Paul Ford: Richard.

Rich Ziade: Yes, Paul?

Paul: I’m still figuring out where I am.

Rich: You’re here.

Paul: Where am I?

Rich: You’re here.

Paul: Okay. So you and I are recording at a studio.

Rich: Mmm hmm.

Paul: Okay. And that’s good. It’s good to see you. Hello.

Rich: Good to see you, too, Paul.

Paul: We’re not alone. We’re gonna explain that, too, in a minute.

Rich: Yes, we are.

Paul: But there’s a, there’s a very brief announcement we should make that’s completely humiliating. Are you ready to make it?

Rich: I love being humiliated.

Paul: I… [laughing] Great. We are changing the name of this podcast for the third time.

Rich: Yeah…

Paul: No, this is the fourth time we’ve changed the name. I’m going to give a really potted history, which is we weren’t quite sure what we were doing and you and I missed having a podcast, because we sold the agency where we had the podcast before. So we started “Ziade and Ford Advisers.”

Rich: Shout out to Josh Tyrangiel.

Paul: Yes, thank you. Gave us a good name. Then we renamed it and we called it “The Aboard Podcast” because we were really committed to this company.

Rich: Yeah.

Paul: But then we renamed it to “Reqless,” R-E-Q-L-E-S-S, the podcast about how enterprise software is changing and changing and changing.

Rich: In the world of AI.

Paul: In the world of AI.

Rich: Mmm hmm.

Paul: Because Reqless means no more requirements. We explained that every single week and nobody really locked into it.

Rich: Yeah.

Paul: And finally we decided, you know—I like it. I liked Reqless. I thought that was a good name.

Rich: It’s a good name.

Paul: But we’re going to take it back, make it simple. And so welcome back, friends, to “The Aboard Podcast.” With the slightly updated branding, but very similar content, a lot of AI, but not just AI. We talk about other things. And we’re at a studio right now. We’re leveling it up.

Rich: Welcome to The Aboard Podcast. Software in the age of AI.

[intro music]

Paul: Okay, that’s enough about us. That is—

Rich: Okay.

Paul: I don’t want to talk about us ever again. Today we have a very special guest.

Rich: We do.

Paul: Patrick Austin. [laughter] Patrick, you are the editor of the cherished publication IT Brew.

Patrick Austin: Hi, Paul. Hi, Rich.

Rich: Hello. Good to see you, Patrick.

Paul: Good to have you.

Rich: We know Patrick.

Paul: We know Patrick. I don’t actually remember how I met you.

Patrick: I cold emailed you.

Paul: Awesome. That’s exactly the way I’ve met some—I think Rich cold emailed me. [laughter]

Rich: I didn’t.

Paul: No, that’s true. I have a very cherished memory of Patrick, which is, he showed up—

Patrick: 14 years ago.

Paul: Yeah, it was a long time ago. You showed up when we opened our last agency, Postlight, you showed up with a bottle of booze. You, like everybody else was kind of—

Patrick: Oh yeah!

Paul: Everybody there—and somebody stole it. I never told you that. Somebody stole the booze. I never got to drink it. But everybody else showed up to see what they could get, and you showed up with a gift, and it was just, I still remember it as one of the sweetest things in the world. And I’m just going to make sure everybody knows that about Patrick. So you’ve been an editor and a writer in many, many different places.

Patrick: Correct.

Paul: You have been around the world of technology and around New York City publishing for many, many years.

Patrick: Sure.

Paul: And now, now you edit a team, and the team is part of the Morning Brew empire.

Patrick: Correct.

Paul: I’m just telling you about your life.

Patrick: That’s fine.

Paul: So that we can sort of just get rid of that and just like, get to something interesting. [laughter] No, we—and you publish a newsletter. It’s called IT Brew. Your team produces it. And it is about the world of information technology. Like, old school, like big systems, so on and so forth. Enterprise software, which is why I’m very excited you’re here because we’re all about enterprise software. It’s very important.

Patrick: Sure.

Paul: So tell me what I got wrong. Tell me, just, like, help the audience know kind of what you’re actually about. And you’ve met Rich.

Patrick: Hi, Rich.

Rich: Hello, Patrick. How many Brews are there, by the way?

Patrick: Ooh, that is a great question. I wonder if I can name them all.

Paul: Come on, let’s do it.

Patrick: We got Morning Brew proper.

Paul: Okay.

Patrick: We’ve got IT Brew, of course, some would say the crown jewel.

Paul: The crown jewel.

Patrick: Marketing Brew, for all things marketing.

Paul: Mmm hmm.

Patrick: CFO Brew, which I believe is one of our newer additions.

Paul: Absolutely. I have that one tattooed on my chest.

Patrick: Retail Brew.

Paul: Mmm hmm.

Patrick: Is retail. Surprisingly interesting. I didn’t think I would find retail so fascinating.

Paul: What do you find fascinating about retail?

Patrick: It’s this strange combination of dealing with people, because you need people in retail for most physical retail, and the management of all of these different facets of retail. Inventory management, security, stuff like that. I find it super complex in a way that I am grateful not to be a part of. [laughter]

Paul: I get that. We have a friend who’s doing stuff like starting up restaurants, and listening to him describe his requirements is wild.

Rich: It’s hallucination.

Paul: You’ve dealt with some restaurant implementation stuff recently.

Rich: Yeah, it’s just, such—tt’s one of the most dynamic environments to be in, right? It’s chaos, it’s weird edge—it’s all edge cases. There is no normal retail. I mean, there can be, occasionally, but it’s really, all your energy goes into the chaos around retail because humans are chaotic and that’s a part of it.

Patrick: Sure.

Rich: Let me guess one: Screensaver Brew.

Patrick: Screensaver Brew…

Paul: I don’t think there are any screensavers really anymore, are there?

