Paul Ford: Hi, I’m Paul Ford.

Rich Ziade: And I’m Rich Ziade.

Paul: And this is Reqless, the podcast about software in the age of AI.

[intro music]

Paul: Okay, look. I am a little under the weather. You might hear it. This is as good as it’s going to get today.

Rich: I like it. It’s cozy.

Paul: Yeah, no, it’s not great. So I thought we could just have a quick discussion, maybe not, you know, I’m going to summon about 10, 15 minutes of really good energy, and then I can melt down again.

Rich: I’ll take it.

Paul: Yeah. Okay. So look, so we’ve had this conversation on and off over the last couple of months about what AI is going to do to the software industry. Because what we’re seeing is that while generating wacky pictures of squirrels riding bicycles is one thing it can do, and generating lots of content that people, you know, like, robots make content and then robots read it, but no humans are in the loop, it’s doing that, too. It can be used for all kinds of sort of, like, additive things and grammar checking and translation. 

But in our world, which is the world of software, which is tens and tens of millions of people working on relatively dry things, not, like, making the new version of a game—maybe that, but also doing things like setting up the backend for the call center for the real estate company, right? 

Rich: Mmm hmm.

Paul: Like that’s, that’s the world that we live in. And that’s actually the world that kind of drives the multi-trillion dollar consulting industry. And it’s big. And it kind of is, when we talk about software changing the world, that’s actually kind of what we’re talking about. Gonna change the whole economy. 

And I had said, I’d been kind of arguing that, like, I think this stuff is coming fast. And you’re like, “Man, organizations move slowly. I think we gotta, we gotta pace it.” And I think there is a lot of truth in that. I think that AI is not going to suddenly go to a carpet manufacturing firm in the next two weeks and then suddenly they’ll fire all their engineers. That is not going to happen. 

But you and I, recently, we’ve been kind of playing with and using more of these tools, and I think we’re a little bit ahead of a lot of people on it. And you have accelerated your time frame. So before—we were talking about this yesterday. Before, you were like, “I think at least five years before, this stuff is kind of everywhere.” By stuff, I mean a programmer sits down, enters stuff into a chat box and gets code back that then gets deployed to production, and it’s maybe 10 or 15 times faster than it used to be to build software. 

So you’ve accelerated, you’ve accelerated a little bit. You’ve reset your timeframe. And I thought I would just kind of throw it to you and let you talk about that for a minute. What are you thinking?

Rich: Yeah, I mean I was thinking it’s going to take years because change comes slow. And then, I think recently, I said to you, I think it’s a couple of years away. 

Paul: Yeah.

Rich: Let’s talk about what “it” is.

Paul: So two years in enterprise software is like the snap of a finger.

Rich: Yeah.

Paul: It takes six months to get a proposal.

Rich: That’s right, that’s right. I think saying that a lot of change is going to come in two years—so what is that change? What are we even talking about here? I think, within organizations, you’re going to see how success is measured and how value is measured out of IT. And—let’s reset this. 

Paul: Okay.

Rich: When you think about IT, most businesses think, “I’m functioning.”

Paul: Right.

Rich: “It’s up.”

Paul: People are using their computers at their desks, filling out forms.

Rich: [filling out forms noises]

Paul: Order forms.

Rich: Transactions are flowing through—and by transactions, I don’t just mean money. It could be support tickets, it could be purchases, it could be advertising budgets getting worked out. It all happens on software.

Paul: Not-for-profits managing donor relations.

Rich: That’s right. That’s most of the world. Most of the world is maintaining, just trying to keep their heads above water. And these are tools, these are not that different, like, “Is the wi-fi slow for you, too?” It’s kind of the same thing, for most people. 

Now what people don’t often talk about is the world of, like, delivering new tools, new capabilities to help you compete and help your business run better. And you’re, like, okay, well, most people just want the wi-fi to work—no, it turns out there’s a multi-trillion dollar industry around getting better tools, because you want to compete, and you want to outpace your competition, you want to innovate, you want to step ahead. You want to eliminate inefficiencies—these are things that are always being clamored for by business. 

Paul: Sure.

Rich: Right? And the truth is, for many, many businesses, it goes down like this. “No. [laughing] We can’t do it.”

