Explaining AI with Josh Tyrangiel
Explaining AI with Josh Tyrangiel
How has public perception of AI changed over the past two years? On this week’s Reqless, Paul and Rich welcome on writer and editor Josh Tyrangiel, who’s been the Washington Post’s “AI tourist” columnist since early 2023. They discuss what he’s encountered in various industries experimenting with AI, and the overall sentiments he’s observed as ordinary people grapple with this technology. Plus: He discusses his recent collaboration with Oprah Winfrey on an AI special for ABC News—and the remarkable lettuce she served him for lunch.
Read the transcript
- Learn all about Josh at JoshTyrangiel.com, and read his AI column in WaPo here.
- Exciting news for Rich: Matthew McConaughey actually did a whole series of AI ads for Salesforce.
- Noah Brier on the podcast this past summer.
- Watch “AI and the Future of Us,” Josh’s collaboration with Oprah Winfrey (which was powered, of course, by Oprah Lettuce).
Paul Ford: Hi, I’m Paul Ford.
Rich Ziade: And I’m Rich Ziade.
Paul: You are listening to Reqless, the podcast about how AI is changing the world of software. Rich, why don’t we play that theme song?
[intro music]
Paul: All right, so, Rich, we have a very special guest today.
Rich: We do.
Paul: But we do before we talk about the world—we’re going to talk about the broader world of AI, not just our little enterprise, nerdy corporate pocket.
Rich: Okay.
Paul: Let’s tell the people about our nerdy enterprise corporate pocket.
Rich: Aboard.
Paul: Aboard. Aboard.com. What happens when I go to aboard.com? What is it?
Rich: Aboard.com is a flexible, custom-software platform for companies. You tell us what you need, and we get you a prototype in days, working software in weeks. Give us a call. We’re using AI to ship software that used to take months.
Paul: It’s true.
Rich: Pretty cool.
Paul: You tell it what you want it to build, and it builds the first version, and then you adapt from there. And we’re doing it for people. It’s real. We’re a real company now.
Rich: We are.
Paul: Not just a podcast studio. All right, so I am excited that we have today in the studio Josh Tyrangiel, who is an editor of great repute, who has recently pivoted a lot of his thinking and a lot of his work to understanding the AI revolution. Josh, hi.
Josh Tyrangiel: Hey, Paul. Hey, Rich.
Paul: Thank you for coming on our program.
Josh: It’s my pleasure.
Paul: Josh is usually on programs, on things that have zillions of listeners, and this is a little, this is, like, I’m glad you came in here. We’re in, actually in a basement surrounded by trash cans.
Josh: [laughing] Well, I would have it no other way.
Paul: Josh, tell the people what you do for the paper.
Josh: Yeah, so I am a professional AI tourist for the Washington Post. I write a column about AI. And basically the reason that I got this job is that when it was going down, and I’m talking early spring 2023.
Paul: Like, Sam Altman’s out there with ChatGPT.
Josh: And people are on fire, kids are cheating, petitions are being signed about doom and species extinction from AI. The Post‘s opinion editor called me. We’ve had a relationship. And he said, “Hey, man, we’re bleeding out. Like, our reporters are trying to catch up. Our readers are furious and don’t understand what’s happening. And we think maybe the solution is not a reporter, but a columnist.” And they caught me at the right moment. I was like, “Yeah, that’s cool. I’m a good tourist. I ask touristy questions. And it may be that a tourist’s perspective is needed on this subject.”
Paul: What was the first column? Do you remember?
Josh: So the first column was a, “hey, what the fuck is this”? Right?
Paul: What is it?
Josh: And it was really as much about the technology as about the culture that had arrived with the technology, which is, as you guys know, a couple of big companies, some of whom were doing old-fashioned skullduggery!
Paul: Mmm hmm.
Josh: Like, “Oh, their product is unsafe and it could eliminate all of humanity.” “No, this product is safe and you should not regulate it.” All of which was super confusing.
Paul: And really new, like, you’re landing at the moment where we’re starting to see the images cropping up that were generated by AI, not knowing what’s real.
Josh: Correct.
Rich: It fell into peoples’ hands. It wasn’t an R&D thing that may come soon. Like, it was very—the way it landed wasn’t the video of the robot dog. It was like, “Here. Play.” And then it kind of shocked people.
Josh: Yeah, and it showed up, and one of the things that, I mean, you guys are in the enterprise space, but in the consumer space, most of the time, software arrives and there’s an instruction manual.
Paul: Right.
Josh: This time it showed up and nobody knew what to do, other than it was super powerful. And so what I can’t overestimate for, like, the straights of the world is how overwhelming and confusing the feeling of AI arrival was. [laughing] And so what happened is what happens when people get overwhelmed and confused. They opted out. And they were like, “I don’t want anything to fucking do with this. I got climate change, COVID. Political extremism. I need to worry that Google’s gonna end the world with AI? No, thank you.”