Rich: Aw, shucks.

Paul: It’s just Apple goes to that…

Rich: Just dated myself. Okay, Is that. It is about 10, a dozen? Or there’s many more?

Patrick: There’s…there’s a few more. Healthcare Brew.

Paul: Mmm!

Rich: Okay.

Patrick: Emerging Tech Brew, now Tech Brew.

Paul: Because it emerged.

Patrick: Because…

Paul: Yeah.

Rich: It’s here.

Patrick: Every 17 years. [laughter]

Rich: Yeah, yeah.

Patrick: We’re launching a new one soon.

Paul: Great. Okay, so IT Brew. Do any of the other ones have funny headlines?

Patrick: Yes.

Paul: Okay. Because I just, people need to know: IT Brew has actually ridiculous headlines. And I’ve written back—I’ll reply every now and then to the newsletter and be like, “Really great headline.” And your boss will be like—and he doesn’t know I know you, and he’ll just be like, “That’s Patrick Austin.” [laughter] Over and over and over again. I’ve done it, like, seven times. And we’ll find some other good examples and we’ll link them, but they’re just like, just imagine the worst possible pun about enterprise software delivered into your inbox in a way that actually makes you feel…like, a little bad. And it’s great. It’s really good.

Rich: I have to ask, is IT enterprise software? Like IT…I feel like, “Call the IT guy, the wifi’s down.” Isn’t necessarily enterprise software—

Paul: Geek Squad kind of stuff.

Rich: Yeah, like what—

Patrick: Sure…

Rich: Give us two minutes on the etymology of it here.

Paul: And also kind of, what’s your beat? Like, that’s…

Rich: Yeah.

Patrick: Sure. Well, I think calling the IT guy is enterprise, because if you’re calling the IT guy, there’s an infrastructure around you built by someone that is not you, which would be your..

Rich: Mmm. Mmm. Mmm.

Patrick: Whoever you’re working for. And I assume they are, they would style themselves an enterprise of a sort.

Paul: Right.

Patrick: But yeah, I definitely think IT is key to the enterprise in general because of how much of our, how much of business is done online, how much data is exchanged with each other, and how secure or insecure it all is. It’s very, like, precarious. It seems very sort of cobbled together, in a way.

Paul: Mmm.

Patrick: And it kind of, yeah, it makes me make that face like, “Mmm.” And I’m just like, “Ah, this could fall apart at any moment.”

Paul: You know you’ve correctly defined enterprise software. [laughter] Tell us some of the big players, like, are we talking Salesforce? Are we talking Cisco? What is your world?

Patrick: We’re talking Salesforce. We’re talking Cisco. We’re talking Cloudflare. We’re talking Crowdstrike.

Paul: Have you been to Dreamforce?

Patrick: No. I would love to go to Dreamforce.

Paul: What the hell? You’re an editor. They don’t junk you out to Dreamforce.

Patrick: I send them out.

Paul: Oh…

Patrick: I send my team of intrepid reporters out.

Rich: Mmm.

Paul: You’re a parent now, yeah, you don’t get to go to Europe for a fun trip.

Patrick: I wish.

Paul: Okay…

Patrick: I’ve to got people going to Black Hat, RSA Conference, CES, AWS Reinvent… Reinvent? Reignite? Reinforce?

Paul: I wouldn’t worry about it.

Patrick: I think those are all real. [laughter]

Rich: Those all sound legit. Yeah.

Paul: Okay, so how big is your team, by the way?

Patrick: Got about four reporters.

Paul: Okay. And so, and you guys seem to publish, like, how many stories a day? It’s a lot.

Patrick: Couple a day. We try to aim for 15 a week or so.

Paul: Okay.

Patrick: 15 to 20 a week. Our newsletter runs four times a week. We’ve got an extra newsletter, our IT Brew cybersecurity edition, runs twice a month.

Paul: What I love is you get, every now and then, they’ll send one, and it’ll be sort of like, “You have to click here if you’re still reading.”

Rich: Yeah.

Patrick: Yeah. Love that.

Rich: It’s like a blank white page.

Paul: No, they’re just like, they’re like, “Don’t—”

Rich: Put your thumb on the letter F on your keyboard.

Patrick: Just jiggle the mouse a little bit, I would appreciate it. [laughter]

Paul: Do not—don’t you, stop pretending that you’re reading this. You read this.

Rich: All right, look.

Patrick: Here we go.

Rich: I’ve been in IT. I don’t say that, because it’s not as cool at a party.

Paul: Fun time’s over, Patrick. [laughter]

Rich: Fun time’s over. I’ve been in—it’s like, “What do you do for a living?” “I’m in IT.” I don’t say that. I say, “Digital transformation guy.” Right? Big things are happening. But let’s face it, I think you’re kind of nailing it. When you go to work, you use stuff that you wouldn’t use otherwise, because you have to use it to get the work done. And the world, I mean, entire economies have been built around these tools that are, whether you call them enterprise software or IT or whatever it may be, it is a world. And then the spaceship landed. AI, right? And do we feed the aliens? Will, are the aliens here to kill us? How long have you been at IT Brew?

Patrick: Over three years. I’m the founding editor, so I started.

Rich: You started it.

Patrick: They brought me on to start.

Rich: And, and timing-wise, has the AI spaceship landed yet?

Patrick: It was, you could sort of see it on the horizon, you know?

Rich: Yeah.

Patrick: It was like that asteroid predictor.

Rich: Yeah.

Patrick: 2.3%, 5.7%. [laughter]

Rich: Where are we at with that, by the way?

Patrick: Hopefully it’s a lot closer. [laughter] Hopefully it’s a lot closer to 100.

Paul: Let’s get this over with. I’m just like, I want this done with. Yeah.

Rich: Okay. So it wasn’t like, “Oh my God, this is changing everything,” yet.