Paul: Can’t do it.

Rich: “You got to give me budget. If you don’t give me the budget, I can’t do this for you.” Or it comes down to large, sprawling, multi-year efforts to “upgrade”.

Paul: Well, and you know the hardest thing in here from years of experience is things that seem incredibly intuitive or not. So I’ll give you an example which is, “Boy, it’s taking so long to update the marketing website. If we could just make this more efficient, we could update the site more quickly and maybe get more leads in.” 

Rich: Yeah.

Paul: Well, maybe. But, you know, I have these other five things I need to do that are directly connected to saving money on our IT spend, rolling out a new parking voucher system for our employees, because parking is such a problem.

Rich: Mmm hmm.

Paul: It’s always weird to go into orgs, because people will essentially be in a zone of despair about the tools that they have to use, and they’ve been prioritized out of existence.

Rich: That’s right. So here’s why I think things are going to accelerate. Today, currently, there is an explosion of innovation around AI in the engineering space. And by engineering, I mean computer engineers. I’m not going to get into hardware and robots and stuff like that.

Paul: People who write Java code for banks. People who write—

Rich: Code. They’re writing code.

Paul: Yeah.

Rich: Yeah. And if they’re curious and they want to fiddle around and they’re watching the internet, the tools are getting incredibly good. Like, they’re just, if you’re willing to almost build a dialogue with them, almost build a relationship with them and understand them, they’re incredibly good. But—and here’s where I go from 5 to 10 years to 2 years—currently they sit in the domain space of engineering. They are inside of the world of engineering.

Paul: They’re built by engineers for engineers. That is true.

Rich: And look, the executives and the business people and the people who treat engineering as a services group or treat IT as a services group are hearing rumblings. You can’t not hear the rumblings. It’s everywhere. It’s on your phone. You know, Apple’s doing it. There’s apps for your phone. So they hear rumblings and they know the engineers are fiddling around, but they are still looking through…you ever see those videos of like the little kid who’s tapping on the glass at the aquarium and the beluga whale comes swimming by?

Paul: Yeah.

Rich: They’re on the other side of the glass still.

Paul: Yeah.

Rich: Okay?

Paul: The engineers are the beluga whales.

Rich: Now let’s blow it all up. And here’s why we will be blowing it all up. It turns out these tools, the utility and the possibilities around these tools—and by tools, I mean these AI capabilities—do not stop with code. They actually work outside of code. And when you go outside of code, you get all these sort of fringe players around how you deliver value to a business.

Paul: Give me an example there.

Rich: Product managers. Product designers. Technical product managers. QA people. Business analysts. The business analyst is a funny one. The business analyst sort of doesn’t want to say the word “product” because they want to be allowed into the room with the executives. These are people who don’t know just yet, and the tools aren’t there just yet, where the same way engineering is seeing acceleration, their world is going to see acceleration. 

But this isn’t about “the business analysts will be able to write their memo faster”. What this is about is the fact that the hops, from business analysis to wireframes to functional spec to unit tests to architectural diagrams to code, those hops are about to get obliterated. And the business analyst can give you a little software, straight-up. 

And that is a wild, wild thing. It is not just about acceleration vertically across professions. It is actually eliminating the key handshakes that happen as software gets built today. And by eliminating, I just mean you could just sort of go. And if you’re not worried about the code, you can get a pretty credible thing going.

Paul: So let me tell two quick stories that might help here. So one is—I’ll do this in reverse order because I think it’ll help people understand. So I decided to do this weekend something really kind of granular and boring using Claude, which is one of the chatbots. And I actually used OpenAI’s ChatGPT, too. And what I decided to do there is a synthesizer that was very famous, called the DX7, and it had a cartridge format in binary that was 4K, just real short little instrument definition files, 32 per cartridge. And you would put the cartridge in and you would suddenly have a trumpet in your synthesizer. Good for you.

Rich: Like a game console.

Paul: That’s right, yeah. Really, very, exactly the same. And so I was like, you know what, I’m curious what’s inside of those and how it might work to take them apart and put them back together. If you’re not an engineer and didn’t used to program, this is kind of a grisly task. Like, the formats of these kind of cartridges are not—they’re 1980s standard, they’re not now. And so I was like, how am I going to do this? And so I tried a bunch of different, I gave it a sample cartridge, I’m like, tell me about each byte in the cartridge and so on and so forth. And I kind of kept hitting all these various walls. And then I said, “Listen, robot, write me a complete byte by byte specification of every element of this cartridge. Make sure to include checksums and other sort of things,” like all these sort of, like, all the sort of gristly material that we have—

Rich: So the output is what from that?

Paul: It is a markdown document that is really impeccable that describes every, the exact sort of, like, structure of the order of each byte and the headers and the footers and the little things that go out to make this work.

Rich: Okay, so just to, just to say this back to you, you fed it an example cartridge.

Paul: I did. But also, then I went and did a little research to figure out—because I couldn’t quite get it. It turned out there were like two little things that it just didn’t see, including a checksum byte. Like, literally it does a little math. And it was like, yes, this is valid. And I didn’t know that calculation. I didn’t know that was there. And it didn’t know either. Once I figured out like one or two little things, I had to do about 10 minutes of googling.

Rich: Mmm hmm.

Paul: I was like, “Write this specification in exacting detail.” And it did it. And then I said, “Okay, read the specification and write me a serializer-deserializer. A loader-unloader.” Right?

Rich: Okay. So you essentially, first made AI produce a spec.

Paul: Yes.

Rich: And then you fed the spec back into AI.

Paul: That’s right. And I’d been dinking around for hours. So actually, so the process was I kept hitting bugs, I couldn’t figure out what was going, it would kind of load and then it would break, I kept getting error messages, I’d give it the error messages and then I stopped. I did a few minutes of just online research, trying to figure out what I was looking at.

Rich: Uh huh.

Paul: I said, “Give me an exact specification,” influenced by what I had just learned. I said, “Now write some code.” And everything worked perfectly in about two minutes. The hours were still there. Right? Like, I still had to spend time kind of getting my bearings. But the reality is, like, I’m going to be a lot f—I’m going to have it write the spec every time from now on.

Rich: Interesting.

Paul: And so we’re going back to, like, what is the business analyst—like right now, we’re in a funny place because it’s very engineering-centric and sometimes it sees the bugs and says, “Do you want me to fix that for you?” But it kind of ends up in a lot of loops. 

Rich: Mmm hmm. 

Paul: But I do think a paradigm is going to emerge where if you specify things really pretty well, the code—and then you say, “Go ahead and make me a prototype based on the specification,” you’re going to get surprisingly close.

Rich: Okay. I think what we’re talking about here, very few people are talking about just yet.

Paul: Yes, very few people are talking about it this way. They’re talking about it as, like, a code assistant, like autocomplete on steroids, or it’ll write you some nice phone functions or so on. But they’re assuming an engineer will talk to it, or that a senior engineer will be able to do the work of lots of junior engineers. 

What I’m realizing as someone who used to operate an agency, is that the thing runs more like an agency. It’s like, you coach it to be a product manager, and then you coach it to be a software architect, and then you coach it to be a senior developer, and then you coach it to be a junior writing unit tests, and then you figure out how to deploy and. Except that instead of that taking six months, you might have a prototype in literally 45 minutes, if things go really, really well.

Rich: Yeah, let me, let me get a disclaimer out of the way. This isn’t like, “Hey, engineers, your days are numbered.”

Paul: It might be. We gotta actually keep that option open. Like, I think it’s—

Rich: Well, I think we can, we can make this constructive.

Paul: I do think change is coming. Let me put it that way. Change is coming and you probably want to be aware of this change.

Rich: Well, listen, man, you didn’t just write a spec—like, if the deliverable was the spec?

Paul: Yeah.

Rich: And then you walk it over to the engineer? That’s cool. But when you had the spec, it just got built.

Paul: That’s right.

Rich: I mean, that’s what I was trying to say before, in a very long-winded way, which is the hops—and listen, man, when you give humans hops, do you know what happens?

Paul: Friction.

Rich: Humans happen. Politics happen. Defensiveness happens. Obfuscation happens.

Paul: Somebody gets a cold, the hop is going to be an extra two weeks.

Rich: He’s in Peru. 

Paul: Yeah.

Rich: [laughing] Our lead engineer is in Peru.

Paul: He’s got other stuff going on.

Rich: He’s got other stuff going on. If you are able, and you’ve said this in the past, and now it’s sort of sinking in for me, you’re eliminating these steps of translation. It is a translation platform. And the truth is, if you feed it clear, clear instructions and details, then it eliminates—because you’re turning a spec into code that then becomes software. If you can make the spec skip the code, because the truth is, if you, if it’s working for you, did you sweat over the code? You’re, like, holy moly, it works. 

Paul: Yeah.

Rich: It works. It works.

Paul: It is good to read everything in outputs, I got to say. Like, it’s not—but, but yes. It’s still going to be, like, literally decades before everything changes. But…

Rich: No, but I think I’m a, I’m a believer in markets.

Paul: Yeah.

Rich: And I think if you are, if you are a young whippersnapper and you want to cause some trouble and find—when you have this kind of opportunity for disruption, opportunity follows it, right?

Paul: Yes.

Rich: There’s, markets get created around it. It may not happen inside of an org, but I got to tell you, the little shops and the firms are, like, “Hey, 90% off. It’s coming. We can do this for you real fast, in record time.” That’s not gonna be fluff. That’s actually gonna be real. And so I think markets are gonna force these changes to occur. I do think some companies will innovate. Like, I think some companies will actually welcome the opportunity to innovate, but—

Paul: I think also it will take a couple of years for the tooling to emerge to do this consistently.

Rich: We’re hacking, right?

Paul: Yeah.

Rich: Like, we’re hosting, like, we’re doing the equivalent of setting up Apache web servers. 

Paul: Yes, that’s right.

Rich: That doesn’t happen anymore. 

Paul: That’s where we are.

Rich: Right? We’re actually, we’re editing httpd.conf right now.

Paul: Wow. That’s like a sense memory. Like, I just—

Rich: Yeah. 

Paul: —traveled back in time.

Rich: I mean, that thing is incredible. Right? And, and I did it. I did it. I used to, I used to, you know, I’m old guard and I used to do that stuff. And then eventually it became three clicks. [laughing] 

Paul: Yeah.

Rich: And that was that. And now everybody’s like, why are you in there? What are you doing in there? 

Paul: Yeah.

Rich: There’s a whole, you know, web admin to do everything.

Paul: Go over here. That’s right.

Rich: Yeah, exactly.

Paul: Use Amazon. What are you thinking?

Rich: Use Amazon. Yeah, whatever. And so I think we like to play. And so we’re inside of it and we’re like, well, I can edit a config file. That’s fun.

Paul: There is an element in our relationship right now because, I’m, you’ve got real work to do, and I’m like off playing with this stuff, and I keep kind of dragging it back into the light and going, “Richard! Richard!” 

Rich: Yeah. No, I think our dynamic, and I don’t mind sharing our dynamic. I think our dynamic is we are just enamored with technological progress. It’s just a joy to us. I tend to think about how to turn it into business. I always bring it back to business. You are constantly poking, like, look at this weird gemstone I found in the cave.

Paul: I’m like, what does this tell us about the past and culture? Right? 

Rich: Yeah, yeah.

Paul: And you’re like, what does this tell us about money? And thank God for that because otherwise I’d be, I don’t even know where I’d be.

Rich: Well, dude, you turned a corner in the cave, and it’s just gleaming gemstones right now. 

Paul: Yeah.

Rich: It’s crazy right now, right? And it’s a wild time. I’m excited about it because with that kind of innovation, innovation leads to disruption leads to opportunity, and I think that’s very cool and fun. And it’s, like, we’re older, we’re not young, but we got, we’re getting to experience this. And you know, the world’s a little topsy-turvy right now, and this is at least a fun little playground to be in.

Paul: One of us was going to have to acknowledge that the world is full of chaos today. But it was you today. I’m proud.

Rich: No, no, I mean the world’s full of chaos. The world is pretty consistent about its chaos. It feels extra, extra sauce right now. [laughing] But!

Paul: Especially for—people in the U.S. aren’t quite used to this level of whack-a-doodle-ness.

Rich: Yeah, and, but on top of that, I think there is chaos coming from within this industry and within this world, and I think that’s interesting.

Paul: I do not think you will be looking at the same kind of IT industry in the next five, ten years. I just, it just will be very, very different.

Rich: There will be new roles, by the way. Like, the way DevOps surfaced and the way… Prompt engineers was a band-aid. That’ll go away as quickly as it came, I think.

Paul: Yeah, I don’t think that’s long-term.

Rich: No, it’s not long-term. You know, AI architects and things like—there’s going to be new roles for sure because the systems will demand different skills. I do think engineers will be able to repurpose their knowledge. But look, I’m not going to sugarcoat it. I also think that for many, they will have to rethink more deeply what they do and how they bring value because of what’s happening. It’s a lot. 

Paul: Well, I’ll be more specific and let’s not end too much on—but here’s what’s real. If you are a junior engineer right now, you’re at more risk than a senior engineer. And if I was a junior engineer, I would really, really get to know these tools. I would, like, double down on them right now, because I don’t think that, hey, you’ve learned a little JavaScript at the bootcamp and we’re going to get you building web pages and figure you out and see if you should stick around?

Rich: Mmm hmm.

Paul: I do feel that this thing is almost like tact—it’s, like, aimed right at that role.

Rich: Yeah, I agree. I totally agree.

Paul: It writes very good React components that can then be dropped onto web pages.

Rich: Yeah. And right now it’s building discrete stuff, but eventually it’s going to keep creeping outward and outward to create more thorough, comprehensive things. That’s just inevitable. There’s just no way around it. You still need a lot of expertise to bring that value back into the fold. And that’s going to get easier and easier. And I think the tools will change. I think that they will evolve. And I think it’s a fascinating time. I think it’s great. It’s nutty.

Paul: The thing that I always try to encourage people to do is just get out of great/not great, good/not good and just kind of look at the thing.

Rich: Yeah, yeah, yeah.

Paul: Figure out what it means for you. Which is very hard today. It’s not really something people want to do. They want to know if it’s bad or good.

Rich: Yeah.

Paul: But, yeah, Pandora’s box is open. All right, Rich, so speaking of opening Pandora’s box.

Rich: Oh! 

Paul: You and I have a company—

Rich: Unboxing?

Paul: It’s called Aboard.

Rich: Yeah.

Paul: Aboard.com. And we use AI. We use also classic great software. We’re working with a couple pilot partners now. We’re a real business now, which is pretty cool. 

Rich: Yep.

Paul: What we do is we help organizations get complicated business software stood up very, very quickly and at much less cost than ever before. We can say ever before with a lot of assurance because we did that for 20 years. 

Rich: Mmm hmm. Mmm hmm.

Paul: And sometimes fractions. So if you’re someone who has to do case management, lead management, tracking all sorts of data across your organization, stuff like that, like, sort of abstract stuff where you just, like, needed to be more accountable and, God, you just need to get it done. You should get in touch.

Rich: We are an AI platform. We deliver stuff really fast with AI. We’re also observ—that’s why the topic of today’s episode is we’re kind of watching the world move so fast around us.

Paul: Yeah.

Rich: And it’s a wild thing. But yeah, we are at the forefront of it and we’re excited to be involved. And if you are looking to do the thing you never thought you had the time for or could afford, you should talk to us because the world is changing really fast.

Paul: Get in there now. Take advantage.

Rich: 50% off.

Paul: All right. I managed to keep the energy up. Okay?

Rich: You did good. You did good. 

Paul: I’m gonna go—

Rich: A little nasally, but you did good.

Paul: Yeah. I’m gonna go take some medicine.

Rich: Okay, do that.

Paul: We have a lot of, we have—you know what I like on the sick days is when you do the, you call into the meetings, and then you’re just kind of blinking for, like, a half an hour. [laughter] Just kind of like, “Yeah, yeah…”

Rich: You’re just staring.

Paul: “Sounds good. Okay. Thank you.”

Rich: Yeah. Yeah.

Paul: So that’s what I have to look forward to for the rest of the day. All right, friends, hello@aboard.com. And we will talk to you all next week.

Rich: Have a lovely week.

Paul: Bye.

[outro music]