Paul: Mmm hmm.
Rich: Is that—I mean, obviously no surveys were put out. But didn’t most—I assumed most people just used it cause it was handy. And then there was a small, really loud minority that was like, “Don’t touch it. It’s gonna kill us all.”
Josh: The polling is actually pretty persuasive. And there have been a couple big polls. There’s been one national. California, interestingly, did a poll prior to their regulation. 40% of people want nothing to do with AI. They are just utterly freaked out.
Paul: Keep it out of my movies. Keep it out of my doctor’s office.
Josh: I want my job, and I don’t need any of these assholes getting richer.
Paul: Mmm hmm.
Josh: And you’re never gonna regulate it, right? Like, who asked for this?
Rich: That’s a giant number.
Josh: It’s a big number.
Paul: But also symptomatic of the larger sense of distrust with giant social media platforms and so on. They saw, they lumped it in with that.
Josh: And by the way, they should, right?
Paul: No, fair.
Josh: Like, I mean, the people who are advocating for the grand use of AI are Mark Zuckerberg.
Paul: Mmm hmm.
Josh: You know, Satya Nadella.
Rich: Yeah.
Josh: Everybody has both AI and a social platform, and it hasn’t gone great. And so I think that there is a large group of people, not just in America, but around the world who are like, “Hey, fuck you.”
Paul: Mmm hmm.
Josh: “I don’t want this.”
Rich: Yeah.
Josh: Now, a lot of what my job and my column sort of evolved into is like, “Hey, but how could we use it?” Like, what, what could we do? Forget the people who brought it to you. Forget—and I apologize to all enterprise software fans. Forget Matthew McConaughey walking through a street for a Salesforce ad where he’s, like, literally saying nothing.
Paul: Oh, I like when he goes in the hot air balloon.
Rich: That’s a very good ad, though.
Paul: He goes in the hot air balloon! He’s very handsome! [laughter]
Josh: No, guys, you are so deep inside the bubble. When you say, “That was a very good ad”—
Rich: I thought it was excellent.
Josh: That is a marker. When the AI police come, they will put Rich against the wall. [laughter]
Rich: Okay, there’s going to be skeptics. And you’re right. I never thought about it, how it’s always adjacent to social media, which is pretty scary. It’s like, we know what you love and we know who your friends are, and now we can think for you. And that’s a terrifying thing. But in terms of the optimistic side of it, or the productive side of it, I should say, how is that translating? Like, what are you hearing in terms of, “Wow, this is going to make things better.” I mean, I think Altman said we’re going to sort out climate change.
Paul: Pause for one sec, right? You’ve now been at this for a year or two.
Josh: A year-plus.
Paul: Okay.
Rich: Is it tipping?
Paul: Yeah. What’s changed?
Josh: Well, I’ll tell you what I’ve found, which actually makes me really optimistic. In things that matter to us—so let’s say government, healthcare, education—what I’ve found is a process that is beginning to make progress. And that usually involves an organization with a leader who says, “Hey, we’re going to be an AI-driven organization.” A handful of people—can’t be a big organization—but a handful of people who are like, “Okay, how do I translate this into my expertise?” And then aiming to hit a single. Like, one of the things that I think people mistake is like, “Oh, we gotta hit a home run with AI. We’re gonna cure cancer tomorrow.”
Paul: Mmm hmm.
Josh: So I was at the Cleveland Clinic a couple weeks ago, and that’s a place where you’re like, “Oh, they’re trying to cure cancer.” No. You know what they’re trying to do? Couple things. One: Bed-management software.
Paul: Okay.
Josh: Predictive analysis of how long—
Rich: That’s so hot. That’s the hottest thing I’ve ever heard. Bed-management software is a thing?
Paul: You have to settle down. If someone had come to us at our agency years ago and been like, “I want to build bed-management software,” we both would have been like, “This is the most important thing that’s ever happened.”
Rich: Pause for ad. If you need AI-driven bed-management software, reach out.
Paul: It’s true.
Rich: And check out aboard.com.
Paul: We’d love to build it for you. All right, so you’re—
Rich: Interesting.
Paul: Cleveland—
Rich: This is a great observation, because I’m seeing this, too, but keep going.
Josh: So they’re just doing it, right? And they did it with a—they did it with Palantir. Palantir built it on the spot. It’s been implemented.
Rich: Wait—
Paul: [laughing] Lot going on—
Rich: Wait a second—
Paul: Let’s unpack that for just a minute.
Josh: Okay. So—
Rich: The drone people? [laughing]
Josh: So Palantir—and I could, how deep do you want me to go here?
Paul: Honestly, it’s really good for everyone to understand.
Josh: Okay.
Paul: Get in there.