Patrick: Not yet.

Rich: Okay.

Patrick: You can definitely see, like, the pitches come in and the stories and the stories go out, just, like, as more and more money was invested in it. And from my perspective as a, you know, thinking about it from an IT perspective, I’m like, this could be very interesting. You could have huge amounts of, like, data service logs and stuff like that analyzed by a robot that’ll do it a lot faster than your dozen-person IT team—

Paul: Yeah.

Patrick: —that has to comb through all of it and look for any sort of weird behavior.

Rich: But let’s specifically, I mean, obviously, I’m thinking about it from, like, just a pure, it landing on everyone’s consciousness, even non-technical people.

Patrick: Sure.

Rich: ChatGPT landing, and it’s, it’s pretend typing back at you and it’s really good, was a moment that’s after you. You had started IT Brew.

Patrick: Correct.

Rich: Okay.

Paul: But that’s a very non-enterprise experience.

Rich: No, I know, but I want to get Patrick’s read on, like, “Oh boy.” Like, tell me your reaction when that thing lands, and it’s, like, starting to consume all the headlines around tech.

Paul: But let me also, the way that I understand you experiencing the world is people—you, you have a team and they have to go do stories. And there’s a few things that are obvious, but a lot of it is just PR just like attacking you.

Patrick: Totally.

Rich: Okay.

Paul: Things are like—so I’m, the way that you would be hearing about AI in this space first, like, you might see some articles about it, you might talk to some people about it, but literally every PR person in the world is going to be going, “We have now enabled our blah, blah, blah.”

Rich: That took a minute. Well, how did it play out? Because you’re not the consumer side—

Patrick: No.

Rich: —where people can spend, you know, a little bit of money and get, talk to the robot. How did it play out? What did you, what was your initial instinct around what was happening?

Patrick: I wanted to avoid it, frankly. [laughter]

Paul: Yeah.

Rich: Interesting.

Patrick: Just because, like, like initially I was, you know, you see the ChatGPT and OpenAI stuff and you’re like, this is like a computer trying to talk to me. This is like a giving me—

Rich: Bots.

Patrick: Nonsense words.

Rich: Bots are back.

Patrick: It’s your AIM chatbot from 1998 come back to haunt me. [laughter]

Paul: You’re a gamer, right? You’ve seen—

Patrick: I sure am.

Paul: AI-style, ridiculous technology chatting for years.

Rich: So you were unimpressed initially?

Patrick: Unimpressed in the sense that we’ve had these technologies before. They weren’t called AI technologies. They were called procedural generation technologies.

Paul: Mmm hmm.

Patrick: You know, you have software that would make the, you know, playing some video game, some Skyrim or whatever, you’re walking through some wasteland or whatever, you have—you’re not going to model all of these stick trees that have been blown away by some nuclear blast or whatever. You’re going to use a tool called SpeedTree, where you feed it your parameters. You say, “Make 1,000 trees.” And it makes all these 1,000 trees based on the parameters you’ve set. It’s procedural generation.

Paul: They’re really a tool called SpeedTree?

Patrick: It’s called SpeedTree.

Rich: Oh, yeah. That stuff’s cool.

Paul: This podcast is over.

Patrick: It’s really great.

Paul: We’re gonna go, we’re gonna go—

Rich: No, like, all the mountains and streams and trees, you see, it’s, nobody’s sitting there sketching those out.

Paul: I’m aware of that, but I didn’t know it was called SpeedTree. That’s really cool.

Rich: There’s SpeedTree. SpeedMountain Speed… Yeah. There’s a whole….

Paul: How the hell do you guys know all about the different Speed—

Patrick: Pops up when you load your game or whatever. All the middleware pops up and shows you.

Rich: Yeah.

Paul: Really?

Patrick: Yeah. So you have, like, the developer, you know, Guy in Boxers Incorporated or whatever, and then you have all the tools that they use.

Paul: I’m so jealous right now.

Rich: I’m like, secretly, I love Blender and 3D modeling and stuff. I’ve always kind of kept an eye on it.

Paul: I wouldn’t say it’s a secret. [laughter]

Rich: Okay, all right. So you’re not impressed.

Patrick: I am not impressed by generative AI in the form of ChatGPT and text generation, image generation, video generation.

Rich: Okay.

Patrick: I’m not impressed with that.

Rich: You’re saying that present tense. You’re still not impressed?

Patrick: No, because they’re doing it in a way that is immoral and unethical. Sure, I could totally do—I could solve the cure for cancer if I could, you know, gather a bunch of humans in a room and experiment on them for 100 years or whatever. But we have rules that prevent that.

Rich: You had a moral and ethical reaction when you first saw these tools?

Patrick: Yes, because I’m a writer.

Rich: Wow, okay. So that’s where we are. What about—

Paul: What about code and consulting?

Rich: I want to hear about both. There’s an impact on industry.

Patrick: Sure.

Rich: Which, park the moral and ethical end of it. Obviously there’s no way we can deny that it’s having a profound impact in a lot of different ways.

Patrick: Totally.

Rich: And I do want to hear your perspective on the ethical and moral side of it as well.

Patrick: Sure. I don’t want to sound like I am completely 100% anti-AI. I think there are many aspects of AI specifically suited for—I think it’s perfectly suited for enterprise stuff. I think it’s a great idea. Use it to detect anomalous network activity.

Paul: We can slide the money across the table now because this is exactly what we think, too. [laughter] I keep going, I’m sorry, for Patrick’s bosses, we’re not sliding any money across the table.

Rich: [overlapping] Many useful contexts in…

Patrick: I mean, you know, use it for, you know, endpoint security on your, on your client-side devices or whatever. If there’s weird activity, use it to support zero-trust login systems and stuff like that to detect if you’re logging in from France or something and you’re based in Milwaukee, Wisconsin or whatever.

