they…

deserve to have good quality information that’s not just entertainment. And I think that’s an opportunity for nonprofits to reach an incredibly wide audience, right? The whole point of being a nonprofit is to be mission driven and to reach people. And so I think it gives me an opportunity to take that platform seriously without having to restrain myself to what would… what the profit model would ask for, which is essentially no original reporting.

Eric Schurenberg (07:24.782)
Julie, I have to tell you that at the Alliance for Trusted Media, which I found and which does research projects alongside newsrooms, a number of the experiments we’re helping them with are designed for short form vertical video. So, you know, read TikTok. It is something that a lot of newsrooms are experimenting with. I think people are rising or reaching rather the same.

Places you if you go where the readers are

Julia (07:55.498)
Yeah, you have to. And I think short form vertical video is particularly suited to news, the short nature of it. You can see that TikTok has already become sort of de facto a bit of a news outlet. It has a little bit of the vibe of Twitter in the old days. And, and the reality is there’s just not enough of the high quality news on there. And I think there are a lot of creators, sorry, there are a lot of creators who are really interested in news…

…but don’t have any training or don’t really know the norms and practices. And so one of the things I’m doing is engaging with them and doing joint investigations so that they can learn.

Eric Schurenberg (08:36.11)
Wow, that’s really interesting. I really want to unpack. There’s a lot there. But first, I want to ask you about one element in your career journey that overlaps a bit with mine. So I’m just curious about it. When I was at Inc., I edited many stories about entrepreneurs. And I know that it’s very different from being a writer. And then when I became the CEO of Inc. and Fast Company, I sort of lived that difference firsthand. You become…

…a manager of people, you have to raise money, you have to keep the thing afloat. And now you are a founder twice over. Does it give you a more holistic view or otherwise how is the job different from you than journalism practiced say in your column at the New York Times?

Julia (09:24.906)
You know, I would have never expected to end up as an entrepreneur. I honestly grew up in an era. Sorry, let me drink something.

Julia (09:39.978)
I grew up in Silicon Valley. My parents had startups. I knew what they were like. Basically, I saw that one or the other of them startup was always in the midst of failing. So I did not go into this lightly, but I went into it with basically two things. One was I was increasingly finding that the kind of journalism I wanted to do was more team -based. It was…

Because of the computational aspects of it, I was often trying to work with a programmer or multiple programmers. And so I was assembling these teams at the Wall Street Journal, at ProPublica. And to be honest, the model in traditional newsrooms is pretty much still solo reporter. It’s not as much team -based, at least not static teams that might assemble for one thing or another. But I was essentially always…

…overusing my allotment of data journalists and like sort of exceeding the quota of what the newsroom really wanted. So I came to the conclusion that I wanted to build a container, a newsroom that was designed around that. And that was really the founding idea for the markup. And I think what I have found is that I do need to build the container for the type of work I want to do. It still doesn’t fit traditionally into…

…a lot of newsrooms, although obviously data journalism, computational techniques have made their way into a lot of newsroom. It’s still not the norm. It’s still not built into the system. You know, my lead editor at Proof News is a data scientist. She’s also a journalist, but you know, I think that is a really unique set of skills that I’m optimizing for. And so,

I wouldn’t have wanted to necessarily become an entrepreneur, but I really felt like I couldn’t find a way to accomplish what I wanted to accomplish. And then what I also found after doing one startup, and I still don’t really understand this because it’s obviously complete and utter hell, but for some unknown reason, I’m very drawn to it and I find it very satisfying. I really have come to a point in my career where I have done enough of the stories myself.

Julia (11:58.154)
And I do want to build up a new cadre of journalists. And I want to find a way to model a future that looks sustainable because right now it’s very depressing out there and the young journalists are feeling hopeless. And so part of what I’m trying to do here is like model some hope.

Eric Schurenberg (12:17.87)
I love that. Let’s talk about not hope in this case, but trust, faith, and we’ll get to charity. One of the principles of proof news that you put in your introductory letter was that you approach stories as a hypothesis, which is a kind of version of the scientific method. It’s…

Julia (12:23.882)
Yes.

Ha!

Eric Schurenberg (12:44.59)
I remember interviewing Steve Blank, the creator of the Lean Startup Technique, who postulated that every startup is in fact a hypothesis. It’s also redolent of Jonathan Rao’s constitution of knowledge, which calls on truth seekers of any kind, including journalists, to treat every claim as potentially false. Everything should be held up to standards of truth. What was your reasoning behind treating stories as a hypothesis.