Josh: So Palantir is a company that was co-founded by Peter Thiel and his friend Alex Karp. They have largely done…oh, military stuff, some of which may make you very uncomfortable. There is a reason the Ukrainians are still in the war against the Russians, and it is not that they’re tough, tough though they may be. It is because they have really good software that integrates all of the ontology, all the fields, all the feeds about where people are, where things are, where munitions are, into iPads in the field, and that software was made by Palantir.
Paul: When I think about Palantir, I think about an old version of the web where everything was kind of linked up and semantically connected. And so, let’s find the terrorist in all of the cell phone records.
Rich: Yeah, yeah.
Paul: That’s their kind of thing.
Josh: Correct.
Paul: But this is different.
Josh: This is different because it’s also, it’s not just taking a variety of feeds. It’s doing the sort of machine learning to be able to predict with some certainty where the bed is going to become available. And so they built it with the Cleveland Clinic. It’s not a huge project. It’s a small project.
Rich: Yeah.
Josh: And not only is it great for the business of the Cleveland Clinic, because now they can take in more patients and know when to move patients, but for a patient in the ER, the wait time to get a room is now reduced by two hours.
Rich: Right.
Josh: So that’s like a small, meaningful thing. They’re also doing—I mean, I found a bunch of ways in which they’re doing that, including elimination of sepsis, predicting outbreaks of sepsis. Sepsis is an infectious killer that arises in a variety of different contexts, and in the Cleveland Clinic, it killed 2,000 people last year
Paul: Whoa!
Josh: And that is a normal number. And so finally someone, a doctor, said, “This is outrageous.” And the doctor basically now splits their time 50/50 as an attending physician and a product manager.
Paul: See, this is my larger point. We get thrown into things like LLMs and these environments, and they’re all-inclusive and everybody—or they’re supposed to be. They’re supposed to be able to solve every possible problem. And they keep building software and tools that are, like, “Upload everything, we’re gonna figure it out for you,” so on and so forth. And they’re kind of competing with the operating system. Like, deep down, ChatGPT thinks it’s gonna grow up and eat Windows alive.
Rich: You know, it’s funny, we heard this from Noah, Noah Brier, who was a guest a few weeks ago, and he said something similar to what you’re saying, which is everybody sort of jumped to, like, the sci-fi ending of the movie. Meanwhile, like, we just got this in our hands, like, a week ago, and I’m just trying to do some basic things with it, and it’s really great at some basic things. It’s helping me do stuff much more quickly.
Paul: I mean, you know, what these things are that Josh is describing? They’re apps. Like, they’re applications that were built and incorporate—
Rich: They’re also smaller hills to conquer, right?
Paul: Yeah.
Rich: Which is, you know, to me, I always, I always compare, like, the Valley’s sort of loftiness and futurism with, like, New York’s, like, there’s a lot of people, there’s a lot of foot traffic here. I could put a hot dog stand right here.
Josh: Yeah.
Rich: I can…go.
Paul: The purest example is crypto, right, which was supposed to take over the entire economy. It doesn’t die. It’s still around. Sometimes you wish it would go away. But it’s nowhere near as big as cash.
Josh: Yeah. One of the things that I keep running into wherever I report is that there’s human friction, which is often overlooked by the kinds of people who build software, right?
Paul: Yeah, but without that friction, we would simply fly up into the atmosphere and disappear to space.
Josh: Right, but as an example, when everybody’s trying to hit a home run, sometimes you forget that each situation brings a different response from a human being. Right? So I was just at a conference with 150 mayors from around the world, and I was talking to, they brought a bunch of their innovation people, who were talking about AI.
Paul: Was our mayor there?
Josh: Our current mayor was not there.
Paul: That’s too bad.
Josh: I believe he might have been doing business in Istanbul, if you know what I’m saying.
Paul: Love that guy. Okay, keep going.
Josh: So, one of the guys I met with is like, “So, of course AI comes in. I’m an innovation officer. Everybody wants to attach a chatbot to what their version of 311 is.”
Rich: [laughing] Bolt on.
Paul: Mmm hmm.
Josh: Immediately, put it in there. “Oh, my God. Think of all the things we’re gonna save.”
Paul: “Help me with my water bill.”
Josh: Couple of other—couple of other innovation officers from cities are suddenly leaning into the conversation. And what they told me, which should not be surprising, is that you know what the number one prompt into those chatbots is?
Paul: Porn.
Josh: Get me a person.
Paul: Oh, okay. [laughter]
Josh: Because—
Paul: I was close.
Rich: Same thing.
Paul: I was actually very close
Josh: I mean, in a way, Paul, you’re right.
Paul: Yeah, okay.
Josh: And in a way, we learned a lot about Paul, right there.
Paul: Boy, did we.
Rich: Puts one word in. Hopes for the best.
Paul: That’s what people put into the box every time! Every time!
Josh: Every time.
Paul: Yeah.