Rich: So monitoring…

Patrick: It’s all about monitoring and detecting weird behavior.

Rich: Okay.

Patrick: And I think that is where the magic happens. There’s, there’s certain stuff, you can use it for coding when it comes to like drafting code and stuff like that. But again, it goes into another issue I have with AI, which is the fact that it’s not—my issue with artificial intelligence is it’s not intellect. It’s using sort of weighted calculations and the data to spit out something that you, that would satisfy you, that it thinks will satisfy you.

Rich: Mmm hmm.

Patrick: So if I ask it to generate some code, what guarantee do I have that this code is accurate or even usable? It’s just based on a billion lines of code it’s already seen. And it’s sort of like this looks right.

Rich: It’s a clever database.

Patrick: It’s a clever database.

Rich: But is that an ethical question or is that a, “I’m not impressed” reaction, or both?

Patrick: It’s overhyped, in my opinion, in that regard.

Rich: In the code gen.

Patrick: In the code gen, it works.

Rich: Yeah.

Patrick: A lot of the time.

Rich: Yeah.

Patrick: But I think it is somewhat overhyped.

Rich: Okay, explain that.

Patrick: If I ask it to generate some code, I still have to, like, check its work. It’s like ask. It’s like asking your kid to make dinner and you’re like, “Did you…you sure you put enough salt in here?”

Rich: The chicken’s kind of pink.

Patrick: Chicken’s kind of pink in the middle.

Rich: In the middle, yeah, yeah.

Patrick: So like—

Rich: Okay.

Patrick: —there are definitely, and I’m sure it’ll get better or whatever. There will be some checks and balances. That’s what they’re spending all this money on. Well, they’re spending all the money, you know, on lawyers, but you know. [laughter] You know, I imagine—

Paul: Definitely not spending it on writers, yeah.

Rich: If that does—do you have, I want to get to the ethical piece soon.

Patrick: Sure.

Rich: Do you have ethical questions with the generating code, or moral questions?

Patrick: Generation of…

Rich: That seems to be your line.

Patrick: Is my line.

Rich: Ah. Explain that.

Patrick: If I’m generating text.

Rich: Mmm hmm.

Patrick: Right? And I’m OpenAI. I go to Paul Ford, author of the amazing novel Gary Benchley, Rock Star.

Paul: Oh boy, that’s a deep cut. Thank you.

Patrick: And I say, “Paul, I’ll give you $1,000 if you let us use this book to train our model.”

Rich: Paul can make that call.

Paul: That’s like $100 for everyone who read the book. [laughter]

Patrick: There you go. That’s a steal. He says yes. Cool. So now every time I ask it to make me a sandwich, it talks about fretboards and guitar strings or whatever.

Paul: Mmm hmm.

Rich: Mmm hmm.

Patrick: To get to the level they’re at now, they’ve essentially stolen so much information and data. They’re not paying anybody. They’re violating copyright law. They’re stealing from artists.

Rich: They’re ingesting everything.

Patrick: They’re ingesting everything.

Paul: I mean, it seems very likely that like Meta just ate all of that giant Russian archive of like all the books, all the papers.

Rich: Is that true?

Paul: Yeah, that was supposed to be the kind of, like, you know, the little secret library that, you know, if you needed a research paper or whatever, and you’re an undergrad with no money, you were supposed to get it. And, and then they just went and were like, “Oh man, we got all the books right here.” So for you, you’re like,”That is my line. I don’t want to use this stuff because it just kind of helped itself without taking care of anybody else.”

Patrick: Correct. And it’s wrong a lot of the time.

Paul: Yeah, that is true.

Patrick: Like it’s super incorrect. Companies are, you know, if you’ve got like Google and Microsoft pushing it out before it’s ready, and you’re like, “How many rocks can I eat every day?” And Google’s like, “Here you go.” It’s really half-baked. And I think it’s a sign, it’s a sign of a lack of sort of innovation and creativity from these companies in the genAI space. And I find it very disheartening what the end result, what they want the end result to be is, which is no more creatives. No more creatives. Just churn stuff out. I am Stanley Kubrick because I typed “make 2001 Space Odyssey” into OpenAI’s Sora, and it makes some video of, like, some weird, twisted video.

Paul: Yeah, but then the, but the monkeys are all Lego, so that’s cool.

Patrick: The monkeys are all Lego, so it is cool.

Paul: Yeah.

Patrick: That’s my issue. You want to, you know, ethically source your data, go for it, and then you can be like, “Hey, write me a story to put my kid to sleep,” or whatever.

Rich: Right. They’ve sort of unilaterally just unleashed it. It’s ingested. That’s why you got these lawsuits. Like, New York Times is suing. I think many other publishers are also suing. Some are signing deals. I think Bezos, the Washington Post signed a deal.

Patrick: I think Vox Media also signed a deal.

Rich: Vox signed a deal to, like—

Paul: I think there will be—look, Anthropic is going to the universities and saying, “I have Anthropic for education.” The universities are in a really strong position to say, “That sounds great. We need structured ways to bring AI into these universities, like us, those big universities, us state schools and whatever. But I cannot—I need to know the provenance of everything. I need a clear citation trail before I can give this to a bunch of undergrads. Can you provide that?” The answer today does appear to be no. Like, they just kind of grabbed everything.

Rich: Yeah.

Paul: And so I think, like, I would prefer that. I would prefer litigation protection and a clear provenance for the stuff that goes into the beast.

Rich: I think this is going to ultimately end up being a negotiation between either sign a deal, here’s some money, and you say, “No, I want more.” And they’ll say, “Well, no, we did it anyway,” and I’ll say, “Well, I’ll sue you.” And then they sue them, and then we waste a bunch of money on lawyers, and then they—which eventually ends up in a settlement anyway. Right? Ultimately, this sort of levels out with money. Sadly.