Julia (13:16.81)
You know, I think it comes honestly from being a math major. I loved math and I love proofs particularly. And when I started journalism, to me, I thought, my gosh, this is just a proof, right? Like this is just a proof with a little bit of different symbols that we’re using. And I think from my point of view, and I realize I’m unique in this baby, but I really always saw it as proofs.

And it helped me frame it in my mind, like basically what level of evidence do I need to support my claim? And I think when you start looking at it that way, you don’t have to be wedded to the like, okay, I need three anecdotes or whatever. Like sometimes, some claims, you really only need one piece of evidence, right? I always think about the first story that broke from the Snowden documents where basically there was one court order from the foreign intelligence surveillance court showing…

…that they were collecting every single American’s phone information. That was enough data, right? Like that one data point supported that claim. And then sometimes there are claims that need like a lot more support, right? And so I basically had always sort of used that model myself. And then as when I founded my newsroom, the markup, I really put it into writing. I was like, every reporter need to pitch their story. What is the hypothesis? What data would you need to support it? Because I felt like…

…that allowed us to be rigorous in our thinking about what are we trying to do here? Because as you saw in my opening letter for Proof News, the reality is we’re not really in a world where the world just needs any old anecdote anymore. Everyone’s newsfeed is filled with anecdotes, right? The hilarious thing about journalism is Man Bites Dog, right? You can find Man Bites Dog anywhere you want.

I think we have to hold ourselves to a little bit higher standard. And to me, that mental model of having a hypothesis lifts it from man bites dog to which is like hypothesis, are men biting dogs? And then you have to say, well, like, honestly, there’s not a lot of evidence of that.

Eric Schurenberg (15:28.302)
Let’s dig a little deeper on the question of trust. I mean, you noted in your letter and in other conversations you’ve had that trust is bumping along the bottom in our profession. It doesn’t help, obviously, that when powerful people call us the enemy of the people, but as a profession, we probably should take some responsibility for where we are now. In your opinion, what… did journalists do wrong? And how does proof aim to avoid those mistakes?

Julia (16:05.898)
So I’m working on a paper about this and I will be publishing it later this year with Harvard Shorenstein Center. But I’m actually really interested in a certain trust framework that you might be familiar with, which is this idea that there are kind of three elements of trustworthiness in a relationship. So for you and I to trust each other, there has to be something that we trust each other for, right? Like I may not trust you to…

…come over and house it, but maybe I would trust you to conduct this interview fairly, right? And so what are the elements that allow me to make that assessment? One is perceived benevolence. So in a relationship, the party who’s extending the trust needs to feel that the other party is on their side. Another piece of it is perceived expertise or competence.

I have to feel like, okay, you’re an interviewer, you know how to do interviews. I’ve seen that you do a lot of them, right? And then there’s this question of integrity. And the integrity piece is really about what I call policing. Will there be consequences for you if you break my trust? And I think that on the, when you think about journalism and their audiences, I have to say, I feel like it’s that third piece that is missing, right? Journalism.

Eric Schurenberg (17:21.198)
Mm -hmm.

Julia (17:33.834)
does not police itself, right? We don’t have licensing, you don’t have membership, anyone can just call themselves a journalist and there are all sorts of legitimate reasons for that. However, what it means is there are people who are writing and calling themselves journalists who have broken readers trust in the past, right? Like who are known to have spread lies or misinformation and they’re still out there…

…calling themselves journalists and that undermines trust in all of us. And so I think readers are feeling less and less like they can get accountability when they feel like they’ve been sort of wronged by an article that misled them, right? And I want to also pause and say, you are correct. There is an active campaign politically motivated to undermine the press and the trust in the press. And so we’re also facing that.

But I think it’s always a mistake to only focus on external threats and not also look inside internally for what we could do better.

Eric Schurenberg (18:39.918)
Let me poke a little bit about that first point, the first component of trust, which is sort of the belief in the audience that journalists are proceeding out of a place of benevolence. And I am reminded basically of the history of our coverage of technology. Before I came to Fast Company, technology could do no wrong…

…entrepreneurs and billionaires were heroes. That’s changed dramatically and changed before I got there. But I’m reminded of listening to a quote from Marc Andreessen who, and I’m quoting from memory, but he used that well -known quote about journalism existing to comfort the afflicted and afflict the comfortable as a weapon against journalism’s trustworthiness.

To him, that meant that newsrooms so oriented are out to get big tech, out to get us. And to be frank, people who win Pulitzers and Loeb Awards, you among them, don’t get them for saying everything is fine and that powerful people are well -intentioned in doing a good job. Do you think that we have a ways to go to rebuild the sense that among our audiences that we’re acting in the public interest.