Josh: So the notion is that it’s a general-purpose tech, but each use is specific, right? So if you’re putting a chatbot in a place where someone is vulnerable, they’re going to reject it. Now, simultaneously, somebody said, “But you know what it’s great for? We uploaded all of our permitting rules.”
Rich: Mmm hmm.
Josh: Which are—
Paul: There you go.
Rich: That’s useful.
Josh: Hundreds, hundreds of PDFs that no human could possibly collate. And so if you’re a contractor and everybody has a housing crisis, all of a sudden I have this accelerant, right?
Paul: That’s right.
Josh: Now, when Sam Altman and Ilya Sutskever came down from the mountain and said, “We have the technology!” They weren’t like, “Man, I really hope the city of Richmond is using it for permitting.” But give people a chance, and what you’re finding is like, “Oh, okay, this is a single.” And a single is really good, and maybe you start to string a couple of singles together.
Paul: I mean, at scale, you know, solving permitting problems is $20, 30 million in the budget that no longer has to go towards the permitting whatever.
Josh: But even more than that. So we’re in New York City.
Paul: Mmm hmm.
Josh: Where we have a housing crisis.
Paul: Mmm hmm.
Josh: And where people currently, to get a permit for construction, will pay someone called an expediter.
Paul: Sure.
Rich: Still exists.
Josh: The expediter will go stand in line at City Hall and you pay them—but the expediter doesn’t even stand in line, they will put a magnet or an insignia of their name and then go to lunch and come back eight hours later.
Paul: Mmm hmm.
Josh: It’s not a great system
Paul: No.
Josh: But when you think about economic power. So if the contractor can now just go and understand that their submission will be 95% right, a human in the loop confirms it. You’re saving months and months. You’re actually solving real problems. So that’s the kind of stuff that gets me excited.
Paul: I think about this, too. I think government, it’s gonna be really hard because government doesn’t like new systems. And they’re very, the way the procurement and the way the contracting works makes it really hard for new tech to land. But the things that slow down government technology could be radically accelerated with this stuff. And you’re talking billions and billions of dollars that just gets, like, thrown at consultants.
Josh: I’ll even see you one bigger, because I’m not in the enterprise software world. We have a crisis of legitimacy with government right now, right?
Paul: Mmm hmm.
Josh: And some of that has been astroturfed by, hmm, one political party.
Paul: Yes.
Josh: But the crisis is real. And if you want to restore and fight that crisis, one way to do it is with competence. And so if AI and the right tools and deployments can lead to competence and some savings, that’s a really important part of the project of a surviving republic. So I don’t want to get crazy about it, but, like, that’s where I get super excited.
Paul: No, it’s true. If we could get away from Intuit controlling your taxes, and you could have a little conversation about taxes and get some relevant forms back, that would beat the hell out of the alternatives.
Josh: Exactly.
Paul: Okay, okay. So here you are. You’re seeing that the grand promises keep getting made. You’re seeing that singles—stop aiming for the grand slam every time. The grand slam gets promised, but the singles are there. They’re happening.
Rich: I don’t think they’re just there. I think that’s all it can really do well right now.
Paul: Well, also, that’s incredibly sufficient and represents trillions of dollars in value.
Rich: We go into organizations all the time and we watch how people work. Two billion people use a spreadsheet to run their business. Like, it’s not anything that sophisticated. What is so tricky, I think, in this moment is that, yes, AI could solve things like bed management and whatnot. It’s just no one—you have to really actually rethink how you approach what problem you want to solve, because for most people who don’t know if AI can help, they just end up, the conversation starts with, “Hey, I’ve been in pain for years. Do you think AI can help?”
Paul: Throw it back to Josh, because he’s out there getting, people are replying to your columns and they’re sending you emails. Like, how are they reacting? Because people should go read your columns. But they’re very like, “I went and talked to this person and I learned this and these are the things that are happening,” and it’s stories similar to the ones that you just told us about the Cleveland Institute and so on. When they reach out, what do they say?
Josh: Well, let’s go to comments first.
Paul: Let’s!
Josh: Because the comments just never change. It’s all just, this is a conspiracy. Someone is out to steal all our money. How can you trust? How can you—
Rich: Comments are great.
Josh: They have not changed.
Rich: Comments are just great.
Josh: Right? Like, you can set your clock.
Rich: It’s the same two dozen people.
Paul: No, it’s the raisin-cake layer of the entire internet, yeah. [laughter]
Josh: Exactly. And it’s hundreds and hundreds of comments. Basically, all—
Rich: It’s wonderful, isn’t it?
Josh: Yeah. But what is beginning to happen, which I like to see, is that there is a community of people who are like, “Okay, it may be a crooked system whereby a few companies are going to make the most profit out of this.”
Paul: Mmm hmm.
Josh: “I may not trust any of the CEOs of those companies.”
Paul: Mmm hmm.
Josh: “However, I do see something useful here.” And those people are beginning to find each other and talk to each other. And almost culturally in defiance of the big stuff, I’m seeing a couple of really interesting ideas percolate. But what I will tell you is the phenotype of the people who are doing that is the same regardless of industry. If I sat down the people I’ve met in education and in sports and in government, they would all be like, “Oh yeah, I know you.”
Paul: Yeah.
Josh: It takes a very specific kind of person, and it’s different than your old CTO type or your old product manager.
Paul: Well, here’s my guess. My guess is that someone who accepts a certain amount of bureaucracy but would like to cut through it is engaged with new technologies but not in love, and knows they’re going to have to do a lot of storytelling to get people on board, but also doesn’t want to get left behind.
Josh: That’s about right.
Paul: I’m saying this because that’s kind of like where I’ve landed with this. I’m like, all right, we’ll go into the orgs and we’ll help them out because we got to help them out. And actually, I see this stuff as so valuable for the underserved, like the NGOs, who could—I’ve seen so much money thrown after enterprise software for, like, juvenile justice organizations and things like that. They’re going to be able to skip steps if somebody lets them and they learn how.
Rich: I think the advice for the non-technologist who tiptoes towards IT to ask for a thing, hat in hand, and expects to get rejected, they don’t know yet that they have immense power and that, like, the tools are already there to completely question the time estimate and the cost estimate that’s been put in front of them for the last 40 years. Word has not gotten out yet. And it’s not about the CTO. The CTO actually, in a lot of cases will be defensive, because they’ll want to protect the status quo in a lot of ways. Right? But the business stakeholder, who is the customer, it’s the guy who thought it was an $8 hose that blew out his car, and then the mechanic’s like, “The transmission’s shot. I don’t know what to tell you.” And you’re helpless. The ground rules have changed, just no one’s gotten word yet, I think.
Paul: No, I think that’s right. So you went out, sort of in context with that, you recently had a column about how you made a sort of short documentary about AI with Oprah Winfrey.
Josh: Yes.
Paul: Who served you lettuce.
Josh: Oh my God, the lettuce.
Paul: I want to just hear about that for one minute because it was just, it was a part that popped up as I was reading.
Josh: So let me just state for the record, she is wonderful, hardworking, lovely—
Paul: You better—
Josh: And as you know—
Rich: Not a lot of wrong going on there. [laughing]
Josh: No. And as you know, a devoted Reqless listener.
Paul: Oh, yeah, no, I know. It’s great.
Josh: So in a prep session, I went to visit her at her home, and we did a couple hours of pretty hardcore drilling and thinking. We had gathered Sam Altman and Bill Gates and Chris Wray, who’s the director of the FBI—
Rich: By the way, the special, it aired already. Is it available online?
Josh: Oh, it is. Rich, I’m so glad you asked. It’s available on Hulu. It’s available on Disney+. And I would highly recommend—
Rich: Hulu-Disney+-ESPN package, by the way, which is a great deal.
Josh: Well, you can see it on two of those three. [laughter] And Mel Kiper is on the other one. So you got that going for you.
Rich: What’s it called?
Josh: It’s called “AI and the Future of Us.”
Rich: Okay.
Josh: And basically, I went to her home, and after we worked, I was ready to go back to my car and depart, like the, you know, producer, gritty—and she’s like, “Oh, where are you going? Lunch.” So I followed Oprah up to…
Rich: Mmm hmm.
Paul: So many people on earth, their fantasy is to have Oprah turn to them and say, “Lunch.”
Josh: Well, it was. It’s worthy of aspiration, I tell you that.
Paul: I get it. I get it.
Josh: So then we were served lettuce, and she put her fork into the lettuce, held it up to the light and said, “Oh, this is Oprah lettuce.” And I said, “My life is complete.”
Paul: Wow.
Rich: Who said that?
Josh: Oprah.
Rich: She calls it “Oprah lettuce”?
Josh: It was grown in Oprah’s garden, served by Oprah’s excellent chef to Oprah, so she should call it “Oprah lettuce”. And I forever will just call every other salad I eat “not-Oprah lettuce”.
Rich: Was it buttery?
Josh: I’ve never had ambrosia, [laughter] the sweet nectar of the gods. But this was as close as I’ve come. It was—in a way, my life is forever ruined—
Paul: Very few people—
Josh: —because I’ve tasted perfection.
Paul: Very few people have been able to taste Oprah lettuce.
Josh: It’s true.
Paul: Good.
Rich: Well we got the title for the podcast.
Paul: Yeah, so wait a minute. Okay, so you made this, I’m calling it documentary. Program, or…?
Josh: It was, in some ways, a very old-fashioned idea, which is a network television special.
Paul: Mmm hmm.