Paul: We’re not going to individual creators, but to large corporate rights holders.

Rich: Well, I mean, rights holders. They’re all bucketed anyway, right? I mean, that exists today.

Paul: So, wait, let me—I want to take this in a slightly different direction, which is, okay, here is Patrick. Patrick looks at this, and he sees a little bit of the void. He’s like, “Oh boy, here we go again.” Now—

Rich: Patrick’s a writer, too, and he works with writers. So I think that probably color some of your sentiment.

Paul: I completely get it right. And I’m not necessarily—I don’t even think Richard and I are very far from you. What I’m, what I’m actually seeing is, like, now you have this job and you have to tell the story of the IT industry back to CTOs and CIOs. And your job has advertisers from giant companies who are probably all in on AI, excited to tell you about it. How do you triangulate that? Right? Because you need to keep your team aligned with what you believe to be true and real. And you also have to make sure that the advertisers aren’t alienated by your product. And they’re not. Like, I see the coverage, it’s funny, it’s bright, it’s not afraid to be critical. It’s a classic publication. So talk about how you manage that on a day to day basis.

Patrick: I really just strive to give it to you straight.

Paul: Mmm hmm.

Patrick: The good, the bad, the ugly, like, all of it. I don’t want to mischaracterize anything or mislead anyone with our reporting. We write a lot about code-generating AI and we talk to people who are pro that, who say “Yes, it helps, we move a little faster,” stuff like that.

Rich: More productive.

Patrick: “We’re more productive,” all of that stuff. I also talk about the negatives, which is that—things happened last year. Samsung employees getting in trouble because they’re putting proprietary code into these AI systems and the AI system is eating it and incorporating it into itself. So—

Paul: Wait, I missed this. Even though I’m an IT Brew very loyal subscriber. So wait, so…

Patrick: Bookmark it.

Paul: I should. Well no, I get it in the, in the….

Patrick: It’s not your homepage?

Paul: I get it in my mail, it comes in the email.

Rich: Homepage!

Paul: So wait, wait. Okay. So what happened? They were—

Patrick: Print edition. [laughter]

Paul: Please. I like the centerfold. The, so tell me about what—so Samsung just started like using…?

Patrick: You know, some developers start using, using the code-generating aspects of these AI models. It’s taking that code and incorporating it into its sort of into its, you know, model, its language model or whatever. Because it has to, because it’s talking to you about, it’s responding to your code. So it has to be able to adjust based on your inputs.

Paul: Mmm hmm.

Patrick: And that’s not code that Samsung wants to feed to an AI model.

Paul: Mmm hmm.

Patrick: So now there’s systems in place where these companies are like, “Do not feed code into these models. We’re making our own internal models that you can feed this code into.”

Paul: Interesting. So at a scale like Samsung they’re like, “We’re just going to have our own little universe. You’re not going to talk to anybody else, and we’re going to have our own stuff.”

Patrick: A ton of these companies are doing it. PricewaterhouseCoopers is doing it.

Paul: I love those guys.

Rich: Because they’re afraid of folding in code that isn’t theirs.

Patrick: They’re afraid of folding in code that isn’t theirs. They’re afraid of the copyright issues that would pop up if they fold in stuff that they haven’t acquired, you know, through sort of legit means.

Rich: Mmm.

Patrick: They want to be sure that the data that is in there is the data that they want to be in there.

Rich: Got it.

Paul: I wonder if the big AI companies end up creating these models that are kind of like, aligned with stuff like this. Like, this sort of like, “Oh, here’s the base model you can build up from, and you can work with this. And these are, here’s every single source that’s in here, and we vetted everything, and the lawyers give you a big thumbs up.” Because I think that is a valuable product. I would not get—we have, a friend of ours gave us a big spreadsheet of really valuable information, and I would not upload that to one of the big alarms. It’s just too risky, right?

Rich: Yeah.

Paul: Because it’s stuff that’s really proprietary to his company, but he wants to experiment with analyzing that data with AI and sort of the log files stuff.

Patrick: Sure.

Paul: And it’s a real quandary because it’s just, it’s not my own—like, I think this is real. Like, I got really into DeepSeek for a minute, but it’s like, I wouldn’t give it my email. Right? [laughter] Like, I wouldn’t… So there are these boundaries.

Rich: Yeah. I feel like we’re in a moment right now where there’s been such a dramatic commoditization that’s happened almost instantly that is going to take us—first off, it’s like stages of grief. Like, it’s, first we’re appalled by it, and then we’re sort of trying to figure out how to—and then eventually, we sort of embrace it. I’m not happy about the fact that musicians make very little money today. I’m not sure how much money they made when there were record labels except for the superstars anyway.

Patrick: Sure.

Rich: But that’s not to diminish the fact that the economics exploded over a few years. Napster on down, the game changed entirely. And I think, you know, it makes me think of, this a weird analogy, but I’ll say it anyway. Taxi medallions. Like, taxi medallions—

Patrick: I was going to say the same thing about Uber. It reminds me of Uber and how Uber grew.

Rich: I mean, there was a day when I think it was a couple million bucks to own a New York City taxi medallion because you built, you could work, work off of it for 40 years until you retired.

Paul: Yeah, but the people who owned all the medallions were such good people.

Rich: Well, that’s the other thing—

Patrick: Salt of the earth!

Rich: Exactly. It wasn’t the drivers— [laughter]

Paul: Simple country medallion-portfolio owner.

Rich: No, no, it’s true. I mean, fair. The reason I want to, I’m saying all this is you know, we can be upset and offended and it’s totally understandable, and then, like, five years pass and I was like, well, this is the new way the world kind of works. Right? And then we just sort of adjust. I don’t—

Paul: Well, the adjustment is interesting there. Let’s stick with that metaphor for a second, because—

Patrick: I think we’re in the bargaining phase right now.

Paul: But it’s not just that.

Rich: No—

Paul: Then you had Lyft, which was a little better behaved and a little more like open about where the money came from. The drivers were like, “I don’t care, man.” But still.

Rich: There was a point where they had, like, five of them.

Paul: Yeah, it was funny. And they, sometimes they’d have multiple phones.

Rich: They’d have multiple phones, yeah.

Paul: Then there were the atrocity apps that the taxi-friendly companies would produce and they’d be like—

Rich: They tried to get back in the game.

Paul: Aw, and it was so bad. It was just like, you just couldn’t—it was like, you can pay, you just have to put your Social Security number in.

Rich: That ship had sailed. It was too late. It was too late.

Paul: And the thing is, you know, what you actually saw is the gap was product. Like, the Uber and Lyft product teams, they’re good, they get paid incredibly well. And then they’d be like, “We can do that. We already have all the taxis.” But that didn’t matter.

Rich: Yeah.

Paul: Right? And I think, so let’s analogize to where we are now, which is, like, there’s very little regulation, things are a little bananas, and they have gone in and they really have, they’ve just said, “I’m going to help myself, thank you.”

Rich: You know what it is? Laws and enforcement of laws are just slow. It’s too slow. And these companies, they know how to play and they just let the beast out. And it’s too late. And—

Paul: Well, the bleak thing is that the timing for a good, solid, commons-protected regulatory framework is not right now. Like today, that’s just not—

Rich: It’s a free-for-all. It’s a free-for-all right now.

Paul: We’re going to have drunken, smoking gorillas screaming all day running the Department of Defense. And like, that’s kind of where we’re at.

Rich: And so let me ask you this. I mean, there’s the looking back on, frankly, exploiting creative works to feed these LLMs, and it’s happening without permission, in a lot of cases. Let’s be real. As you look forward, put aside exploiting the past, but looking forward, do you see this change, this technology as positive as potentially something that is gonna, like good, positive things will be born out of it?

Patrick: I see it as a productivity tool more than anything else.

Rich: Okay.

Patrick: I see it as the drafting table for the final product you want to make.

Rich: Okay.

Patrick: I see it as something that should not be public-facing in terms of, like, text and imagery and all that stuff. It shouldn’t really be public-facing.

Rich: It should be a tool.

Patrick: It’s a tool, it’s a concept, it’s a conceptual tool in my, in my opinion. And I think that’s where it would shine the most. You want to be Stanley Kubrick? You want to go to Universal and be like, “I need $10 million to make this movie” or whatever, “here’s what I think the movie is going to look like.” And then you have your OpenAI Sora 5-minute pitch deck generated, and they go, “Cool, go.” And then you hire your DP and your director and blah, blah, blah, and you go make your movie.

Rich: What if you don’t? What if you just go back to Sora?

Patrick: What are you pulling from? Are you pulling from, you know, Kubrick’s archives or whatever? And then he’s like, “Do I get a cut of this?”

Rich: No, no, you’re not pulling from—

Paul: It would be very strange if he did that. [laughter]

Rich: Well, I am the future director. I am the future director of film. I’m just really good at prompts.

Patrick: Aw—

Paul: Let me, hold on—

Rich: I want him to react to this. Because someone’s gonna make a movie off of prompts.

Paul: No, but, like, the Kubrick thing is throwing us—

Patrick: Prompts, man. Prompts.

Paul: Let me, let me throw this back to you, which is cutscene in a video game.

Patrick: Somebody’s gonna make that.

Paul: Somebody uses a tool that can actually do some 3D rendering, like—

Rich: SpeedTree

Paul: SpeedTree, but coordinated with AI. It’s, like, the boundaries are fuzzy.

Patrick: I mean, sure, but SpeedTree is a—

Rich: It’s math.

Patrick: It’s math. It’s—

Paul: Not to the trees.

Rich: Fibonacci action.

Patrick: So you can—right, you can, if I were making a less than stellar argument, I’d be like, “SpeedTree is basically AI,” but it’s not.

Rich: Yeah.

Patrick: AI is based, again, based on incorporated data that is sort of measured and weighed differently and output based on your prompt.

Rich: Mmm hmm.

Paul: SpeedTree is a procedural algorithm for creating a bunch of trees.

Patrick: Correct. And they use it, they use it for, not SpeedTree, but there’s procedural generation for levels for like, like, we’re just saying mountains, whatever, blah, blah, blah.

Rich: Like, there’s like literally a database of like different kinds of trees.

Paul: But I’m literally trying to imagine Patrick going to, like, a congressman and trying to explain this to them. I don’t know if we have any hope for this level of—

Patrick: Congressmen don’t even know how email works.

Paul: They don’t know how plumbing works. [laughter]

Patrick: It’s a series of tubes.

Rich: Yeah.

Paul: Exactly.

Patrick: So, like, I think that’s definitely—I mean the issue is that lack of technological literacy when it comes to these, when the people, when it comes to the people making these regulations.

Paul: Yeah.

Patrick: So these companies, you know, OpenAI and Anthropic and all these guys, they have the advantage. They’ve got the nerds and they’ve got the lawyers that are nerds. And you’ve got this 80-year-old Congressman who’s like, “Does when I type, check my email into OpenAI—”

Paul: “Where are my pants?”

Rich: Look, man, this is power and money kicking in. I mean, you know, the big AI firms are going to have lobbyists.

Patrick: Sure.

Rich: Are the small publishers and the writers going to have lobbyists? Probably not. Let me ask you something. I’m a writer.

Patrick: You’re really good.

Rich: Thank you.

Paul: He actually is good.

Rich: Appreciate it. I’ve been writing for years and my writing is really the sum of all the knowledge I’ve taken in reading other people’s stuff. I’ve never plagiarized anything unless I put it in quotes.

Patrick: I was just thinking about this in the shower today.

Rich: Yeah. AI—

Patrick: You were gonna make this argument.

Rich: Yeah.

Paul: You’re a morning shower guy?

Patrick: I am.

Paul: Okay.

Rich: When do you shower?

Patrick: Morning.

Paul: I’m a morning guy.

Patrick: Morning Brew, Morning Shower,  Morning Coffee.

Rich: I shower at night.

Paul: A lot of people shower at night because then the bed doesn’t get messy.

Rich: Also, people go to the gym late.

Paul: Yeah, that’s right. It’s about the gym for me.

Rich: Cool. This is sponsored by Irish Spring. [laughing] It’s fine. It’s not actually spitting back out what it learned. It’s inspired by it.

Paul: That’s a great devil’s advocacy.

Rich: No, it’s inspired. I’m a writer. All of my writing is driven by and really a synthesis of all the knowledge I’ve taken in in my life and the things I read and the people I talk to. Am I plagiarizing? Am I exploiting that knowledge? Is AI really doing that? It’s not—it’s inspired, for lack of a better word, by the knowledge it’s taken in. And it’s now putting out new artifacts that, yes, are inspired. Like, just as a writer is copying Paul Ford’s style, but not copying his words and stealing them.

Patrick: Well, we have a word for that. It’s called being derivative.

Rich: Okay.

Patrick: I don’t think the AI is being derivative. I think it is, it’s making calculations based on the input you give it with the prompt, with the data that it’s using.

Rich: Yup.

Patrick: With the weights that make it more or less, what was the, what did they say the other day? They made it, they made it less sycophantic.

Paul: Oh, yeah, it was getting real bad. [laughter] ChatGPT. It was just kissing—

Rich: It’s going through it right now.

Patrick: Yeah, it just reminds me of, you know, inspiration. You’re reading a book, you’re inspired. That is not you doing, running an algorithm in your, in your brain. Although if you’re, if you’re the devil’s advocate, you’re going to be like, “The chemical reactions in your brain are kind of an algorithm” or whatever. [laughter] But, like—

Paul: That is—

Rich: That was the “mocking devil’s advocate voice.”

Paul: That is what, yeah, no, but that’s what, like, AGI people love to say stuff like that.

Patrick: Yeah!

Rich: I mean, that is their ultimate defense is AGI, is that it is mimicking the thinking and reflection of the human mind. Right?

Patrick: Like, that is alive. It’s got a soul. Remember that guy from Google who got fired? He was like, “The AI is alive.”

Paul: There was a lot going on.

Rich: Wait, what?

Paul: We’ll talk about that some other time. [laughter] It’s not a comfortable subject. No, hold on.

Rich: But that, so that doesn’t persuade you. It’s still, you still have strong.

Paul: Not only are you not going to persuade Patrick, but actually what I would say is I actually want the editor—I don’t agree with you on some of this stuff. Some of it I do.

Patrick: Sure.

Paul: But I want the editor of my publication about this industry to be deeply suspicious. I actually think that’s super valuable.

Rich: I agree. I agree.

Paul: And so like this filter that you’re giving me, it’s funny because like, even though I don’t necessarily agree—

Patrick: What don’t you agree with?

Paul: I think my line on generation is different. I think a lot of it, I think a lot of what you’re saying does sort of accord. And I think that the hogs at the trough aspect of the companies is rough. And I’m really not surprised to hear that big organizations are building their own LLMs under kind of like with litigation—

Rich: Walled in.

Paul: Yeah, litigation, people don’t understand how much litigation runs the world that you’re in. Like, everyone is afraid of getting—they’re vulnerable to being sued all the time, in a way that somebody like typing stuff into ChatGPT never has to worry about. I think the utilities are broader. I think that these tools are really good at two things. Syntactic structural tasks like writing code or making outlines or to do lists.

Patrick: Sure.

Paul: And at bureaucratic tasks they’re really good at, like, “Please, I need to be ISO 9001 certified, look at these PDFs and tell me how to do that.” Not—the creative stuff, to me, I almost see it as weirdly orthogonal, because what I think will happen is nobody will want it. Like, it’s not unless it’s like really like specific—

Rich: There’s already a sameness.

Paul: There is. It’s the glaze—

Rich: Uninspired.

Paul: Yeah.

Patrick: It’s shiny and latexy.

Paul: The word that’s starting to show up is glaze. AI glaze. And I think that’s real. It’s just everything is, like, vibrating and smeary. And so like, to me, I almost see, like, yes, there’s this moral problem over here, but it’s almost more of an aesthetic problem overall. And I just assume that humans will reject it because we get so easily bored anyway. So I’m like, “Eh, don’t even bother.” But focus on the parts where it can actually make really boring things more efficient.

Patrick: And that’s the part about AI that I love.

Paul: Yeah, me too.

Patrick: That makes the boring stuff more efficient. That make a draft, make an outline for me, check if I’m compliant. These are the things that we should be using it for. But the problem is all the attention, all the resources, both sort of computational and environmental, are going to these AI-generating, or text-generating, video-generating models that are, that are so fraught with, with controversy and don’t really give you anything interesting anymore. It’s all homogenized.

Paul: I know, but Veo 2 gave me an image of a woman cradling a pygmy hippo that was really funny.

Patrick: It’s so cute.

Paul: Yeah, I know. Okay, okay.

Patrick: It’s not real though! [laughing]

Paul: All right, so, so…

Patrick: And another thing is, is the contamination of actual real data, real images. If you type, you know, “goose” on Google Images, are you sure that’s a real goose or is that a, is that a, is that a real bird? You type, you type, like, “colorful bird” in Google. You get, it’s all AI, it’s all generated stuff.

Rich: Is that true? In regular Google search?

Patrick: Yeah.

Rich: Like, it’s polluted?

Paul: This is the, this is the other aspect.

Rich: Mmm?

Paul: This infinite, recursive, content-generating polluting disaster.

Patrick: It feeds on—

Rich: Is that because Google made the birds?

Paul: Everybody’s making everything.

Rich: Or Google indexed other birds that were made that are not real birds?

Paul: All of the above. It’s bad out there, man. It’s going to be, it is a recursive garbage fest. The entire information commons is becoming a Superfund site and there’s kind of nothing to be done. Except to me, I’m just like, “Let’s ride it. Let’s see what happens.” Now look, what do we want to know? Patrick can see more of the IT industry than we can. What do you want to know?

Rich: Well, I actually, I am in IT. I was about to say I’m not in IT, but I am in IT.

Paul: We are. We are.

Rich: What’s the advice you’d give someone in IT, in the context of everything we’re talking about and the way the world’s changing?

Paul: And they’re trying to figure out what to do about AI and they’re trying to figure out what to do next?

Patrick: I would just say learn. Like, learn—

Paul: Oh, come on now.

Patrick: Sorry.

Paul: Learn what?

Patrick: Learn, learn how to operate without AI and learn how to operate with AI. Because the problem that we’re going to, the problem that I see, you know, that I foresee in the next couple of years. I’ve got siblings going to college. One left college.

Rich: Younger.

Patrick: Younger, yeah. My partner, college professor, every time she comes from class looking at these papers, half of these are generated—

Paul: Yeah, we’re hearing about this.

Patrick: —because when you do a real in-person test or you hand them a piece of paper and tell them to write it, it’s nonsense.

Paul: It’s bad stuff.

Patrick: Compared to the, the DocX file they put on Blackboard or whatever.

Rich: Scary.

Patrick: Right. And there’s going to be a lack of, I mean, you see it, the atrophying of society, in a sense, the atrophying of technical skills. I’ve had interviews with like CSOs and I’m like, “What Is the future of, what is the future of AI? What is the future of the worker as they relate to AI?” And they’re like, “Well, they don’t really need to be as skilled as they are now because AI will handle the rest.” And I’m like, “That sounds like a terrible idea.” [laughter]

Paul: It is real. Like, it’s like, imagine—

Patrick: That sounds like the worst idea imaginable.

Paul: Imagine how different America would be if everybody would just read one more book. Just one more.

Rich: I mean, let’s dream.

Paul: Yeah, that would be 300 million more books that were read.

Rich: Yeah. I mean, we’re out of time. I think it raises an interesting question about, has the horse left the barn here in a lot of ways? I don’t know. I mean, I think what you’re saying, when you say learn, it was interesting when you clarified it, which is like, “Learn how not to use it.” Meaning essentially, go build some actual skills and use your own muscles.

Patrick: Yeah. And use the AI tool, like I said, as a sort of conceptual—

Rich: Helper.

Patrick: Conceptual aid.

Rich: An aid.

Patrick: To draft your ideas.

Paul: The wacky thing is it’ll do a good job of teaching you.

Patrick: Sure.

Paul: If you want to learn some of the skills.

Patrick: Absolutely.

Paul: I mean, this is why I keep learning piano, even though I’m really bad at it. I could also just go listen to music. All right.

Rich: Patrick, thanks so much. Really interesting conversation.

Paul: Thank you. If somebody wanted to get in touch with you, how could they reach out?

Patrick: They could, I guess you can email me.

Paul: That works.

Patrick: Yeah, you can email me at patrick@morningbrew.com.

Paul: Okay.

Patrick: I’m on Bluesky. I’m on Instagram.

Paul: They should subscribe.

Patrick: Where I’m at—yeah, frankly you should subscribe to IT Brew.

Paul: And the thing is, if they want to reach you, they can literally just hit reply to the newsletter.

Patrick: I get it.

Paul: Yeah.

Patrick: I check it. I’m like, “Oh, that’s nice.”

Paul: Yeah. [laughter]

Patrick: So, yeah, subscribe to IT Brew.

Rich: Very cool.

Patrick: Yeah.

Paul: Well, thank you for coming on. Will you come on again sometime?

Patrick: Oh, absolutely. Happy to be here.

Paul: All right, good.

Rich: This is The Aboard Podcast, Paul.

Paul: It really is. It’s The Aboard Podcast now.

Patrick: What’s that mean?

Paul: Well, you know, Aboard, thank you for asking. Aboard is a platform that uses AI to accelerate software development, frankly, along a lot of the patterns you just outlined—

Rich: A tool.

Paul: The good ones, like, it helps you get everything started, but the last mile is still really carefully crafted human software.

Rich: Yeah.

Paul: And so because we feel the same thing, which is you just can’t trust AI with your whole world.

Rich: Yeah. Check us out. Aboard.com.

Paul: New website coming soon. We’re doing an event really soon. We’ll announce it in our newsletter, so you should sign up for that. It’s on June 3rd at our office at 215 Park Avenue South. You are invited. Let us know if you want to come and where we’re going to launch and show some new stuff and maybe even just build some software for people when they show up.

Rich: Give us five stars.

Paul: Aw God, I love stars.

Rich: Click on that.

Paul: Patrick, thank you for coming in.

Rich: Thank you, Patrick.

Patrick: Yeah, happy to be here. Check out itbrew.com.

Paul: God, you should—

Patrick: Should I look at—is this my camera? Is this my camera?

Paul: There you go. There’s your camera.

Patrick: Check out itbrew.com, your number one IT and Cybersecurity B2B publication. We publish newsletters four days a week. We have a great team of reporters, and we’re only getting better. All right, I’m done.

Paul: It’s got the goofiest headlines, too. It’s so good. All right, thanks, Patrick.

Patrick: Thank you.

[outro music]