Julia (20:11.338)
Yeah, I prefer George Orwell’s quote on this, which is that, journalism is what somebody doesn’t want published and everything else is public relations. I think that our role in society is to hold power to account. And so I don’t really love the comforting and afflicting model that, Andreessen used and has been used before because…

…I don’t think we are in the business of punishment, right? Like affliction is, it sounds like we’re putting people in jail, we’re not, right? We don’t play that role. We provide the information that maybe some lawmaker might use to punish someone, but we’re not the punishers. And I do think it’s important to recognize our role in society, but our role is not adversarial to the public. It’s meant to be on their behalf, right?

And so this is where, for instance, I think that you’re right about awards, right? They do reward a very specific style of journalism that is very much about a specific kind of accountability, which is I’ve found a bad, powerful institution or person doing a bad thing that they were trying to hide from you. And that’s important, but I also think the public wants other things, right?

That benevolence, they would like us to tell them, like this morning, I was going to vote in the Democratic primary. I couldn’t find any local news to tell me anything about these candidates, right? Like local news is so devastating that I’m in New York City, right? So I think there’s a bit of service journalism, which unfortunately has not had the prestige and awards, that also contributes to benevolence. And that’s another reason that I…

…really like to frame my thing around hypothesis. I actually really believe in explanatory reporting as being radical, as being public service. And I want to sort of steal the mantle from traditional investigative, which by the way, I’ve been a big participant in, but I also, and successfully, and I’ve been, I was a trained animal for that particular thing, right? And to be fair, it was,

Eric Schurenberg (22:20.366)
very successfully.

Julia (22:29.418)
hard for me to stay, I founded proof news with a new ethos, which is I’m not playing the awards game anymore. I just want to serve the public. I actually just really, I’m old enough. I’ve done enough. I don’t need any more awards on my wall. And I also find the younger journalists are not as caught up in the awards thing as I was when I was their age. They’re more mission driven. They’re more like, look, the world is.

burning up, like, I don’t know if I’m going to have a future. Like we got to get on this. They’re, they’re more activist than I was. And I want to talk about that because a lot of people say that’s a problem. And I say that the scientific method and having rigorous processes allows us to overcome the inherent biases that we all have. And so I believe let’s not punish people for their activism. Let’s build journalistic processes that stand up to rigorous scrutiny and ask people to judge us on our work alone.

Eric Schurenberg (23:32.878)
Let us pivot a bit to politics. Speaking of activism, what we’re about to, in fact, we’re already kind of swimming in the middle of a interminable presidential campaign. What should be on the agenda for reforms or regulation of tech? You have written about the EU’s Digital Markets Act and Digital Services Act. What can US regulators learn from that?

Julia (24:03.05)
I think there’s a bunch of things to be learned. One really important thing to be learned is that…

…laws can look really pretty on paper. The EU is really good at writing laws that look awesome. Their privacy laws look awesome. The Digital Services Act looks awesome, but enforcement is key. GDPR, the privacy law that went into effect, I think in 2018, has mostly led to these consent boxes that pop up asking you to consent to cookies…

…turns out that that is not really meaningful privacy, right? All that is is just everyone has to click to get through because no one has time to really negotiate each and every one of these policies. But there’s a really wonderful counter example in France. The French Privacy Enforcement Office, the CNIL, they have said their interpretation of GDPR, because the way EU works is each state can, each nation can sort of like tailor their own version…

…is that you have to offer a meaningful cookie consent box. So you can basically click, I don’t want any of these cookies and I still want every single bit of functionality on this site and websites have to offer that. So in France, there is actually a meaningful privacy thing. And I think that goes to enforcement, right? They are really out there cracking down on websites, demanding these changes.

And so we’re in a situation right now with Digital Services Act and Digital Markets Act that we are waiting to see how heavily these are going to be enforced. A lot of this is going to come in this summer when the first reports come in from these companies where they are supposed to rate their own algorithms for whether they are harming democracy, whether they’re harming youth. These are going to be self -reports. And it’s on the EU to say whether those reports are acceptable.