Josh: That is in the public service. And it’s like, okay, this stuff has been out here for a year. People are saying artificial intelligence, but don’t know what it is. No one’s really gathered the important stakeholders and asked them important questions. And so that’s what we did.
Rich: It’s a 101.
Josh: Yeah, it’s a series of one-on-ones. And she is, what I would say is, like, if you are worried that people aren’t asking enough hard questions, you know, you may phrase them a different way. Oprah has body language that renders a sort of judgment from God.
Paul: Mmm hmm.
Josh: And so her—she is the perfect proxy.
Paul: Yeah, I wouldn’t want that 360—
Josh: No, no. She’s a great proxy for ordinary people and their concerns about AI.
Rich: Amazing, yeah.
Josh: And it was really great to do. It was fun.
Rich: It was really good. It’s worth watching. I’ve watched it, and it’s really, really good.
Paul: Where did you find yourself having to truly boil stuff down? And I asked this not as, like, there’s a context where you could be like, boy, the audience is dumb. But what I’m finding is I’m going out more and talking more about this stuff, and I am realizing, I don’t think of us as being on the bleeding edge, because we’re kind of out there looking at stuff and, like, there’s a lot happening in this space, and there’s tens of thousands of startups and so on. But it actually turns out you can get to the bleeding edge of AI really, really quickly, and it’s very, very confusing to people. So—
Rich: You know what it is, also, we were, like, looking for, we did, like, competitive analysis, like, who’s in the space? It’s so early and there’s so much land that we couldn’t—we found a couple of companies that are sort of doing what we’re specifically doing, which is specific. I think it speaks to just how open-ended it all is right now.
Paul: I mean, sit down with a smart friend who doesn’t know anything, and you’ve got, like a minute, you know, you had to process this, to do this, to do this program. What would you tell them?
Josh: I think you really have to make this a sequence of knowledge, right? And it doesn’t have to be a long chain, but it really does start with what, what is artificial intelligence, exactly? Why is it different than previous—you know, what makes it different from previous software?
Rich: What makes it different from Google?
Josh: Right. And so one of the things that I think Sam Altman did incredibly well is in about 30 seconds, he explained how the collision of a new idea in computer programming and an incredibly powerful bit of compute had to dovetail to get us where we are today. What I would say to people, and I have, is that there’s this great quote from Penn Jillette, the magician. The comedic magician, Paul.
Rich: Like Penn and Teller?
Josh: Yeah. And so what he said—
Paul: The libertarian.
Josh: Yes. One of the things that he said about magic is he’s like, sometimes what people call magic is just an individual having spent an unimaginable amount of time on a task. [laughter]
Paul: Mmm hmm.
Josh: And so take that and apply it to compute, and all of a sudden you begin to understand, like, “Oh, right, this ChatGPT response is actually just an amount of compute that is preposterous applied to my question about how large is Mexico City, for example.”
Rich: What I’m hearing is that it’s worth it to take a minute and explain how it’s doing what it’s doing to calm people down. I’ve seen people react to the thing. They’re like, “Oh my God, what is happening?”
Josh: Well, it feels like omniscience, right? And that spooks you.
Rich: And it feels threatening because you’ve been, you’re a really great paralegal for the last 20 years, and now this.
Paul: It’s also designed, it’s designed to simulate human speech with its output, which if it was producing rectangles with, you know, little sliders in them, we wouldn’t be as scared.
Josh: Yeah, I mean, if you think about War Games, right? A movie important to our youth. Like, when the computer says, “Would you like to play a game?” You don’t think, oh my God, it’s a person.
Rich: Yeah.
Josh: You think, that is a computer in the uncanny valley, right?
Rich: Yeah, yeah, yeah.
Josh: This is different. Especially once you get voice, attitude. It’s like, shit, that might be a person. And so demystifying it to explain why it’s not.
Rich: It’s also really freaking friendly. It’s like, “Yeah, absolutely!”
Paul: It’s so polite.
Josh: And by the way, like, that is the part that spooks me. It’s like, I don’t need more friends, okay? I’m good. [laughter]
Paul: It’s dangerously and disturbingly apologetic. It’s like, “Of course, you’re right. I did that totally wrong. I’ll do it again.” Yeah.
Josh: No one in my life talks like that.
Paul: We don’t have that. We living in New York makes that—I feel that maybe—
Rich: I need it to say, “Not right now, Rich,” every once in a while.
Paul: I think it’s a California thing. [laughter] I think that, you know, they’re just sort of like, “Hey, the vibes are amazing. Let me get that for you.”
Josh: Yeah, yeah.
Rich: Very interesting.
Paul: All right, so I know how to close this, because this is how you have to close one of these. Let’s make some predictions.
Josh: [defeated] Ugh. Oh man.
Rich: [equally defeated] Not again.
Paul: Oh, yeah. Where are we headed? Like, and you know, let me actually frame it this way. I don’t. Don’t. Don’t tell me what magical toy is going to show up. You’ve got kids.
Josh: Yeah.
Paul: What should they be learning and doing with this technology?
Josh: Look, I think you guys are a good example, right? You came of age in a computer age and then gradually moved into the internet. You understood how to use it as a tool that was really important to your, not just your ability to earn a living, but to your self-satisfaction, to your expression. And so the first thing is like, learn to use the tools. It’s the fastest way not to be afraid. I’m looking at stuff like Khanmigo, which is the AI extension of Khan Academy, and it’s like, okay. That’s good product. It doesn’t take the place of a teacher-student relationship. It’s really adjacent. And so I want to onboard—
Paul: Well it’s also, like, Khan Academy was purely about, let’s scale up through video who gets access to education. So this is one more layer on top of it.
Josh: Exactly. And it’s a nonprofit. And so I would say, first and foremost, like, for kids, and for the future, like, find safe spaces where they can experiment with it and get to know it and be really fluent with the tools. I think for us, what will be really interesting is two things: One I’m super afraid of, and one I’m super positive about. I’ll start with the super afraid, which is robotics combined with AI is coming. Nobody knows when. When it does show up, that will make this look pretty simple, because the complexity of having robots that can think for themselves out in the world, and they, they may not be on the city streets where you’re running into them, but you can bet they’ll be in factories. You can bet they’ll be in buildings. Don’t know what that’s gonna be. That’s gonna be weird.
Rich: Oh, it’s gonna be, it’s gonna be, the headline is, “Chipotle robot slings a burrito and kills five people.”
Josh: Yeah, and that’s, by the way, that’s just a headline you see today. We just sub the robot for the human. [laughter]
Paul: It’s true. I’m just imagining a world of terrible, like, sudden, avocado-based death, globally.
Josh: Yeah.
Rich: I think it’s gonna be really weird—like, we chuckle at the weirdness now, but when it’s physical weirdness, we’re gonna be like, “Whoa, whoa, whoa, whoa, whoa, whoa. Calm this guy down. Unplug him.”
Josh: Yeah, I think we’re…five? Ten years away. You guys might have thoughts on that, too, but that is actually too much for my brain at the moment. The stuff that I’m excited about is seeing places figure it out faster, and get to a place where, like, your healthcare experience is actually better. Where I’m not fucking around with United Healthcare on on the phone—
Rich: 40 minutes for—
Josh: —for 40 minutes—
Rich: —prescription approval.
Josh: Where, like, appointment, man—like, there’s a lot of stupid friction in our lives. And as you guys have seen and are, you know, practitioners, like, that stuff can be eliminated. Like, if you hit 15 singles in a row, you’re likely to win the game.
Paul: Well, it’s also one where the market forces are going to get really obvious, like this, I got insurance over here in five minutes. Right?
Rich: I think thinking in singles is the theme here, which is, for us, we’re trying to figure out, like, our sell, which is, because it’s, the aperture is so wide. Thinking and, like, solving smaller things and then building on that is a great way to think about AI for someone that isn’t an expert, a technologist, or an expert or anything like that.
Paul: So how, you know, the PR people listening to this are going to want to pitch you—
Josh: Oh God.
Paul: And they should do that right now. You should give them— [laughing] What’s your home address?
Josh: It’s, ah…Ftrain? [laughter]
Paul: Yeah, that’s right. What kind of, like, what signal are you looking for out there now? Like, where, what are you, what are you sort of, like, who do you want to hear from in the world? I don’t mean, like, Sam Altman, but I mean sort of—
Josh: I want to see, I really am interested in people who don’t just have an idea, but have an implementation strategy. Ideas are—
Rich: Something real.
Josh: Yeah, ideas are pretty cheap, and particularly when you introduce this thing with a huge bang, everybody’s got an idea. What’s interesting is implementation.
Paul: So following the doctor around.
Josh: For instance, going back to the Cleveland Clinic.
Rich: It’s real. It’s out there.
Josh: It’s real. They’re doing it. And they have found this method of pairing clinicians with product people and turning clinicians into product people. It’s fascinating. I wrote about baseball. I’m gonna go back out after the World Series, which I hope the Yankees are not in, and if they are, that they lose.
Paul: How’s that going, Rich?
Rich: Let’s go Yankees!
Paul: Ugh…
Rich: They look real good right now.
Josh: Yeah.
Rich: My friend in Cleveland is pretty bummed out.
Josh: Well, that’s just because they’re in Cleveland. [laughter] So I am going to go back out there to meet a couple baseball players who figured out that there’s biomechanical software which can absolutely understand what each tiny muscle group is doing, and you pair that with visualization of the output, and they can actually change how they play the game, which will then change how they get paid, right?
Rich: Great success stories of actually applying this amazing new tech in the real world.
Josh: Correct.
Paul: Some of that, by the way, is not LLM stuff. Like, some of that is just classic machine learning and visualization.
Josh: Correct.
Paul: Yeah, that’s the thing. I think you and I, obviously, Rich, because we’re building on top of these technologies, but there’s a whole world of stuff that now, that probably needs to come more and more into the light from the last 20, 30 years, that is insanely powerful and it’s been kind of locked inside of places like Google, and at scale, and now that scale is getting smaller.
Josh: Yeah. The ability to, I mean, what AI is sort of secretly best at is translation, right? Like, it can take anything and integrate and translate it with another thing. So suddenly, visualized biomechanical data is a thing. Like, you can actually see what individual—
Rich: It’s useful data.
Josh: It’s really useful.
Rich: That you can apply in other ways, yeah.
Josh: And you can see it and understand it in ways that before were just table after table, and good luck integrating that in your brain.
Rich: It’s so early, as we talk about this, it’s, a lot of people don’t see it through this particular lens.
Paul: Why is it so hard? Last question: Why is this, like, because I’m finding, you know, I have a certain profile as a person who really likes tech and really likes to get involved. And there’s a few people kind of in my cohort, and I feel that we’re all making eye contact and going like, “Whoa, this is real.” But many other people in our cohort are not. Like, when do you think that tipping point comes where people are like, “All right, I better deal with it?”
Josh: I think most people will not.
Paul: Okay.
Josh: I think what they actually will do, and this is one of the mistakes, you know, I’m not, I’m not advising any of these companies on communications, but like, at a certain point, could you just stop saying the letters “AI”? Just call it software, call it product. Many people don’t give a shit what the guts of a thing are, right?
Rich: Or call it something in a narrower, applied way that is relatable.
Josh: Well, I think we get hung up on, like, it’s a thing and we’re going to name it and we’re going to stuff it down your throat. I was giving advice to a friend of mine recently who’s having some concerns and issues that were in the public space about their company. I was like, “Stop talking about your plumbing. Nobody cares.” [laughter]
Paul: Yeah.
Josh: Right? So, like, we can get excited about AI, but for the most part, I think people will not engage until you present them with something that is persuasive and changes the way they do their lives and their businesses. And then all of a sudden it’s, like, oh, this is cool, right?
Paul: That’s true. For me, it went from, this has a few ways, and there’ll be lots of applications to, okay, this will utterly transform the way people program on a daily basis.
Josh: Yeah. So I was talking to president of a city council from a city in Mexico City, right? And she was asking me, “Well, what do you do about the political problem? I have constituents who are freaked out about privacy, and all this stuff with AI.” And I was like, “Well, stop talking about it. Build your trash pickup app.”
Rich: [laughing] Yeah. Make the world better.
Josh: Yeah, and when they don’t see bulk trash on their neighbor’s lawn for two weeks, they’ll just thank you. And if, at that point, you want to say, hey, you know what? AI helped us build it. Do it or don’t.
Paul: They won’t, though, because now they’re thanking them. That’s right. Use the robot to get stuff for yourself.
Josh: Right!
Paul: Yes. Okay. Is AGI, you worry about it at all? You ever think about it?
Josh: Um…
Paul: Robots gonna take over? You talk to these people.
Josh: Yeah, I I’m less concerned about it, largely because I think there’s less money in it, [laughter] and I think that these companies are driven by money, and so—
Paul: Yeah.
Rich: There’s a lot to go get, right?
Josh: There’s a lot to go get.
Rich: What’s at hand. For sure.
Josh: There’s a lot of cash sitting out there. AGI down the road will be a problem, but they’re not, that’s not where they’re driving yet.
Paul: Let’s kick that can, baby.
Josh: Exactly. Exactly.
Paul: All right. All right. So, Josh Tyrangiel, people want to get in touch with you. What should they do?
Josh: They should call Paul.
Paul: That’s fine.
Josh: And then Paul can be like, “Ah, I don’t know…”
Paul: I’ll screen. That’s fine.
Josh: Yeah, I mean, that’d be great.
Paul: Okay. They can email you through the Post website.
Josh: Oh, I don’t, I don’t—I’m an IC. I don’t, yeah.
Paul: Okay, that’s fine. They’ll figure it out.
Josh: Yeah.
Paul: That’s good.
Rich: No one’s figuring anything.
Paul: [laughing] Yeah, that’s true.
Rich: All right, Paul.
Paul: Ah…
Rich: Wrong way to close it. Josh, you can duck back into the shadows.
Josh: Thank God. [laughter]
Paul: Thank you for coming.
Rich: Thanks for coming.
Josh: My pleasure, guys.
Paul: All right. Rich, people want to get in touch?
Rich: Hello@aboard.com. We love to talk to people.
Paul: And we love people in general.
Rich: Check it out, also, at aboard.com.
Paul: All right, let’s go.
Rich: Have a great week.
Paul: Bye.
[outro music]