This Week in Startups - Cade Metz on his controversial Slate Star Codex article, new book “Genius Makers,” dangers of algorithmic moderation, a16z vs. legacy media & more | E1187

Episode Date: March 19, 2021

Buy Genius Makers: https://rb.gy/s84drl FOLLOW Cade: https://twitter.com/cademetz FOLLOW Jason: https://linktr.ee/calacanis

This week in startups is brought to you by LinkedIn Jobs. A business is only as strong as its people, and every hire matters. Post your first job for free at LinkedIn.com slash twist. Vanta. Compliance and security shouldn't be a deal breaker for startups to win new business. Vanta makes it easy for companies to get a SOC2 report fast. Twist listeners can get $1,000 off for a limited time, at vanta.com slash twist and digital oceans app platform a new platform as a service solution to build modern

cloud native apps with app platform you can build deploy and scale apps and static websites quickly and easily use app platform for free and receive a $100 credit for any upgrade at dio dot co slash twist that's DO.co slash twist. Hey, everybody, welcome to this week in startups. Today on the program, we're going to talk with basically the nicest guy at the New York Times, who is also become the most hated person at the New York, second most hated person in New York Times by the tech industry. You can tell by my laughter that I'm joking.

Cade Metz is back on the program because he's got a new book out. He's got a new book out. And because he did a little story. That went somewhat viral, and we should just get into it. It struck a nerve. That's the way I prefer to think. So you've been on the pod before. I know you as an intelligent, considered kind,

how dare I say, soft-spoken, intellectual. You wrote a piece about Slate Star Codex, which I've heard of exactly prior to your whole Brouhaha and the Donny Brooke. ensued, I think I'd heard about them maybe four times from the Peter T.O. You know, A16Z, I don't know what you call those dark web kind of thinkers. Libertarian, I don't know. There are a lot of names for it. And that's what the piece is about, right? It's about this mindset. And we'll go into that. Yeah. How did you first become aware of this blog, the SSC?

Slate Star Codex. That's right. Well, I've covered AI for a long time. First, it Wired Magazine where I was for about five or six years. And then I moved to the Times about three and a half years ago. And if you cover AI, the blog comes up every now and again, right? As I describe in the piece, it's at this point, you know, the epicenter, let's call it, of a community called the rationalists.

and this community focuses on AI in many respects, and it's one of the communities in various parts of the globe that believes AI is a real danger, that it could in fact destroy humanity one day. That's one of the central beliefs. And so if you cover AI like me, every now and again, you get a link to Slate Star Codex. You develop sources and you have friends who read this on a daily basis

We're talking blog posts that are 8,000 words long, and there are people who read every single word of it. Is it well written? Is it well written and is it intellectually coherent? Do you find it appealing to read it? It depends who you're talking to. Well, I'm talking to you. Okay, well, this is one of the things I'm sure we will discuss ad nauseum.

I'm a New York Times reporter, right? So my aim, when it comes to writing a story about this or writing a book, I am not a player, right? At least... Your objective. Your goal is to be objective. If I can help it, right? I don't want to be in the story.

With the story, I had to be in there, unfortunately. But I do want to, even if I'm in the story, be an objective, you know, neutral point in the story. And I want to tell the story with fairness and rigor. And so it's not about my opinion. It's about everyone else's opinion. Let me tell you, opinion is divided. on whether even it's coherent.

So you and I are part of a group of journalists that were trained, I think, in the 90s. I don't know when you started. I'm 50 now. I think you might be the same age as me. A little bit younger, but yeah. A little bit younger. Okay. You come across as much more considered and mature than me.

So you're a Gen Xer. Yes. And we were taught, hey, straight down the middle. Your opinion doesn't matter. The facts are what matter. The facts tell the story. you would admit, I think, you would concede that the journalist in this, you know, I'm painted with a broad brush here, but journalism has become more opinionated where people pick aside and maybe do what I would consider more advocacy journalism. Do you think that's the trend? More opinion is inserted and we are the old school to this new school version of journalism? You definitely see that, right? And what bothers me is that all journalism

gets painted with the same brush, right? So there's this wide range, and there's always been a wide range, right? You know, we've been talking about yellow journalism, you know, since William Randolph Hearst and before. And, you know, what ends up happening is that as people, you know, embrace journalism as an advocate, then, you know, the New York Times gets accused of this

and everyone gets accused of this. And I can speak for my own work. I can speak for the Times work to a certain extent. But, you know, that's where it gets dicey is when, you know, people have an issue with certain types of journalism and then and then seem to see that in another type of journalism. Would you concede as well? And I'm setting up this line of questioning because I think you and I are sympathetico and I just want to sort of lay the groundwork here. And it's super inside baseball. But I think nobody's really had a rational discussion about your

peace. And nobody's certainly from my side of the table because I am a Silicon Valley living investor in technology companies, but I also started my career as a journalist for 15 years. So I feel like I might be able to help navigate this just a wee bit. Would you also concede that younger people are more interested in advocacy journalism than what I'll call the neutral straight shot journalism that you and I were trained on? Not necessarily. So people make that claim a lot, right? A journalist is a journalist, whatever age they are. Okay. And of course, we're going to see differences from generation generation. We're going to see, we're going to see differences from different parts of the country, different parts of the world. Okay. But look,

I have colleagues who are much better at this than me, who are much younger than me. Got it. And I have colleagues who are older, you know, who do things differently than I do, and who do things differently in New York Times. It's a wide range. Would you say that the Trump presidency had an impact on journalism where journalists felt they needed to take a stand? Because I do think that this might be part of the contentiousness,

which I seem to feel like got very acute during the Trump presidency, that people said, you know what? I normally wouldn't have an opinion, but I need to make an opinion now because this person is an authoritarian, a liar, and, hey, you know, this could be the next Hitler or we really are fighting for democracy here. So now is the time to not be precious and be straight down the middle journalist. Now is the time to actually advocate for what is right. Did you feel that you, that was a general thing that happened over the Trump presidency? I think there are two things that happened.

That's one of them, right? You see people do that. You know, I'm not going to deny that. You see people at other publications from my own who certainly do that. The other thing is that in the past, and as a Gen Xer who's been covering the technology industry for a long time, I started out as a researcher in New York at PC Magazine, which was... Oh, really? Yes. Wow.

P.C. Mac. Ziff Davis. On Park Avenue and 32nd Street. You got it one Park Avenue. know it. I've been in the building. That was my first job right out of college. I came to New York with no job and I was hired as a researcher, which meant I fact check stories. Wow. Fact checking. Whoa. Fact checking. Back in the day when you had money. When people had money for fact checking. You got it. And also the cadence, let's be honest, you know, we were monthly. It's in the name, right?

Like, or weekly or a daily. There was actually a little bit of time for fact checking in this new on-demand world. fact-checking is just not as much of a practice. It's true. Sadly. Sadly. Sadly. But there are other ways to fact-check, and we can talk about that, too, if you are.

We will unpack that one for sure. But Jim Seymour, John DeVorek. I know all those guys. Michael. Michael Miller was the editor-in-chief. When I met Michael Miller, when I was coming to the industry, at PC Forum, or one of those, I was shaking. In the 80s, I was shaking.

In the 80s, I grew up reading his column, Jim Seymour, Rest in Peace, John Dvorak. Who are the other columnists? There were like four columnists. Yeah, Bill Howard was a long time one. There's another Bill who I'm blanking on. But yeah, those, you know. It was incredible. And you had the lab.

You guys would benchmark PCs versus each other in their speed test. Had one in New York, one on the West Coast in Foster City. And my point, though, is that, you know, from there I worked up to be. a writer at PC Magazine. I later moved to a very different publication called The Register. Oh, yeah, Register, adversarial tech journalism from the UK, can't get any advertisers, bites the hand that feeds them. Bites the hand that feeds them. And so, look, very good. Is that literally their tagline? That is literally the tagline. And here, you know, my point with the Register,

great, great training ground for reporters. And it might not seem that way. Why? Well, let me explain it to you. Let me tell you who came out of there. Ashley Vance, who's at Bloomberg Business Week. He was my colleague for two years. One of the best reporters I know.

Dan Gooden, another incredible reporter who has been at, you know, business 2.0. Oh, wow. He's now at ours. And he was at the Wall Street Journal. You know, Chris Williams, you know, who went on to write for some great British newspapers. The list goes on. The reason why it's a great training ground is that it forces you to go around the companies. The companies will not talk to you.

Yeah, the register calls. It just block. You could get nothing, which meant you had to build the story yourself, right? Go to sources. You had to build the real story. Now, my point is that, so I've seen both sides of it. I've seen PC magazine, the PC magazine side. And let me just be frank about this, right?

There were times when the directive was, don't get the advertisers mad. You better not be critical here. Yep. That's wrong. That is not. Let's just say there wasn't a Chinese wall, you know, or if there was, it wasn't that high. You can step over it. And, you know, at various times of my career, right, there is, whether it's spoken or not,

you know, there's a director to be more optimistic and that kind of thing. And, you know, as you said before, as a journalist, your aim should be something different, right? Your aim should be to step back and look at this thing and look at the reality of it. And so what I have seen, and this is a long way of asking your question, but I've seen everything over the years. what I think has happened also recently is that the tech industry as it as it grew up over the past couple of decades you know it got a lot of favorable coverage and that was sort of the way it worked right is that you sort of presented what you wanted to a lot of these you know a lot of these publications and you got what you wanted and then at some point that started to change right

But I would argue before you ask the next question, it needed to change in that respect, right? The new year is here and that marks a fresh start for your small business. We're hiring a bunch of people at launch in 2021. We need a second producer, a third video editor, a community manager, operations people, and more. Things are going gangbusters for us. The podcast is sold out. We're going to three, four, five days a week. The syndicate is blowing up in a good way.

And our fund is hard at work doing the launch accelerator. So we need help. you know where we're going to find the most qualified candidates. You know it. I know it. We all know it. LinkedIn jobs, of course. We love using LinkedIn jobs at launch because we can manage all of our job postings and contact candidates from a single view. Whether you're shifting business hours or hiring more remote employees, one thing that remains unchanged is the importance of having the right people on your team. When your business is ready to make that next hire, LinkedIn jobs can help by matching your role with qualified candidates so that you can find the right

person quickly. LinkedIn now has over 722 million members worldwide and they mean business. So post a job with targeted screening questions and LinkedIn will quickly get your role in front of the most qualified candidates. You need speed and you need quality. Speed and quality. And that's what LinkedIn jobs is all about. When your business is ready to make that next hire, find the right person with LinkedIn jobs. And now you can post your job for free. Just visit LinkedIn.com slash twist and post your job for free. to use that special URL, LinkedIn.com, duh, slash twist, TWIST for a free job posting right now. Terms and conditions, of course, apply. Oh, I'm not going to disagree with you. I think, you know, having been there for it, and I did Silicon Allie Reporter, as you know, in the 90s, and I was

considered more closer to the register than, you know, a cheerleader, although some people said I was a cheerleader. I was more an interested party, I think, because I ran the publication in a very Machiavellian way. I actually picked the winners and losers. I took a Jan Wenner or a Graydon Carter approach, which was, bring me all the stories, and I'll spike whatever story I don't like, and I'll tell you which story to write,

and I'll pick who's on the cover, and I'm going to pick the headlines, and it was more me with a bunch of people supporting my worldview, which is unique in publications as well, but you are correct, and I think part of it was, the nature of technology was it would be introduced and it would make our lives better.

And I think up until that point in time, you know, airbags and PCs and GPS and whatever technology we found, digital cameras, MP3 players, laser printers, they all added to our lives. It was hard to find a moment in time where any of those technologies were having an adverse effect. Name a technology before 2000 that you felt had a horrible effect on life. Can you think of one? I'm trying to think of one. Yeah. Hacking, video games, I don't know.

They all seem quite nice. There are always drawbacks to technology. You know, whatever time period you're in, you know, there are positives and there are negatives. And a lot of it depends on who you are and how you see things. But I will say the technology is getting more complicated. Yes. Certainly. This is key.

Right? It is key. If I asked you before 2000, what's the worst impact you saw firsthand with technology and then after 2000? Right. I think it would be two different size lists and two different magnitudes. Let's do it right now. Post 2000, what's the worst you saw in technology? Well, I mean...

Well, again, my, you know, my aim is to stand in the middle here. But, but... What would most people say were the downsides to technology post-2000? Not you or me, but most people. Social media comes to mind off the top of my head. Well, look, anyone who's lived over the past four years has seen the effect of social media on their daily lives, right? You know, I was just talking with a friend of mine who's running a Facebook group over

the past 10 years and he saw how it changed and he saw how people reacted differently. He saw how the algorithms would push certain things and not push others in ways that he did not expect in ways that a lot of the creators of this technology did not expect, right? Or did not care, potentially. They're all different things, all sorts of things going on here. And then so the job of the journalist is to look at all that, right? The job of the journalist is to look at where these things are going wrong, where they might go wrong. And as the technology gets more complicated, and we're going to see this more and more in the years to come.

Sure. Facebook is relatively simple when it comes, you know, relative to a lot of this AI stuff that I talk about in my book. and how that has evolved and how it will continue to involve and how it could exacerbate a lot of these problems we're now dealing with. And I forgot to say, the name of the book, Genius Makers, the Mavericks, who brought AI to Google, Facebook, and the world, which is coming out right now.

It should be out. So go buy it, buy the audio book, by the book, and we'll get into that in the second half of this episode. So I think we're kind of an agreement and alignment. There wasn't really a lot of downside to the tech early, it's gotten when it hits scale, and because of AI, because of the feed, and specifically because of one company, Facebook, which I believe is a bad actor in the whole space. I think Mark Zuckerberg is a bad actor. I can have that opinion. You're a journalist. You might not have that

same opinion. But I believe he over and over and over again, when he faced an ethical or a moral dilemma, he picked what would grow the company, not would grow society or what would protect people. He always picked his own self-interest of removing friction and making the monstrosity that is Facebook, that frankincite of a horrible monster that he created. He only wanted it to grow. And that amoral nature of Zuckerberg, I believe, was the moment that somebody pissed in the well of technology and poisoned it for the whole village. That's what I think was a turning point in the industry, was his personal bad behavior. I'm going to put that aside for now. That's my opinion, not yours, but let's wrap up the

blog post. Okay, you find this blog. It's well read amongst the AI community. It's not well read amongst Silicon Valley. So when I saw the title of the piece and it said Silicon Valley Safe Space, was that the title of the... That's right. Did you write that title?

You know. No. Okay, great. I know that answer. I'm a poker player. If you take that long to answer and you look up, I can tell you're, you didn't write it. Let me ask another way. Do you stand buy that title. Of course. I stand by every word in that story. But no, did you write the title or not?

Did I write it? You know, I honestly don't remember, but it is certainly, I think it's a great title. Catchy. I stand by every word in that story. Like, you know, if you read the story. I've read the story, but hold on. Let me just start

the title here, though, because I think it's important for people to understand. And I'm actually, this is like I'm kind of advocating for you in this one. So, Take the win. All right. I don't, I know you didn't write that title. And I know that New York Times journalists don't write their titles in overwhelming number of cases.

They get written by editors and by now a social media group that is looking for clicks. And it is designed by a different group of people than the journalists write the stories. This is true. No, no, no. My editors and I choose the title. Okay. The titles are chosen to get clicked, correct? No. The titles are chosen by me and my editors. No one else. And it's the best title for the story. And let me point out here.

Best to find how. Accurate or appealing to Laurel the reader in? What are you guys thinking about when you write a title? You're thinking about all sorts of things. But if you think that the aim of the New York Times in titling stories. In Thailand stories is to get clicks. You have never been inside the New York Times. It is astonishing to people who haven't been in that the currency is not how many clicks you get. You're telling me people at the New York Times don't get paid by their social media following and how many clicks they get for their stories? No, we don't. Okay, you don't think Kara Swisher's deal is based on her Twitter following.

Ah, so this is a great thing you bring up because I'm, you know, this is one way. this is one way I want to convey, you know, what's going on. And she's a friend of mine and she's worth every penny. I mean, they're paying her a million dollars a year to do those podcasts, I'm sure. She's not an employee, I understand that. And podcasts are slightly different and you guys got your asses kick with the Caliphate podcast. No, but there is a difference. And this is, I understand why this is hard for people to grasp or if they don't grasp.

But I understand why. There is a newsroom where I work. I'm a reporter. and we have editors in the newsroom, and we have strict rules, a lot of which we've talked about during the course of this podcast. Separate from that, there is an opinion section. Yes. Kara is on the opinion side.

She writes op-eds. That has nothing to do with me. What percentage of users do you think understand the difference today? I don't know what percentage is, but a lot of them don't understand it. But what I will say is it? Whose responsibility is it then to correct that misunderstanding? The readers or the New York Times? When we publish an op-ed, the first word in the headline is opinion, right?

Okay. We clearly identify the opinion, but people still don't understand that there's a difference between what I do and what Kara does. Do you feel that the New York Times could do better at this in really making sure the, we both know the audience doesn't know the difference in most cases, I would say the majority of cases? and we both know that the opinion is gone wildly to the left and that there really isn't a space for right-leaning opinions at the New York Times opinion page anymore. So do you think the New York Times opinion page going so far left and not having room for right voices makes your job harder?

Because it exacerbates this image that the New York Times has gone full MSNBC, even though it's not accurate. I can tell you, I can tell you, you know, In all honesty, I don't think about the opinion section. That is separate from me. And people are always coming to me and asking me questions like that. They're always, they won't one introductions to people. All I can say is I work in the newsroom and that that's my foot.

So that is incredibly magnanimous of you. I think the New York Times reporters who are on the news side resent and are disturbed by the fact that the opinion page has gone just buck wild and can't even keep someone like Barry Wise, who is, you know, barely on the right or, you know, barely right-leaning, and that you can't keep that balance there. And I think it makes your job a lot harder, but that's my opinion, not yours. And I think it's why the New York Times is losing its status as the paper of record is because the opinion page has gone buck wild. I think because of the Trump derangement syndrome, which I think I suffered a little bit from, I'll be totally

honest. Like, I was appalled by the Trump presidency. It really disturbed me. I'm not afraid to say that. If you weren't disturbed by Trump, I think there's something. wrong with you. Like his behavior was so disturbing that having Trump derangement syndrome, probably the logical thing, like given what he did to this country and how, you know, unless you got a couple things right, maybe. Hey, everybody, I thought I would bring Christina Casioopo. I pronounced it correct. I'm hoping. Christina. You got it. Yep. All right. You're the founder of Vanta. People have been hearing your ads on the pod for the last year. And I thought it'd be fun to have you on and you to explain why you created Vanta and what SOC2 is. And what SOC2 is.

and why it's important people get it right. So let's start with what is SOC II for people who are just realizing they have to become SOC2 compliant? For sure. So SOC2 is at a high level. It's sort of a customer asking you to prove your security. Now these audit firms that you partner with, you prepare everything, but you still need to have an auditor. So who gets the order? You or the company that is engaging Vanta? Yeah, absolutely. So one part of what we do at Vanta is we've, we've built a network of audit firms, and there's a couple dozen we work with today. So we're happy to broker introductions, help companies kind of choose what sort of auditor or firm would be best for them. And then the other part of the pitch is, you know, that auditor knows Vanta, understands and

trusts our data. And so the audit will be faster and cheaper if one of our network firms are used. All right. Fantastic. Well, thanks so much for coming on and telling the audience why you should get your SOC2, when you should get it, and how you should do it. And you've been very nice to our audience, giving them $1,000 off, which is a really significant and generous offer. Go to vanta.com slash twist, V-A-N-T-A-com slash twist to get $1,000 off your sock, too. Thanks, Christina. Appreciate it. Thank you so much.

Cheers now. Well, so let's get into the story, because here's where I think it gets hard for you. I believe that you try to do this straight down the middle. And I think I kind of know you, and I definitely know your archetype as a journalist. Because I consider myself part of that, and I candidly hired people like you. specifically to do what you do. Like, you're the old school, now that we're Gen X.

We're old school. There's two generations behind us who have a lot to say about us. We're almost boomer territory now. But I think that this specific story becomes very difficult for you to navigate because everybody assumes that it's a hit piece. You did not have an agenda

where you wanted to docks this individual or uncover their identity. You just wanted to tell the story of the most influential AI blog in the world. Am I correct? Well, let me step back a little bit from that. So you brought up the title, right? Silicon Valley Safe Space.

You know why that's a great type? Because that's why I forgot why I found it offensive. I am as connected as anybody in this place. I'm the most connected in all likelihood. Like top 10 most connected people here. Very few people read it. It is not Silicon Valley Safe Space. It's an inaccurate title.

It's a safe space for like 20 people. who are Peter Teal's minions and, you know, that whole circle of Balaji and Andreessen Arts and all those weirdos. Does the title refer to the block? No. The title is Silicon Valley's safe space. It's talking about a subset of Silicon Valley, okay? And that is undeniable.

There's this certain mindset among, you're right, a subsection of people in the world and in Silicon Valley. And that's what I'm trying to do. And this blog is a window into that mindset. And it ends up talking about a whole lot of different people and groups. And there's a lot of names for this mindset. You alluded to this. You can call them the rationalist or there are people who self-identify as the intellectual dark web. So weird.

The Lunar Society on Clubhouse. There are all sorts of names for this. Who are these people? Let's describe the people. because I, we're going to disagree about how the title is interpreted. And of course, you know, this is a Roshaman type issue. There's five versions of the truth because we all have a different angle. From my angle, I was like, oh my God, you know, Kate is painting with this brush that like I think this way.

I don't think like these weirdos from Peter Thiel and Mark Andresen's like libertarians. I'm kind of libertarian, but I don't think like these guys do. They're kind of weirdos. I'll be totally honest. And the intellectual dog rub is super weird. These are like weird people. But who are these people? Okay.

If we have to make a composite, who are these people? What is a rationalist? Okay. I think it's very simple. And again, like, I think the story speaks for itself. Like, it doesn't say all Silicon Valley think this. It says many Silicon Valley leaders, right? In the deck, as we call it, like the subhead of the story.

It's a subsection. But the way I've been in describing it to people is, you know, we tend to think, about the world in terms of liberals and conservatives. We as human beings, we like absolutes. Left and right. We like left and right. Red and blue. And to paraphrase my favorite playwright, Tom Stoppard, this story is about the third thing

when you thought there were only two, okay? It's this third group that is somewhere in between those two, in between left and right. And what it's about is, you know, people who might seem progressive, might seem liberal, and they are in many ways. Socially liberal, right? Right. But they had this belief that any idea should be, you know, should have a space. Any idea should be discussed, no matter how extreme. And there are people in my story who are on the record, you know, saying this, right?

that even if you venture into like anti-feminist thought or if you venture into race science, you know, we should be allowed to go there, they say, right? We should be allowed to discuss this, you know, especially in Silicon Valley. You know, we're about disrupting and we have to have. So we should be, their thesis is we should talk about difficult topics, even if they are outside the Overton window. And sometimes those topics are really outside the Overton window. And so one of the reasons that the Slate Star Codex blog was a nice window into this is you can read the blog as sort of a, you know, something you sort of dive into here and there.

You might get a link. You read it. That's not going to give you the full picture. If you go into the comments, okay, it's a different world. Yeah, they're anonymous, they're pseudo-anonymous, yeah. But also, there are known white supremacists, like names you know, known eugenicists, right? It's like this, you get, you do get these really extreme views. And so then it becomes a question of...

Do those people write comments on the New York Times? Oh, well, it's, you know, I'm sure that they do, but this is different, right? There is a concern. How is it different? I'll tell you. It's a real concern. And this is where the rationalists come in. It's sort of this global society, let's call them.

But they really believe in this notion that, you know, you should give everyone their say. So there's this concerted effort to do that, right? I've talked to so many people inside this community, on the fringes of the community, right? You know, they really believe in this that everyone should have their voice. And you see this time and again where, you know, in these other Silicon Valley groups, they make an effort, you know, to give these extreme voices their due. Are you saying that they don't just allow extreme voices like the New York Times does, clearly, it's a comment section, that they might actually encourage those or in some way endorse them? Well, you know, definitely encourage like all voices, right? and this came out in like...

So they believe everybody should have a voice. Everybody should have a voice. Does the New York Times believe everybody should have a voice? Of course we believe everybody should have a voice. What's the difference then? What's the difference? Well, there are all sorts of differences, right? Is the difference here that you don't like those voices?

No, absolutely not. Is it the amount of those voices versus the amount of them on a percentage basis versus the New York Times? In other words, they have a lot of people who are racist. they have a lot of people who are sexist. They have a lot of people who want to talk about intelligence being inherited or a trait rather than a learned behavior. And the New York Times has less of them. So it's the percentage that disturbs you when you read it?

No, it's not necessarily about percentage either, right? You know, the New York Times job is different, right? We're journalists. We're looking at the world, right? We are not trying to cultivate a particular point of view or, you. you know, cultivate, the New York Times is not looking

to cultivate a point of view. My job is to give people an understanding of what is going on in the world, right? So, you know, whether it's, you know, what's going on with you

as a podcast or an investor or, you know, what's going on with this blog. My aim is to, is to look at that, right? Except for the opinion page, which is absolutely encouraging people

to have opinions and a diversity of events. And that's separate. So the opinion page of the New York Times does have a perspective. You pick who you want for president. You pick who you're going to fight for, or you'll pick winners and losers on that page, just like this blog will.

And you'll have a range of comments. But so this blog is very similar to the opinion page, but your reporting and the reporting in the newsroom is very different. It's just so we're clear. I agree with you on all this. And the reason I'm asking this in a very concise way is because I'm just trying to unpack exactly. I think it's good. Yeah. It's good.

And let me add one thing. Is you're right. There are some cases where the New York Times. we'll have an editorial, right? The editorial board on the opinion side will give their opinion, but then we have other voices that we ask in, right? All sorts of people. Except for the right-wing ones, which you can't seem to keep at the paper. Any thoughts on that? Again, that's not my area. I'm in the newsroom. It would make your job easier if they could have at least a, I don't know, 25% of the opinion page be right-wing or, you know, leaning right? Wouldn't that be helpful?

in terms of this discussion? My job is hard no matter what, and I'm glad it's hard, and like, that's the way, and it gets harder. Let me tell you, it gets harder. We can go into that. The job is just hard, especially in this world we're living in, of social media. And it's hard for a lot of them. And facts being debatable. And the search for truth is hard because the algorithms push to the top of the feed, whatever is the most engaged, not whatever is the most truthful.

And, you know, and there are reporters who have it a whole lot harder than me. So like, look, I have no complaints. I met the job. I'm, you know, I have the job that I want. This week in startups is brought to you by our friends at DigitalOcean and their app platform. This is a new platform as a service solution to build modern cloud native apps. With app platform, you can build, deploy, and scale apps and static websites quickly and easily. Simply point your GitHub repository.

and let the app platform do all the heavy lifting. Since DigitalOcean runs app platform on their own infrastructure, your costs will be significantly lower than with any other product. And no big price jumps as you scale, right? You're not going to get that surprise bill. It's built on top of DigitalOcean Kubernetes. Kubernetes, if you're a developer or you're in tech, you know what that means? If you don't, look it up.

It's K-U-B-E-R-N-E-T-E-E-S. And that provides a source. smoother migration paths so you can take more control of your infrastructure setup. There are three basic tiers, starter, basic and professional starter is great for basic sites. Basic is better for prototyping apps and professional works best for production apps. Use app platform for free and receive a $100 credit for any upgrade at DO.co.co slash twist. Do as in digital ocean. Dot co slash twist. Okay, so this blog exists. It's fascinating. There's a range of outside the Overton. window, this makes it the perfect story for a journalist to talk about. It's absolutely fair game.

And then comes the collision point. The author who was anonymous on his blog, but maybe not that hard to figure out who it was, had safety concerns for his lifestyle and, you know, the content being outside the Overton window. And thus the debate starts of should the person be named in the New York Times, like you and I are named on this podcast or you are bylined, this person has chosen to be anonymous. Why did you insist upon or the New York Times? Well, I think it's you who insisted because it's your piece and you could bowed out and not have published a piece.

So why did you insist on including his real name despite the fact that he asked you and begged you to please not do it? And not only did he beg you and plead with you not to do it, he turned off his entire blog in protest because of his. safety concerns. Why did the New York Times and why did you decide that you had to print his name? Well, this is complicated. And let's say, let me just ask you one question before I go on. How do you know that all that happened? Well, factually, I know he took his blog down and I know that he took it down because of the story. Those are just observable facts. He said it. I know that he has an.

alternative lifestyle because he said it. I know that you all insisted on, I know that you all printed it. So the assumption I'm making here is that you didn't quit the New York Times over it and you didn't publicly, to the best of my knowledge, object to his name being printed there. So then I assume by the fact that you didn't quit the New York Times over it, take your name off the piece that in fact, if you left your name on the piece and you still work at the New York Times, that you would endorse it. Okay. Well, is that logical? Because that's how any logical person would look at it in my mind. All right. Well, let me just go back. What am I missing? Let me just go back to Ground Zero. We can talk about this because there is a difference, you know, of worldview here, right? So I'm working on this story.

And you're right. He is pseudonymous on the blog, okay? But it is, but his real name has been shared for years, okay? Inside the communities that follow this, everybody knows his name. It's been shared online. He had even published. published one of his blog post in a science journal under his own name. When I finally, and I told him this, and when I finally sat down to find his name, it took me a matter of minutes. Like it's, you know, like when you did the Google Auto-complete, like his full name was in the Google Auto-complete, right? So not too hard to find. It's not too hard.

But he has safety concerns and he told you those. Hold on. Okay. Hold on. We take the safety concerns very seriously, right? And let me tell you, we understand safety concerns. concerns, and that is always a discussion, right? People often say, you can't put my name in the story or you can't make me the lead of the story, etc. It's not your choice. Because there

are safety concerns. Do we take that seriously? Absolutely. And we have conversations constantly about that. Is there a safety concern here, right? For lots of people involved. Do you think he has valid safety concern? What was your determination in this case? because I know your editors were involved in it. This is a very controversial case. So this must have gone to the top of the New York Times. What was the top of the New York Times position on this and its safety? I will tell you.

So what happened, though, let's get this right. Like what happened though is I went to him, right? Once, you know, people started to tell him about the story I was pulling together. And I thought, well, I need to contact him at this point. Yeah, they're back channeling it. Sure. Yeah. And so I thought I needed to contact him and see if, you know, we could talk.

And, you know, eventually. It's a right thing to do, yeah. You know, get an interview, meet him, you know, sit, send down with him. Unfortunately, I cannot tell you what he said because he has to go off the record and I need to respect that. But I can tell you what I said, okay? And, you know, on my side of the conversation, you know, what I said was is, you know, I could not guarantee that his name would not be in the paper. And in saying that, you know, I tried to meet with him and talk this through and try to understand all the issues. But, you know, what happened was is that his, you know, he took down his blog.

And at this point, I have not published a story, right? I'm just researching the story. Right. But he goes nuclear. He pushes the nuclear button. Right. And so, you know, he took down the blog and he said it was because of X-Ect. Y and Z, okay.

X, Y, and Z being you in the New York Times doing the story and outing him and his personal safety. Right, but it's phrased in a certain way, right? You know, it said that I threatened to dox him. I did not threaten anyone. You know, I did not threaten to dox anyone. His name was already out there. Doxing has a particular meaning. Some people think it means something different.

But, you know, what we feel strongly about at... Doxing for so people who don't know is sharing somebody's specific address. And in this case, revealing a pseudonym isn't exactly doxing. But I guess the second order impact could be, it could lead to doxing. I think that's the most accurate way to say it, right? Yeah. Yeah. But in any event, like, you know, it's cut and dry.

Do you understand when he says, when you said to him, I'm writing a story and I can't guarantee your name won't be in it, but let's talk. Do you not see that that could be interpreted as a threat that if you don't talk to me and then you put in the same sentence that I can't guarantee it? Well, if you're the writer, why can't you guarantee it? Well, because at the times, you know, we're practicing journalism, right? We are, my aim is to show you what's going on in the world. And if someone is worthy of a story, if they have influence. And let me tell you, there is influence here.

And you could see that in the reaction to him taking down. his blog, right? And you could see it. Then I didn't even publish a story and there was a huge reaction and there was a reaction when I did publish, right? And if someone, you know, is worthy of a story, you know, we can't just leave out facts that are right there out there in the open, right? And other people... It would look like you were being, to say it another way, it would look like the New York Times made a rule for Scott that they hadn't made for other people who had made the same request to remain anonymous? Well, you know, we, we're just doing our job. And then amidst that, right, we think about safety concerns. We absolutely think about what were your, what did you,

what was your assessment or what was the final assessment? Would he be unsafe because of the writing of this? Or would you think he was overreacting and using that as a technique to intimidate you? Well, I'm going to dox, you're doxing me and using that term then makes it look like you're a savage who wants him to be attacked. I'm not, you look, I'm not going to tell you what our conversations were inside the times, but I'm, you know, but, you know, this is, this is, this is, you know, cut and dry. But what I will say is that, you know, when I didn't dry that he's a public figure with great influence, therefore he doesn't get privacy.

Well, also. He's given up his privacy by being a public figure. If you write a book where you host a podcast, you are now a public figure. That's the determination you guys came to. Well, hold on. Also, and somehow this gets left out, we didn't publish the story until six, you know, six months later or more. He had already relaunched his blog and given his own name in the blog. Right.

Okay. So do you think this whole thing, do you think his whole nuclear thing with you was sincere or he was doing this to kind of undermine your piece? I have no idea. But what I do know. That's the right answer. You can't get in his mind. But I can't get in his mind. I can't get in anybody's mind that I write about.

But what I do know is that he published his own name on his new blog. And then we published his story. And, you know, so we certainly did not docks him. He had already published his own name. Again, the internet is a complicated place. Well, I mean, I think we've spelled it out pretty clearly here, the order of events. Right. Your piece did not come out with his name until then.

he reacted in a way that he felt was best for him, which is to take his blog down and not have people linking to it. And he said, well, you know, if it's going to come out anyway, I might as well own it and have my name come out. Do you think he's racist? Again, I can't get inside anyone's head. Do you think the things he's written are racist? The story is not, this is the other thing.

The story is about, you know, this mindset where you allow any idea, right? Some people think that that is absolutely fine, and they're really intent on making that happen. Other people think that there's a real risk there and that you can lead people to extremist views if you're giving airspace to those extreme views. So now I see it. So by creating a forum where anybody can talk about even the most controversial topics, the bell curve comes to mind, Charles Murray's book, in a claim in that book, I believe most accurately is that there was a difference between the IQs of different by race.

And it was a slight difference. And they talked about that. And if you talk about that, it might encourage people to think that black people are less intelligent than white people. Think about it like this. Is that the position people have? Is that by creating that forum for that discussion, you are encouraging people to then go to their worst demons and maybe exacerbating some racist tendencies? Right. Some people are concerned about that and it's not limited to this blog. There are other, you know, you know, situations where this happens. For instance, like there's this new publication, Quillette, right? And yes, they will publish a wide range of things, but some of it is in this very area you're talking about. And so, you know, that's the concern. And there's a really interesting study on this. There's a researcher based in Switzerland who I quote in the story. And he's done this.

this really interesting study where he looks at people who would view YouTube videos from the so-called intellectual dark web, right? These, Eric Weinstein, Sam Harris, who was never part of it. It was a personal friend of mine. We have the same book agent. We've been friends for a decade. He said, I don't want to be part of the intellectual dark web. Thank you. No, thank you. But those folks then lead to other folks. What the study shows is that people who viewed those videos, it was kind of an on ramp to more extreme videos on YouTube, and they can, they can track it, right? You know, user by user.

And that's the type of concern that is discussed in my piece. Whose fault is that? Because now there's dovetails with your new book. Whose fault is it that a perfectly reasonable Sam Harris, who is incredibly intellectual, incredibly fair-minded and rational, then leads to, Ben Shapiro, who then leads to Info Wars, right? Like, this is the, this is what people are claiming

the algorithm does. It takes somebody from, you know, the New York Times, then it hops to one of the subjects, maybe Star Slate Codex, maybe, you know, a Sam Harris video. Then that leads to an Eric Weinstein video, which leads to an Info Wars video, and then down the rabbit hole, you go. Well, these are like the biggest questions of our time, right? This is why I wrote the article. Who, but in your research, who sends people down that rabbit hole? It's a very easy answer. Well, it's not Sam Harris.

It's not Eric Weinstein. There are lots of factors here, right? It's, you know, and I'm looking at this writ large, but I think, you know, what you're saying is that there are, you know, there are people involved. There are technologies involved, right? We're talking about a vast ecosystem here. And you're right. In my book, this is one of the things I get into. And not just with this question, but so many other questions about where our world is moving, when you mix in this increasingly

complicated technology, and not just increasing complicated, but technology that you don't, where it's hard to understand what is happening, right? You're speaking specifically of the algorithm and how the recommended video is on the hand side of YouTube serve up the next piece of content. Anybody who has listened to, you know, an Eric Weinstein video or Lex Freeman interviewing Eric Weinstein, we'll see on the right hand Ben Shapiro interviewing Eric Weinstein. Now, Ben Shapiro, who is a devout Jew, doesn't believe in transgender, doesn't believe in gay marriage, etc. You know, based on what he said. I've listened to his podcast. He's, you know, he's pretty devout religious. So you have

scientists talking about science, and then Ben Shapiro maybe overlaps a little bit, or he's interested in their topic, and then InfoWorse maybe, or Dave Rubin kind of overlaps there and Cernovich, and then all of a sudden you're on an InfoWorse page or something even more pernicious, and that's all done by the algorithm, correct? Well, look, I think you raise a good point here, but what interests me here is how ideal spread from person to person, and they can spread in lots of ways. They spread through technology, and they spread in other ways. We tend to believe, you know, what's around us, right?

And now what is around us is, you know, the space is so vast, right? And technology enhances that. So when you use Twitter, you know, Twitter is a part of what's going on there, but also all the people around you are part of that and what they are saying and how many of them there are, right? And there are just countless ways, you know, and you know, you see this, whether you're a journalist or a VC, if you're in Twitter, you can, you're inside this bubble where you can think that that is the sole reality. And it's hard to step outside that and realize that that's just a small subset of what's going on. And the world is bigger than that. You know,

it's not just about technology. It's about the blending of technology and people and what they're saying. So there are two vectors here, if I'm parsing this correctly. You have the content being created. It has adjacencies as it would. I mean, a New York Times story might lead people to, you know, something further left, like MSNBC, which might lead somebody to something completely socialist. And then that might lead somebody to something completely communist. Are you a communist?

No, I'm not. But are you a socialist? Again, I don't think my political views are relevant here. Okay, I'm going to guess you're on the left, but we both live in the Bay Area. It's kind of a, we both live in California. Sorry, if that's docks in you, I could beep it out, but I think it's on your byline. It's all good. We both live on the West Coast.

We're both near the, we're closer to the Pacific than the Atlantic. Right. Please don't show up at our houses. The, but to the left of you is communism. And then to the left of communism is suffering. in pain and authoritarianism and communism is their authoritarianism. So you could go further left, too. People could wind up reading, what's the socialist publication, Jacob something? It's like a magazine Jacobs. Anyway, now I'm going down that rabbit hole to the left because of the Twitter algorithm

because I think socialism and communism is garbage. Jacobin. Yeah, J-A-C-O-B-I-N. I would ever known about that, but I started mixing it up with the AOC crowd, which then goes to the left a little bit more. So you can go left on this journey, can't you? As well as right. Of course. course you can't, right? Just so we're clear. But it's also like, you know, let's talk about just all the misinformation that, you know, get spewed on a daily basis. And that's, that's a huge problem as well. It's not just about, you know, moving to one side or the other based on on the facts you're seeing. It's based on all the misinformation that gets spread. And, you know, what's really interesting, and this is only just part of what my book gets into is these AI technologies that

are making it easier and easier and easier for machines to generate the misinformation, to generate images that look like the real thing, videos as well as tweets, blog posts, and it's not perfect yet, right? But we're moving towards a world where it's going to be so hard to tell if anything is real or not, whether it's the written word or whether it's an image. And if you think we have problems now with this, as the technology improves, it's going to get worse. It's a whole different problem. So we have the algorithm sending you down a rabbit hole.

And we know the more extreme of you is, the more energy emotionally it triggers inside of you, the greater the chances that you will watch more. We know this just from outrage culture, which we used to call it in the 80s or 90s, people who hated Howard Stern would keep listening because they wanted to hear what he would say next. The same with that reality television, right? The more outrageous, the more you tune in. So if somebody does something completely offensive, oh my God, we're all going to talk about it. And obviously Trump was part of that playbook, which he stole from Howard Stern. That was a whole thing in and of itself. And now we see the algorithms understand that part of human nature, which is the more emotion that is elicited by a headline that you may.

may or may not right. And on the YouTube side, which is just the best example of it, I think, is if they're going to suggest a video after you watch an Eric Weinstein, Lex Friedman video, where they're talking completely intellectually about something, if they have the choice of sending you to an MIT courseware on AI video, or an AI video about AI taking over the planet and, you know, Terminator, or, you know, some intellectual dart web video, they're going to show you the one that elicits the most emotional. because in their selfish purposes, I remember in the early days of YouTube, they said,

all we care about is time on site. That's our North Star metric. All the algorithms are told to increase the number of minutes you watch. They told me that. And Zuckerberg was very clear about that. We just want you to engage more, post more, read more, log in more. Isn't that the core of the issue here is that we've given over to the algorithm curation? Well, I think it's one of issues, honestly. Is there any issue bigger than that? That seems to me to be the number one issue. No, there are so many issues, right?

What's a bigger issue than that? Well, again, it's not just about the technology. It's about the people. And it's so easy to say, right, I'm just doing this, right? You know, my intentions are good. And this goes for the leaders of the companies, you know, the people who are generating the content, whoever you want. You can say, I am just doing X.

okay and I I care deeply about this and I don't want to cause problems but you have to think about the consequences right we do we we live in this world where there are consequences um to our actions and in ways that we don't always expect right got it unintended consequences of the content we produce right whether it's personal or whether it's a company right you know we all have to think about this stuff and it's very easy to try to come part-mentally analyze and say, you know, I'm just doing this in this tiny portion of the universe. But there are all sorts of people and technologies to think about and forces. And you can't just say, you know, it's just about the technology or it's just about like the data or just about rational thought.

There are, you know, there are emotional truths and historical truths that need to be thought about political truths. You know what I just realized as well, we're having this discussion, there's the algorithm, but then there was also the grifter crowd. There's a group of people who know that being in the info war zone of conspiracy theory is a great grift. If you talk about conspiracy theories, you will get more page views. You will get more ads. You will sell more product. So there's actually even more prentious aspect to this. Not only do you have the algorithm sending people down the rabbit hole. If it is sending you down the rabbit hole,

those people get rewarded with more money. And therein becomes this like crazy devil's bargain where you kind of start leaning towards that. I mean, I literally tweeted myself with red eyes, you know, the Bitcoin image. I just retweeted it to get the Bitcoin people out of my social feed

because they felt I was too critical of Bitcoin and they were bullying me. I was being bullied because I said Bitcoin you know, is very speculative and chances are it'll be replaced with a better technology. I mean,

you're talking about thousands of tweets very personal, like you're fat, your hair lines receding, you're ugly, you're dumb. And I'm like,

okay, I know, this is stuff I know. I've got a mirror. Like, I have my transcripts. I know I'm not the smartest smartest guy in the class,

but I just tweeted the image to get those people off my back. But there are other people who are tweeting those images, in other words, to make money, right?

There's a bit of a grift going on here, I believe. I think you make a great point, right? There are some people who, you know, their intentions are good. They just get caught up in it. There are the people who are, you know, actively trying to grab hold of that, right? And use it. Of course. Of course.

And sometimes it's different to tell, it's difficult to tell the difference between one and the other, right? It is. How do you get inside someone's head? Now you're talking about intent, which was where we started. Like, did you intend to docks him or does he, do he, do he, intend to blow it up to make you bad. You can't read in somebody's mind. You can't get inside someone's head.

That I know. Let me ask a question that's come up a bunch. Scott, you know, who's writing this blog, he has to be subject to public scrutiny if he's going to put himself in the arena and write blog posts. We all agree on that. This is, he's a public figure and there's no way around it. Some people don't agree with you, but I hear you. I mean, it's unrealistic.

that the president of the United States or a New York Times journalist or a podcaster has to own their words. And somebody who's massively influential with millions of people reading them every month doesn't have to own their words. You don't get a free pass on being a public figure, even if you use a pseudonym in my mind. So journalists, too, should be prepared to be under scrutiny. You certainly are prepared to be under scrutiny when you have a byline, correct? Absolutely. So why then is the New York Times? so offended by Taylor Lorenz being called out by another journalist Tucker Carlson to the extent that they feel they need to write a public statement about it and tell him to stop harassing her.

Why is it harassment when she is held accountable for misattributing quotes and, you know, the New York Times feels the need to come to her defense? Again, that's not my area. I'm a reporter at the New York Times. I've only just recently gotten to know Taylor, you know, and as far as statements, you're going to have to ask somebody else about that. Got it. You know, I'm a reporter. Got it.

Way to deflect that one. Well, I mean, I understand that you have to, but said another way, journalists should be prepared if you're going to be a journalist, putting her aside and Tucker aside, because I think both of them are, to be totally honest, self-absorbed and like the attention. They wouldn't be full contact on social media. media. Like, Taylor would not be attacking people all day long on social media if she didn't want attention. Now, that doesn't mean she wants harassment, but she could very easily do what many journals do, which is don't have a social media presence and just retreat from that or have a

pseudonym and just be on servers without people knowing it's her. Hold on. Here's what I will say, right? Harassment online is a huge, huge problem. And, you know, and it comes. And it kind of comes from so many different sources, from so many different directions. And it is intense. And in some cases, it is really, really horrible. And it's scary. I mean, what was your life like after the Star's blog came down? Did you get death threats? Did you get, did you have personal safety concerns? Well, let's put it this way, right? I have no reason to complain. There are, you know, colleagues of mine at the times who are far more talented than I am who get this so much more than I did, right? You know, it's intense. And, you know, a lot of people get it all the time.

And, you know, that's just a fact. And, you know, so many other people in the world are getting this. And it's a real thing. And if you haven't experienced it, you know, it's hard to understand. until you have. I think it's very astute point. I had a death threat last week. I've had three stalkers. I've had people show up in my office. And generally they fall into the, you know, mentally ill or incredibly effervescent, really major fans who just don't have, who have boundary issues, right? So there's like a boundary issue. But then there's also like, I literally got a death threat last week from Australia. And I'm like, now I'm like talking to the Australian authorities about how credible this developer who's apparently somewhat high profile. And they're like, you know, in our

country, we don't have laws against people saying they want to kill you or shoot you, which is what the person said to me. I literally had somebody docks my address and pictures of my house over the July 4th weekend. And they were a private equity person in New York who just disagree with something I said. In fact, it was over me criticizing Taylor Lorenz. I had criticized one of her pieces where she said, or she said on Twitter that, like, people were. being ridiculous that they were going to work and how selfish that was. And I was like, unless you have a family and you don't have a job and you go to work during a pandemic, in that case, that's okay. And you work on a keyboard. So it's really, really not cool of you

to dunk on poor people who have to drive buses, you know, to take you to work. And then somebody literally did a dox me. I mean, this stuff gets crazy. And for women, I do know it's 10 times worse. But all journalists are subject to this. And there's no reasonable world where a journalist could use a pseudonym, right? I mean, it would be... No, no, yeah. I mean, and... How would that even work? I mean, like, you have to own your words. Right. I mean, we live in, we live in complicated times, right? And, you know, I think journalism is sort of an extreme version of what I think everybody, you know, should go through

and hopefully does, right? You, you, you sit down and you think about what you've seen and what you've heard and you reach a decision and you act on it, right? And hopefully you stand behind it. And that's certainly, you know, what we aim to do. What gave you the idea for the book? Was there like a moment in time, a spark that made you feel you had to write this book? The book, I started work on it when I came back from Seoul South Korea. I traveled to Seoul in 2016 when DeepMind, the AI lab in London owned by Google, built a machine to play the ancient game of Go and they took it to Seoul to take on Lee Seidel, who was the best Go player of the past decade. And Go is like an ancient game that is far more complicated, exponentially more complicated than chess. And people thought that a machine

that could beat one of the best players in the world at Go was still decades away. And that week, when that machine defeated Lee Cedol was, and I keep saying this to people who ask me about, it's one of the most amazing weeks of my life. And I was not a participant. I was just an observer. It was unbelievable. The whole country and vast swaths of Asia, you know, China and Japan were focused on this. And you could feel, being in soul, you could feel the whole country sort of ebb and sway, you know, ebb and flow with this match. and it was a inflection point for AI technology. And when I got back, what I really wanted to do was write about the people building this. Deep Mind.

Well, I wanted to write about Demis Asabas, who's one of the deep mind founders, who's an incredible person. And so he became one of the characters. But then as I sold the book to my publisher and as I started to write it. Who's the publisher on this? It's Penguin Random House. Great. And it comes out tomorrow, March the 16th. Perfect.

everybody go buy it. You know what to do if you're in the audience. Go buy it. I appreciate it. But it really ended up being about these incredible people, these largely academics

who believed in this one idea that's driving most of our AI technology. They believed it for decades when no one else did, right? And they were based out of Europe, right? They were in London and they were like a dozen academics and I remember Elon told me he had funded them

and then Larry Page funded them and they got to some crossroads. I don't know if you cover us. in the book, the history of it. It's all in the book about how all this plays out with Page and with Elon Musk. Because Elon, I don't know if this is public, but he really tried to convince them to not sell, and he offered them a large amount of money to not sell to Google and stay independent. Interesting. Well, you know, we don't know, we, that's not in my reporting, but there's a lot in there that is. And it's not just deep mind. It's there, you know, the central character became

this guy named Jeff Hinton, who's a generation older than Demis, who believed in the idea of a neural network from the early 70s. And this is a, this is an idea that drives, you know, face recognition, you know, the speech recognition on Siri and your iPhone. It's essential to self-driving cars, what robotics are doing nowadays. The list goes on. All these technology we're talking about that that can generate images and generate tweets and blog posts. It's all, based on this one idea. And Jeff was one of the few people on Earth who believed in this idea. And, you know, for decades when no one else did, and then it started to work around 2010.

And there's this incredible moment when it starts to work and the biggest companies on Earth realize it. And Jeff auctions his services off, the services of himself and two of his students to the highest bidder. And we're talking about some of the biggest companies of Earth, Google, Microsoft, a major player in China. China was involved from the beginning. Yep. And that's the- We're talking about tens of millions of dollars. For, yes, for three people.

Three people. And that set the price for the talent. Which is everybody knows AI people who are developers who are worth their salt get paid low millions of dollars, one, two, three million dollars a year. Yes. Is that right? That's the moment when the, when the. price was set and it just went up from there. And it's fascinating to watch an entire industry see, you know, this, this what would seem to be this obscure mathematical idea working and then

like going all in on it, right? And one company will, you know, will ante up. And then you've got Facebook coming in, right? And they're raising. And it's really interesting how this industry works. And then who is in your estimation? Because for people who don't know Go and why it's so complicated, this is a game where stones, just consider them like checkers. You know, black and white ones. They're on a grid of 19 by 19, I believe. So it's 361 pins. And when you put two different colors on either end of a line, it flips to that color.

The number of permutations, this cannot be brute forced. it is not a brute force type AI, which chess is. Chess is a finite game. All the pieces are known, and it's a very small grid. Poker also has some misinformation, because you don't know what two cards you're holding versus the two cards I'm holding. And there's also bluffing,

which is another weird thing to put in a game. That bluffing actually works. So how does a computer know to try a bluff? Is a person going to call them down or not with a weekhand? It's very hard to know the randomness of that. But solving Go, we were very far off in terms of our estimation of when Go would be beat, right? You're exactly right. And you raise a great point.

The best Go players in the world play by intuition, right? They can't look too far ahead. You're right, because the game doesn't work that way. It's too big for the human mind. And so if you're going to build a machine that can beat the best players, you at least have to mimic that human intuition, right? they make moves based on field, the top players do. And you can make a move in the middle of the game. It'll have repercussions like dozens of moves later.

I compare to geopolitics, right, where one tiny thing will happen in one part of the world, and that will have a knock-on effect years later. One vendor has their fruit stand seized upon, which I think was that the Egyptian revolution that was based upon that? I don't know. One of the Middle East revolution, I remember, the whole awakening in that specific country was because somebody who was a vendor had their,

um, you know, fruit stand taken from because they, they didn't have an official license. And the person I think then did an act of self-immolation, you know, in protest. And then that ripple then created an entire revolution,

which is what happens in human rights or, you know, actually George Floyd would be the perfect example of it. Like we, we, we haven't confronted black lives matter. and that whole movement and equality. And I just had a number of,

I had two different African American and venture capitalists on my podcast recently, and they both pointed to the George Floyd moment as the moment that people and LPs took them more seriously as venture capitalists. That's really fascinating. And I was like, what? It took a murder.

And I asked them to explain, and they were like, yeah, I think it was when they saw a nine-minute murder occur in front of them, they just couldn't deny it anymore. And it just made them say, things have to change. You know, it's a really interesting thing that you bring up. And like, there's a, there's a prime example of this in the book as well. This guy, Jeff Hinton, I talked about, right? He had this idea that would, he was one of the few on earth who really nurtured this idea that would become so important decades later. In the 80s, he is a professor at Carnegie Mellon University. He'd immigrated from England. Okay. And there came a point where he realized the only way to do AI research was to take money from Ronald Reagan's Defense Department. And he did not want to do that. He and his wife did not want to do that.

They left the country and he went to the University of Toronto. Okay. The reason there were no neural network researchers, meaning, you know, important AI researchers in the U.S., in 2010, when all this started to work, is because he made that decision. Wow. How faithful. It's incredible. So the center of gravity for this research was in Canada and it was in Europe, as you said. It was not in the U.S. And so then it hits like two decades later, and all these giant, you know, U.S. tech companies have to go elsewhere for the talent. It's a fascinating thing. What do you think is the end game for corporations and AI? What are they so enamored with that they are willing to fight in that auction and put so much behind it?

What do you think they're trying to solve? They're trying to solve everything. So think about this, right? This one idea, it works with facial recognition technology, right? It's what drives Google photos, right? Where it can recognize what's in your pictures and you can sort them. Simple things like that. Siri, Siri now works with this idea.

That's why it can recognize what you say. You need it for self-driving cars so it can recognize pedestrians, street signs, line markings on the road, flying drones, like self-flying drones. You need that type of image recognition. And now, we now have these giant language models, they call them, like GPT3, which came out of OpenAI, this lab that Elon Musk helped found. What this system does is it's a giant neural network. A neural network is just a mathematical system that looks for patterns and data. So what you do is you take this neural network and for months, you give it Wikipedia articles, self-published books,

all sorts of other content from the internet, it learns to recognize the patterns in the English language. It learns how to piece language together, right? That means it can generate tweets, like I said, blog posts. It can learn to carry on a conversation. What could go wrong? Right, what could go wrong, but also it's certainly, everything to go wrong, but also it is vitally important to the future of Google. Like, if you want to build a chatbot, if you want to improve your search engine. This technology is all already helping to drive the search engine. So there's this force on the one hand that is forcing this technology into the heart of Google. And on the other hand, there are all these problems, you know, problems of bias against women and people of color,

as well as the disinformation problem. And you're not just saying that as like virtue signaling New York Times writer. It literally, if there is racism on the internet or systematic bias in the Wikipedia, because it's edited by a bunch of white graduate students in England or, you know, like that whole group of Wikipedia editors, wherever they came from in Jimmy Wells' circle, became the predominant editors. Now it's going to, because the data set was biased, biased and biased out. Is that the bottom line?

You are spot on, right? We as humans are flawed. We have biases, right? And we exhibit those biases on the Internet all the time. And so this system is going to pick those up. And a really good example is people have shown this with these systems, if you start talking about a programmer, it always refers to the programmer as he. Sure. Right.

And it applies the female pronoun to other things. You know, subtle biases like that. And then, you know, you get toxicity as well, right? Hate speech, you know, comes out of these systems. And we've seen that before. And we're going to see it again unless these companies. Why would we think anything differently? I mean, it's like, in one way, it's surprising and in a other way, it's just confirming.

Like if you, I mean, look at the history of cinema. If we gave it the entire history of cinema, you would, and the Overton window is opening and closing during that whole arc of the hundred years of cinema that we've experienced, you're going to have some, you know, things set in blazing saddles that wouldn't be said in the 40-year-old version, but then maybe South Park would say 10 times worse. You're going to have this ebb and flow of politically incorrectness or comedy or racism. And the machine is not going to know the difference. Is there something like Go that they didn't think would fall that you think in your research for the book is going to fall now? There's one that there's one that fell this past year. And, you know, this is probably the most important result of the decade. Deep Mind, who we talked about, who built that Go machine and had that really impressive result.

But it's a game, right? A game is not real life. They had a result this past year, end of 2020. It's called, their system is called alpha fold. And they cracked what's called the protein folding problem. And this is a problem that biological scientists have worked on for decades and that they didn't think would, you know, would fall for decades to come, just like the GoPro. This, anyone who has lived over the past year can relate to this and why this is important. The protein folding problem can help us repurpose medicines that we have, apply them to new diseases and viruses.

It can help us develop vaccines. It can accelerate this process. And them having solved this problem, this biological sciences problem, might make it easier for us to deal with the next pandemic. And if we've got medicines that have been improved by the FDA for other things, We can better understand and more quickly. Yeah. So Alpha Fold is just amazing. And by the way, if you go to DeepMind's website, they got a huge, they publish a lot of this.

They do. I noticed that Sam Altman from OpenAI said they were going to stop publishing some of their work because they felt it could be dangerous. Who owns these innovations? Are they patentable, trademarkable? Or are these all going to eventually be commoditized, like? computer chips and memory and cameras are commoditized, or will we see one company win it all in your estimation?

This is another great question, and it's so interesting to me, and this is a real thread throughout the book. Like I said, it was academics like Jeff Hinton, you know, who worked on these ideas, and then they were sucked in the industry. And what that meant was, is they brought an academic sensibility to this work. Academics published their research. and the big companies started doing that. So all this AI stuff, they openly publish it, which means everyone has access to it.

So the currency now in this field is twofold. One, it's the data. Do you have enough data? Okay. Well, actually, there are three things. Do you have the data? Do you have the talent, the people who know how to do this? And do you have the processing power?

And we're talking about a ton of processing power. So this type of stuff that Sam's working on, he got a billion dollars in funding for Microsoft just so they can have enough processing power to do this stuff. They got a billion dollars in cash from Microsoft or credits on their network. It's promised. You know how these deals are. Who knows what the time horizon is. A hundred million over 10 years.

Right. But it's a lot of money. It's a big number up front. And it's not the kind of money that everybody has. And what it's about is building these giant data. centers essentially that can process all that stuff. There are not going to be a lot of companies or organizations on Earth that can do that. So at least in the other times. That's interesting. So Google has all the search information and intent. Facebook has everybody's behaviors and personal

information. Tesla has all those cars and all the data or Uber has that as well with their apps, you know, on the phones. They have that massive data set. So the data set plus the talent, let's assume there's a lot of talent in the world, not infinite, but there's enough to spread around. there's going to be pockets of data that each person owns. And then the compute platforms, Amazon has a big one, Google has a big one, and I guess you can buy them, Microsoft has a big one. Is there going to be a moment where they are able to brute force certain things like, say, cracking Bitcoin or cracking deep encryption levels?

And did that come up in your research, that this could unleash a level of security concerns and hacking that we have never seen. before? And is AI being deployed by the Russians, Chinese, and hacking efforts yet? Well, that is a concern, but it's in a slightly different area. The concern there is with what they call quantum computing, right? And that's sort of a separate area. There's some overlap, but the real worry with a quantum computer is that if you can get there, and we're not there yet, the worry is that you can crack today's encryption. And that means you're going to need new types of encryption. Now, there are people who are also working on those new types, which also involve

quantum mechanics, right? This type of physics that drives a quantum computer. So it's another arms race. It's an enormous armist. Having written the book, Genius Makers, The Mavericks have brought AI to Google, Facebook, and the world, do you feel that this bringing it back to the AI rationalists, where did you fall personally on the belief that we could have a cataclysmic AI event versus that being the stuff of science fiction? So I'm fascinated by this whole dynamic. And it's, you know, what I tell everybody is, you know, very often in newspapers and, you know, and the like, you get these headlines, we go back to headlines.

You get these headlines that say, AI experts say. Well, AI experts are not a model. thing. And you have incredibly bright, incredibly well-educated, your top people in their field who believe that AGI, as they call it, a machine that can do anything the human brain can do is right around the corner and it's a huge danger. And then there are people who, you know, who are equally well-qualified and equally bright who think that's ridiculous. And it's really a belief, right? General AI you're referring to. Yeah, they call it artificial general intelligence. And basically means things like a human.

Right, things like a human. But a very powerful human. Right. And it's really, you know, the chapter in my story that goes into this, it's called religion, right? Because it's really about do you believe in that idea or not? And again, it's so much. Where did you come into the book with it? Where were you when you came into the book?

Do you think it was farcical to think that would happen or you just had an open mind to it as a possibility? You know me. I'm a New York Times reporter. I stay out of this. And I just observe. I observe. I observe. Straight down the middle.

You observed. Okay. Now, here's what I will say, right? What's fascinating to me is none of us know what's going to happen in the future, obviously. We could have AGI in five minutes. There could be an announcement in five minutes, and it could be there. It would do the announcing. It would take over all the radio stations and all the podcasts and replace it with its announcement.

Right. People of Earth. Right. Because we don't know, we can make any claim we want. And so that's the battle that you have. And you know this about Silicon Valley. Like, if you're going to build something, whether it's Facebook or some piddly little app or AGI, you better believe in it.

You've got to be delusional. Delusional, as I tell founders, is a superpower because being delusional means you don't quit. You got it. If any rational person knew what it took to build a self-driving car, they would not start. You got it. Now apply that to AGI, which is so much harder to do than a self-driving car, right? So it's the same playbook applied to something that is just astronomically more complicated. The Chinese are, you know, it's the CCP, it's a communist Chinese party.

They have a different view of humanity, human rights, how people should be treated. Human rights organizations universally agree. There is a genocide occurring with the Uyghurs in China at this very moment. And the Chinese have made it a national effort, I believe, to win. the AI race. What did you learn about China's efforts? And then part two, when you put a communist country that is currently involved in the one genocide that we know of on the planet, like at-scale genocide, like legitimate, plus AI, what concerns do you have? Well, I learned a whole lot in writing this book. And one of the things I alluded to earlier is that,

you know, China was involved in this race from the beginning, right? When I talked about Jeff Hinden auctioning his company office. Bidu was one of the companies bidding, right? The Google of China, they call it. And, you know, that's worth remembering, right? China is a big player here. We talked about the talent being important and the data. Oh, boy.

China's a big country, right? Yeah, and they get all the data. There's no privacy in China. Exactly, right. Or on TikTok, less do you think. And there are some just incredible characters in my book, Chi Liu, who's a China-born technologist who was one of the top executives at Microsoft. There's this incredible story about him and why he really left Microsoft.

Let's just leave it at that. But others lead to it. Could have been influenced? Could he have been influenced? You're like, you can't believe the story. It's just like too good to even, you know, even hint that here. Go, go read this. I can't wait to read it.

I mean, if you look at Jack Maude disappearing for six or eight weeks, like the Chinese have a different view. of the God King CEO than we do. It's true. And the other thing that you mentioned, you know, was, you know, the situation, you know, involving an ethnic minority. And AI plays right into this. One of my very, very talented colleagues, Paul Mosier, has covered this for the New York

Times in Asia, right? Everest to use face recognition technology, you know, to identify this ethnic minority. and, you know, there are pointed ways of using this technology for surveillance and, you know, in ways that concern a whole lot of people, right? And that's one issue. And there are lots of concerns here, you know, not just surveillance, but autonomous weapons. And, you know, we got to think about that, too. If we ban autonomous weapons here, they're just going to be. It doesn't mean China is not going to build a tank that can drive itself.

and figure out which buildings to blow. And, you know, if you just think, I mean, if you ever, I don't know if you ever saw the movie, The Lives of Others about the Stasi. Absolutely. What a great movie. Great movie. But East Berlin, I forgot how much I enjoyed talking to you.

God. Oh, it's great. This is a great conversation. You know, if you just think about the Stasi with AI, you know, like, you're just listening to everybody. And then all of a sudden, you're, this is 1984. Like, everybody's got a device listening to their thought crimes or a potential discussion occurs.

And now the government's like, hmm, where are the hotspots for, you know, people who believe in freedom of speech? And okay, then give me the network around that. Okay, bring them all in. And let's have a little interrogation party and then we'll get everybody to confess. I mean, this has really dark, dark implications.

I mean, just meeting people, if they could know the proximity you are to a revolutionary to somebody who's of a certain minority or a certain religion. And just your proximity to them could make you guilty. I mean, it gets back to the precogs in Minority Report. And I think AI could actually predict pretty well. If we know that people are led to a life of crime, you know, and it starts with this

indoctrination, this, you know, a gang indoctrination, a mafia indoctrination, running small crimes, eventually becoming big ones. The pre-cog concept is pretty obtainable by AI today. Well, people are already deploying much simpler algorithms to try to do this type of of thing in the justice system. We've written about this. And, you know, it's an issue today and the algorithms are very simple. Once you start applying more complicated things, it gets more complicated. And then that's just one, you know, one issue to think about here globally. And these are, these are global issues and they're complicated. Like, you know, people, again, like to think in

absolutes and that. Yeah, no silver bullet here, no good and evil. There's just a spectrum. Exactly. And let's not forget. that, you know, people will say, you know, we need to close down our borders, say, to our, you know, to talent from our rivals because we're worried about, you know, espionage. But then if we do that, you know, we're just shooting ourselves in the foot, right? We need as we've showed during this conversation, the U.S. needs foreign talent. It's essential to us, including Chinese talent. So you- We should be getting every Chinese great scientist into our university system and then get their

families over here and get them paid, get them stock options on a five-year vest and make it impossible for them to go home. Like we should take the approach that an authoritarian country would have to not make people leave. We have to take the capitalist approach with honey, not the stick and the shackles, but the honey and the stock options to get them to stay here. This is a national security concern. Pay them off with citizenship, equity, and a lifestyle of freedom and prosperity. Again, let me tip my hat to my colleague, Paul Mosier. I was lucky enough to work on a story with him about this very thing. And we talk about, you know, the complications here.

And, you know, I think it's, you know, not a lot of people understand it, right? It's very easy to say, you know, we don't want, you know, researchers from ex-country coming in the U.S. It's not the right way to think. No, we need to, speaking of like security, honey trap would be good. Okay, as we wrap here, and Drason Harwood's back substack and Clubhouse, those two platforms are explicitly trying to steal journalists and make them independent. We have this tech versus media back and forth that you and I, you know, I think can reasonably discuss because I've got my feet on both sides of it and you're a very reasonable person

and your old school kind of classic journalist. What is the vibe inside of journalist circles, not your opinion because you're very good about not having one and sticking to your knitting of being a reporter. But I'm curious, you know, you had Balaji, who's been on my podcast a bunch of time. He's not at A-16 anymore, but he had a quote in your story that was particularly gnarly where he said, let's figure out what journalist, I think it was, if things get hot, it may be interesting to sick the dark enlightenment audience on a single vulnerable hostile reporter to docks them and turn them inside out with hostile reporting sent to their advertiser friends' contacts.

You viewed that email. This is a correct email. You would not have printed otherwise because that would be a massive liability, correct? Oh, absolutely. And this was an email that was to Curtis Yarvin, who is whatever, I guess, NeoReaction. I didn't go to graduate school,

so I don't understand what Neo-Reactionary means. But what is the vibe when you read that email and you realize, hey, Andresen Horowitz and this group of people is actually wants to, is in a war with you, and it's become so personal now that I believe they are backing substack in Clubhouse explicitly, not just for the opportunity there,

but to dismantle their adversaries in the press. And Drason Horowitz is saying they're going to be doing publishing and they're going to hire reporters and they're going to create their own new press corps to replace what they believe is too critical of them and they're doing it where it hurts most. They're trying to steal top reporters.

and make them independent. This is a explicit strategy in my mind. How do journalists at the New York Times or other places, back channel, back offices, when you're having cocktails or talking in whatever secret slackroom you're in, what is the vibe about Andreessen Horowitz's explicit efforts to attack journalism and dismantle it? Again, you can't get inside someone's head, right? So, you know, I don't know. These are powerful people putting a lot of money out there. You know, but I can tell you what we, you know, have talked about recently, this has gone on for a long time, right? You know, you know, for years, you know, parts of Silicon Valley have, you know, have said, you know, we need to build our own infrastructure to get our, you know, our voice heard. And, you know,

and it might be getting more extreme now. But this isn't, this is an old thing. And, you know, how much of it is just, I never saw them offering cash money advances to dislodge journalists. Right. That's new. Well, you know, what I will say is that, you know, parts of the, parts of the journalism industry, so to speak, are doing very well. Like, New York Times is a subscription business off the charts. It's subscription, right? It's not about clicks. It's about subscriptions. And there are other, you know, important papers that are moving in the same direction and are really healthy, right? Yeah. What really worries me, and I've noticed this recently, as I kind of do about bliss for my book, is that, you know, I went through like my, my, old hometown newspaper in Raleigh, North Carolina, right? It's not doing as well. Like, local journalism has dried up. And, you know, that's a real concern. And you might have some

big players, which are doing well, but we need more than that, right? We need a wide variety of journalism. And that's a concern. But, you know, people are always going to, you know, want to build their own ways of doing PR or whatever you want to call it or or journalism you know that's just going to happen it's happened for for for years they have a folks 90 minutes with Cade Mets go by genius makers the Mavericks who brought AI to Google Facebook in the world and thanks you've been on the pod this is your fourth appearance if you didn't know 2015 2016 2018 episodes 573 6703 670 73, 808, where we just chopped up the news and talked about it.

It's a great knowing you. Congratulations on the book. Thank you. And thank you for having a reasonable discussion about these issues. I think it's, I would like to see, you know, now we're past the Trump derangement syndrome, sort of horrific thing the country went through. I'd like to see the, you know, the tech industry and tech journalism maybe find a little common ground here.

And, you know, I think part of it is the percentage of coverage. And I've just been asking journalists, like, maybe write about some of the companies doing good things, not just when we fuck up. Just like, remember the old school profile of a cool company doing something fun? Like, can we get one of those for every five times we screw up? Literally, my entire life is like doing crisis management for whatever company if I've invested in has had problems that year. Company screw up. Buy the book, everybody. All right, man.

I'll see a boy chick bagels, maybe get some bagels at some point. Right on, it is a deal. That New York, see, this is, my faith in the New York Times has been renewed by you, and the fact that this story, which I thought was fake news

and a fake headline and link bait, and I ordered the damn bagels from boy chick bagels, and I have eaten a bagel in a half a day for five days. Amazing. Now we get to the important subject. These are legit bagels.

Yeah. I mean, you've had them. Boy chick bagels you've had. Like I said, I'm a stone's throw from them right now. I might have to go right now and get me one. I mean, if you get, I mean, I toasted up an egg bagel from there and a pumpernickel one. I put a little butter, a little cream juice, got a little locks. They got some caper berries going.

Amazing. There you go. Amazing. They did a good job. You did a, whoever wrote that story at the New York Times. Right on. A great job. Right on.

All right. All right. We'll see you all next time. Bye-bye.