Julia (26:07.722)
That is a real test because if they just accept all of these self -reports and they say, yeah, you guys are doing great. And like, they don’t really have the technical capacity to audit meaningfully. Then that law won’t really have done what it was intended to do. One thing about the U S is we’re actually pretty great at enforcement. We just don’t have any laws in this space, right? We just like have chosen to entirely not regulate this sector. And so I think what we need is some basic…

…laws here, California has started with a very strong privacy protection act and it’s in the first ever privacy protection agency in the nation and they are really leading the way on privacy, but there’s more to be done. And I have, I’ve written a bit in the New York times about my pet peeve, which is I want control of the algorithms. So much of what is happening is that we can’t control the algorithms…

…feed, you follow people, which is your attempt to control your algorithm. And then they don’t even show you the people you’re following. They show you whoever they want. Right? This was sort of the innovation that Elon Musk had when he took over Twitter. He basically just said, no, no, no, you’re going to read my tweets and the people I like. And that was sort of his deal. And then it was a pay to play, right? He was like, actually, you have to pay me a fee to get into that special feed. And so what that means is we’re no longer…

…picking the news we want at a newsstand, right? The analogy I have is like, this is our news, this is where we shop for news now, but you can’t choose the magazine, right? You’re being forced into certain one. And so to me, a lot of the issues that we have with these platforms would be not totally solved, but would be mitigated if we had some dials and levers on our algorithms. If we could say, I never want to see another one like this, and I want to see a lot more like that, right? And that’s technically very possible.

and they don’t offer it.

Eric Schurenberg (28:06.798)
Hmm. Just what does that look like? Is up voting or down voting particular?

Julia (28:14.794)
Well, one way it looks is on Blue Sky. I don’t know if you’ve been on Blue Sky. That’s one of the social networks that competes that rose up after the Twitter demise. I mean, Twitter became X and then a lot of people left, especially journalists like myself, because Elon Musk chose to punish and kick off journalists’ coverage he didn’t like. So it became a very much a not safe space for journalists. So Blue Sky is one of these competitors and they do have…

…sort of the first version of this that I really like. So I can just choose feeds. So I have a bunch of feeds. I have one called tech news. I have one called puppies. I have one called happy news, you know what I mean? And then I can just choose those feeds, right? I just like, I look at my who I’m following. Then I like scroll over to tech news. I scroll over to the main news and the puppy news. And then I have a feed that like combines little pieces of all of them.

I would like to be able to, for instance, also get news curated by a trusted source. So for instance, maybe you as a person who runs a trust Institute would curate a trusted news source feed, or maybe I’m a fan of the ACLU and I want their feed or I want Heritage Foundation. Right?

That would allow us just so much more control and it would be much more similar to how news used to be consumed, which is you kind of knew the explicit bias that you were buying into. Right now, the problem is there’s sort of like the myth of objectivity, like, this is just the news because it’s what shows up in your thing and it’s really not.

Eric Schurenberg (29:50.414)
The big fear about AI in the election is that we’ll be inundated with deep fakes and AI generated fake news. What role do you think AI is gonna play in the election?

Julia (30:10.026)
Well, I’ve been tracking the use of AI in election contexts just through repeated testing of the leading AI chatbots. So starting in January, I held a forum where I actually brought in election officials from across the US, sat them down at computers, put in front of software that I had built that allowed them to enter the kinds of questions that voters often ask them…

…into a prompt and then they would see the responses from the leading five AI models. So Google’s Gemini, OpenAI’s Chachi BT, et cetera, Meta’s Llama. And so then they rated those responses for accuracy. And what we found was they were highly inaccurate. So majority of the answers were incorrect. And that was particularly true for

Some models more than others. So ChatGPT did better than the others. I was only like 20 % incorrect, whereas some of the others like Gemini and Meta were much more in the 60 to 70 % incorrect range. And then I’ve been repeatedly testing since then, and I have found that some of them are improving, which is good to know. But the problem is we don’t know, we have no insight into what types of questions people are asking to these models.

You know, with Google search, at least they put up the search trends, like what types of items are trending, what kind of queries people are running. It’s not a full data set of queries, but it’s something. There’s no information about what kinds of queries people are asking of these models. So I’ve been testing sort of what people in elections tell us are the most popular questions and the things that they most often hear voters getting wrong, but there could well be way more.

But that’s not the only risk, right? Of course there’s the risk of deep fakes, voice cloning, image cloning, fake videos. And I think a lot of people are worried actually post -election that during the ballot counting, there will be a lot of disinformation created around challenging the validity of the results, right? Which we already saw last time, which was a pretty, but it was a cruder attempt.

Julia (32:29.226)
They didn’t have as many of these tools. Now, if you want to make an image that shows, you know, you literally could just put in your two favorite election officials from your state and ask the thing to make a picture of them burning ballots, and it would look pretty reasonable.

Eric Schurenberg (32:47.246)
That is terrifying. By the way, I thought the proof news story about incorrect election information was very well done, very clever to bring in the experts. And I appreciated your following up. I guess it was a month, maybe six weeks later to see if the companies had improved.

Julia (32:49.226)
Correct.

Julia (33:09.93)
Yes, and we are going to continue to keep testing. And yeah, because I want to make sure, I mean, I want them to feel like they’re in competition, but this is one of the rare places where there’s massive competition and it should lead to better quality results, but that would only be the case if they’re being held accountable, right? And so part of my mission here is to continue doing this comparative testing so that whoever…

…does the best, gets some bragging rights, and then the other people feel like they have to catch up because in the absence of any regulation, it feels like market competition is going to be our best bet here.

Eric Schurenberg (33:49.07)
As long as we can really get it going, that is, I believe in competition. Let me ask you about another question about AI. So the New York Times is suing OpenAI for using what the Times believes is copyrighted material to train their AI models. And a few other publishers like the Atlantic, Reddit, Axel Springer have struck licensing deals with AI companies. What do you think AI companies owe publishers? and are the AI companies underpaying for that content?

Julia (34:26.698)
I’m not sure exactly what they’re owed, but it seems pretty clear to me that they’re underpaying. So, and here’s why, because these models are based on, they need high quality content to train and they stole this content, right? Like they went and took it without consent. you know, in cases like the New York times where it’s mine to pay well, we don’t exactly know how they did that, but.

Eric Schurenberg (34:31.822)
What do you say?

Julia (34:56.65)
they got behind and scraped everything. The terms of service are clear in a lot of these cases that you can’t do it, right? There’s the database of books that was scraped. My two books are in there. Every book of every author I know is in that database and that was scraped or I don’t even actually know. They must have actually scanned in all the books and then fed them into the models. So this was like a large scale theft. And so I guess the question I have is like, okay…

…once your stuff is taken and it’s been included in a product that they are selling and no one is offering to remove it from that product, what’s the compensation? We’re not even talking about, this isn’t a licensing deal. This is kind of like reparations, right? They took your stuff. They’re making a huge trillion dollar business off of it. What do you owed?

I would say probably some percent of sales, right? And that’s how with these deals, these are flat, single digit million for your entire corpus, which we already took, you know?

Eric Schurenberg (36:01.486)
Yes. Right. I was going to ask you what social media platforms you use, but you’ve already just given a nice shout out to Blue Sky. Are there others?

Julia (36:12.49)
I mean, I try to use so, I use so many of them. I can’t even, it’s like breaking my brain. I think this morning I posted on Macedon, Blue Sky, threads, Twitter, and I haven’t yet posted on LinkedIn, but I’m going to, and I know I’m supposed to call it X, sorry, but I just always forget.

Eric Schurenberg (36:30.99)
We knew what you meant, Julie. How about news? Where do you get your news and how do you know in this chaotic information environment what’s reliable?

Julia (36:41.162)
Well, I feel like I’m probably not the normal person. I do feel like I have more news literacy than, you know, others, but I do find that I still prefer my trusted, people that I follow, particularly in blue sky. they share information that’s relevant to me. I almost always find something that I wouldn’t have found otherwise. Right. So.

I have started going more often directly to news sites than I used to back in the sort of heyday of social media because I still get confused, especially with breaking news. Twitter was sort of this ultimate breaking news service and sometimes I find I just can’t figure out what’s going on. So I’ve been relying more often during breaking news on the New York Times live blogs or Washington Post live blogs because I find it’s actually really hard to

Eric Schurenberg (37:33.102)
Mm -hmm.

Julia (37:36.81)
threads, for instance, the one that Facebook puts out. Most posts that I see are two days old. It seems to be anti -breaking news.

Eric Schurenberg (37:50.158)
Hey, one final question. Are you, from your point of view, optimistic or pessimistic about Americans getting back to a shared reality?

Julia (38:05.674)
Well, I will say this. I’m an unreasonably optimistic person, which I think you have to be to be an entrepreneur. You have to be slightly delusional, I think. And so I know that the evidence is strongly against us having a shared reality. And yet I’m putting every single moment of my time into trying to build that. So I guess I am trying to make it happen. And I really.

I think we have the tools. I think we have the ability to do it and I think we can and we have to try.

Eric Schurenberg (38:42.446)
Julie Angwin, I will leave it there. That is great and optimistic sense, courage in the face of gathering clouds. Let’s just put it that way. What else can you do and what is the upside of pessimism?

Julia (38:50.41)
What else can you do, right?

Right. Exactly.

Eric Schurenberg (38:59.406)
Julia, thank you for being on In Reality. This was a great conversation.

Julia (39:03.434)
Thank you, it was such a pleasure.

---

Created & produced by: Podcast Partners / Published: Jul 9 2024

---

Share this episode: