Code with Jason
Code with Jason
316 - Adapting to AI in the Agency World with Errol Schmidt
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
In this episode I talk with Errol Schmidt from reinteractive about the evolving role of development agencies amid rapid AI advancements. We explore the impact of AI tools on productivity and the future of software development, as well as the cultural differences in responding to AI's rise.
Links:
Snail Mail Newsletter Pitch
SPEAKER_03Hey, it's Jason, host of the Code with Jason podcast. You're a developer. You like to listen to podcasts. You're listening to one right now. Maybe you like to read blogs and subscribe to email newsletters and stuff like that. Keep in touch. Email newsletters are a really nice way to keep on top of what's going on in the programming world. Except they're actually not. I don't know about you, but the last thing that I want to do after a long day of staring at the screen is sit there and stare at the screen some more. That's why I started a different kind of newsletter. It's a snail mail programming newsletter. That's right. I send an actual envelope in the mail containing a paper newsletter that you can hold in your hands. You can read it on your living room couch, at your kitchen table, in your bed, or in someone else's bed. And when they say, What are you doing in my bed? You can say, I'm reading Jason's newsletter. What does it look like? You might wonder what you might find in this snail mail programming newsletter. You can read about all kinds of programming topics like object-oriented programming, testing, DevOps, AI. Most of it's pretty technology agnostic. You can also read about other non-programming topics like philosophy, evolutionary theory, business, marketing, economics, psychology, music, cooking, history, geology, language, culture, robotics, and farming. The name of the newsletter is Nonsense Monthly. Here's what some of my readers are saying about it. Helmut Kobler from Los Angeles says, thanks much for sending the newsletter. I got it about a week ago and read it on my sofa. It was a totally different experience than reading it on my computer or iPad. It felt more relaxed, more meaningful, something special and out of the ordinary. I'm sure that's what you were going for, so just wanted to let you know that you succeeded. Looking forward to more. Drew Bragg from Philadelphia says, Nonsense Monthly is the only newsletter I deliberately set aside time to read. I read a lot of great newsletters, but there's just something about receiving a piece of mail, physically opening it, and sitting down to read it on paper that is just so awesome. Feels like a lost luxury. Chris Sonnier from Dickinson, Texas says, just finished reading my first nonsense monthly snail mail newsletter and truly enjoyed it. Something about holding a physical piece of paper that just feels good. Thank you for this. Can't wait for the next one. Dear listener, if you would like to get letters in the mail from yours truly every month, you can go sign up at nonsense monthly dot com. That's nonsensemonthly.com. I'll say it one more time. NonsenseMonthly dot com. And now without further ado, here is today's episode. Hey, today I am here again with Errol Schmidt. Errol, welcome.
SPEAKER_00Yeah, thanks a lot, Jason. Great to be here.
SPEAKER_03Great to have you once again. Tell us a little bit about yourself. And uh I'm familiar, of course, with ReInteractive, but for anybody who's not familiar, tell us about Reinteractive as well.
SPEAKER_00Yeah, for sure. So ReInteractive is a uh development agency. So we produce software, primarily Ruby on Rails. I say primarily, almost exclusively Ruby on Rails. Um and we operate around the world. Um, so large enterprise, right down to startups and and smaller businesses. Um it's an interesting space to be in, Jason, at the moment, because uh, as you are well aware, the whole industry is shifting and the subject of what we call a developer is changing, or the definition of what we call a developer is changing. So, along with that, the business model has to change with it. And um, like many other development agencies, we're grappling with that at the moment and figuring out what's the path forward.
SPEAKER_03Yeah, tell me about that. I I haven't had my finger on the pulse too much for a long time regarding agencies. I've worked with agencies a lot in the past, but I'm really curious how they're handling now this big shift that's happening.
SPEAKER_00Yeah, look, it's it's been, you know, it's a slow burn, I guess, because uh over the past probably 12 months, the uh the quality and the output from the tools that we use, Claude, Codex, Copilot, uh, have been improving, you know, markedly over that year. And we started off about a year ago with a I'd say a 20 to 30 percent improvement in output and quality. Um, and that was reflected in the estimates that we would provide our clients. You know, they were, you know, instead of four to five weeks, they were three to four weeks sort of thing, you know. Um but I think in the last couple of months that has improved significantly more. And I think uh, as you noted, Jason, in a tweet today, I think, that will improve quite a lot more um in the next short period. So for a development agency, that means we can't you know make our living out of what we used to deliver in exactly the same way. Our developers are still gonna output the same quality, but they're gonna do it so much faster. Um Right.
SPEAKER_03And the perverse thing is that you get penalized for efficiency.
SPEAKER_00Yeah. Yeah, that's right. Yeah, because we get we're gonna get paid less. So one of I think the advantages is a very good developer who understands the code, knows how to check the output and read the code should be able to do instead of one stream of work, possibly two or even three streams. This is untested. I'm just literally, you know, this is an idea. So a developer might be able to do two or three streams. Now we won't be able to charge as much for that. You know, maybe we charge only a third as much because their attention is on your application a third as much. Or maybe they've got agents running different branches at the same time. Perhaps they do it that way. Um, so we're gonna charge less but do more. That might be a way to do it.
Vibe Coding And Professional Guardrails
SPEAKER_03Um as the productivity is going up, you know, pep people know about AI, even if they're not developers, they know all this stuff is happening, and so their expectations are gonna change. That's right.
SPEAKER_00Yeah. Uh one of the things I've noted, which I think is extremely relevant to your listeners, is uh the term vibe coding. I hate the term. I think it's it's got a pejorative to it. It's it's a negative because it it it really is someone who doesn't know what they're doing telling a coding agent, I want you to do this. And they might get something out of that, but it'll be insecure, it won't have tests, it won't, you know, it won't run the edge cases, it will um, you know, it doesn't have the guardrails that a professional developer would build. A professional developer using those same tools is going to produce an enterprise quality application, and that's what we want. So we have to coin a new term. It's not vibe coding, it's something else. I don't know what that is yet.
SPEAKER_03Yeah, and the the okay, so something I talk about a lot is the difference between things that happen to be true at this moment in time because of where the technology is right now, and things that'll just always be true. Um and something that I look forward to slash fear is the fact that AI tools seem to be heading in the direction of doing work that is in many ways just as good as the best developers. Um so, like thinking about those edge cases and writing tests and and all that stuff, and that's a much different picture. Even between now, I can tell a difference between now and just months ago. Um months ago, I would give a complex task and it would say okay, I'm done, and it would just be like junk. Wouldn't even not only would the code be crappy, but it wouldn't even really like get the thing done. But now it's it's very different. And it's like, wow, okay. This this is a different picture. And the question a lot of people are asking themselves, I think, is where does this leave me when the computer is just as good as I am? Where does this leave me?
SPEAKER_00Yeah. Well, in the it does take a good developer still, so it's still somebody who knows what they're doing. I'll give you an example. Um, I haven't done professional development for many years. So I have been for many years a professional developer and made my living doing that. And then at some point I sort of transitioned into doing more sales and management roles, um, as many developers do, you know, like you, you know, expand your horizons. And so for the last probably five years, that's almost exclusively all I've done. I've done no development at all, like hands-on, actual development. So in the last few months, I have started doing that. I've started finding things that I want to do, and the things that I've tackled, so two things. One is electronics, because I wanted something that had a hands-on approach. So you can see like the screen behind me here is run through an uh a microcontroller that needs to be programmed. And you know, so there's like a blend of hands-on electronics and programming. The other thing I've done is taking one SASTful that I use at a time and rewriting it and canceling the subscription. And so far I've done that with two applications successfully. Wow. Okay. And I think that that there is there is something of interest. So the first one I chose was um I didn't want to pay for my accounting software. I've got all these old clients that have still got domains and and you know things that I have to keep billing for. I hate uh, you know, I hate doing it. If any of them are listening to me, I'm sorry. But I I hate billing them every 12 months for the domain name that they signed up for 20 years ago or something. So so I cancelled like the subscription to the software that I was paying for wasn't even matched by the income. Like it was ridiculous. You know,$60 a month for you know, and I'm invoicing$30 a month. It was ridiculous. So I rewrote it using exactly the features that I wanted with the exact amount of security that I wanted, wrote it in Rails, so I knew what I was looking for, you know, knew exactly what it was that was running, launched it, and I haven't turned back. Like it's running, it runs the you know, runs the jobs overnight, it you know, it's doing everything that it needs to. And the clients are not the wiser, of course. They just receive their invoices and they pay them and everybody's happy. So that that that's a good example of you know of the future. Now, can just anybody do that? Like can a um can um a person who knows nothing about development write, rewrite an entire accounting software themselves? I would say no.
SPEAKER_03I agree. Yeah, yeah, yeah, that's a good point. Um and and something that is true of programs built with AI is like, okay, let's say you want to build a car. Currently, only humans are capable of designing and building a car. You know, the humans build the factories, even though the factories are at least partly automated. A person has to do all this stuff. But then let's imagine a world where you can do the equivalent of how you can vibe code now and just say, like, hey computer, hey robots, make me a car. You could say that. You could say like, make me a car and just just don't bother me and make the car and tell me when the car is ready for me to get in, get in it and drive. But what car is that gonna be? You know, you're not specifying is this gonna be a sports car, a pickup truck, minivan? Like what is it? Is it gonna have a leather interior or fabric, like all these things? Um and so people it it AI like really democratizes coding in the sense that it really lowers the bar and gives people access to be able to do this thing, but it still matters uh what you want to build. You still have to specify what you want to build, and one person's gonna say something different, and it takes skill and taste and all those things, and I think those are things that are never gonna change, no matter how good AI gets.
SPEAKER_00Right. Absolutely right. Like, you know, you can say to the machine, develop the UI for this application, and it will develop the UI, but then it still takes a human to look at that and say, that's not aesthetically pleasing. I don't like the look of that. Or I don't, you know, these buttons are not going to work there. They need to go to the top, not the bottom, or something like that. That yes, you're absolutely right, or or you know, change the entire style of it because this isn't working for me.
SPEAKER_03Something uh related that I've seen, and I wonder if you've dealt with this, is um it it's an it's another economic question. Um, you know, productivity is going up, uh, there's maybe I'm just guessing, but maybe there's downward pressure on agency rates and stuff like that. But another kind of adjacent phenomenon I've noticed is is the job market and how all that works. I guess it's kind of analogous to agencies too, because it's like uh employers hire programmers on a permanent basis, but also people hire agencies. Um and that whole marketplace is just so different. Like there was recently a job I saw filled, and the number of applicants before it got filled was 2,000. Um Yeah. Wow. And you know, I've heard Was that a developer role or something? Yeah, it it was. And I've heard that now it's normal to have like you know several hundred applicants, a thousand applicants. Uh this is the first one that I've heard that's two thousand. That's the biggest number I've heard so far. But basically employer employers are getting DDoSed with applications. And I think that's both because the economy is is tough, the market's not what it used to be, and there's AI spam. People are are submitting all these AI submissions. Anyway, any direction you want to take that? I'm just curious if you've had any experience of the markets changing.
SPEAKER_00I think, Jason, yeah, what what you're saying is correct. I mean, 2000 is outrageous, of course, but we are living in the tumultuous times. We're living in the period where we're not entirely sure how this shakes out. Uh we're we're figuring it out as we go, like literally moment by moment here. We're trying to figure it out. And there will be a point in the future, and I'm I have some certainty on this, that uh well, a couple of things actually. One is that for sure the number of applications and the amount of software that the planet requires dramatically increases, like by several times over, simply because we're becoming more and more tech uh more and more of a technology society every day. So that increases. So therefore, the value of a good developer who can read a piece of code, understand it, QA it is worth more. So because there's more software out there, a developer, even if he's running several things in parallel because he has his agents running them, their agents running them, uh I I think there's still plenty of work there, is my thought. That's my thought. Now, you know, I could be completely wrong about that. I just don't know. But there is definitely a human in the loop still. There there's, you know, there's no doubt about that.
SPEAKER_03Yeah. Yeah, I share that view, and and I have for a long time. And then I got surprised. I don't know about you, but I got surprised when these shifts happened in the job market, where it changed from a seller's market to a buyer's market. You know, there's there was a time that that I remember where if you were a developer, you held all the cards. Um and and you could, if you didn't like your job, you could very easily go get a different job. Um, and now that is so not the case. That that's the change that surprised me. I it never I never would have guessed that you'd have people with 10, 20 years of experience just uh unable, like more than a year without a job. Like, this isn't just one or two cases I've seen. This is like I've seen a lot of cases of really experienced people basically just not in the game anymore because the market's so different. It's so crazy. I never saw it coming.
Past Tech Disruptions And New Roles
SPEAKER_00Yeah. Yeah. Yeah. I I I I read a lot of books on this subject, um, and I can't even remember particularly where I picked up this concept, which book it was, to be honest. I think it might have been the coming wave, but I can't remember. Um and it's it it's that point where people were riding horses and everything was done by horses. Work was done by horses, travel was done using horses and carriages, and then almost overnight engines and cars and tractors and so on became a thing. And it happened almost immediately. And there were no horses, or very few horses. So, what happened to all of the people that would look after the horses and could train the horses and run the horses and stabled the horses? Like what happened to all of those people? They all had to find something else to do because their profession didn't exist anymore almost overnight. Um and there's been periods like that in history before. Um the invention of the loom was another one, you know, where individuals had to weave clothing and and rugs and so on, and then almost overnight this machine took over their job and could do it, you know, much greater output for less human interaction. So that this thing has happened many times in our history. Um somehow new jobs were invented along the way. For example, there was a time when social media didn't exist, and so no one would have expected that there was a job of social media manager. Like, why does a business need a social media manager? Well, all large businesses do, maybe many of them. Um and there was a time when this didn't exist and now it exists. So what will those future jobs be? I I don't know exactly. You know, like the role of a project manager probably still exists in the future, the role of a CTO almost certainly exists in the future. What else that can be invented? I I don't know. You know, I remember when when when I think it was GPT 3 maybe, and people were saying that there was going to be a job called prompt engineer, and there just wasn't. There's never gonna be a thing. But yeah, we we don't know, but there's definitely there's roles in the future that we don't know about yet.
SPEAKER_03Yeah, I just want to dwell on that for a second. There's this whole thing, prompt engineering, uh, where you like basically get really good at prompting. And it's like that can't you don't really need that anymore because you just say what you want and the technology is better now. You don't need to have that skill.
SPEAKER_00Yeah, what what I enjoy is I can I can send it off with all my typos in it. Right.
Prompting For Clarity And Agent Feedback
SPEAKER_03Yeah, and it still understands. Yeah, yeah, and and what I like to do, this is a little tactic that I've adopted, is I I tell it uh uh to just anything I I I give it, um scrutinize it and like repeat it back to me in your own words. Uh and if there are any gaps or anything, uh again, scrutinize it and ask me questions and stuff like that. So I'll give a quick example. Um I'm working on one of my uh genetic programming projects that I've been working on for many years now. Um and I have these little uh I'll I'll say I have these little critters floating around on the screen and in in this particular version of it. Uh and the critters have basically eyes. They're they're these hexagons, um, and so each critter has six eyes, one on each of the six sides of the hexagon. And in order to um be able to understand what's happening in this program and in this world, I have this. Monitor where I can pick a certain critter and see what it sees. So there's on the left side, imagine a like moving conveyor belt where it's like a table with it's a table with six columns, one column for each eyeball. Um and and uh each eye is currently seeing just the the what whatever's directly in front of it. Right. Um it'll just see exactly very low resolution, all it sees is a single color. So it might be seeing blue or black or red or whatever. And so imagine a row of little rectangles where it's like a blue one, a blue one, a red one, a blue one, six of them, and then they move down over time, and so you see this pattern where it's just you're seeing what the critter has seen as it has moved around in this little world. Way more background than I intended to give. But here's the feature that I added. Um I said give each critter a a two-character ID and label them all so that I can keep track of them. And I want to be able to somehow select a specific critter and have the monitor change so I can see what that critter sees.
SPEAKER_00Okay.
SPEAKER_03And so the AI asked me, this is an example where it took my kind of ambiguous uh input and and figured it out and didn't just take it verbatim and try to build it. It said, like, well, how do you wanna how do you want to communicate that? Like, do you want to click on a critter? Do you want to like use some kind of keyboard input somehow? Like, I need to know how this is gonna happen. So I thought that was really interesting because like in the past, like, I don't know that and without that instruction to like scrutinize and ask for clarification and stuff like that, I'm I'm not sure it would have done that. And this is very reminiscent to me of what a very experienced, conscientious, good developer would do. You know, don't just take the instructions and say, yep, got it, and then go build something. Be like, hang on, let me make sure I actually understand. Did you mean this or this? How do we handle this? Blah, blah, blah. It's it's really incredible.
SPEAKER_00Yeah, that that reminds me very much of the um of um the recent article from DHH, or actually it was a Twitter post, I'm sorry, uh, where he mentioned that he'd said his um uh open claw agent uh was looking into API documentations at hey.com. And um it reviewed something and then unprompted came back to him and said, the paths are confusing. The paths are too complex. That's not good for an agent and it's not good for a human. You should rewrite it very simply with just an ID. And he looked at it and went, Yes, of course. And let it go off and do that. So it was an unprompted recommendation. That was a very smart, accurate recommendation, which I think is a very like it's a surprising result, actually.
SPEAKER_03Wow. That wow, that is really interesting. That that's something DHH put online. I'm gonna go look that up later.
SPEAKER_00Yeah, I'll uh I'll um I'll I'll I'll find the the link and send it off to you.
SPEAKER_03It was very show notes too.
SPEAKER_00Yeah.
SPEAKER_03Yeah, that is really interesting because because one thing that AI lacks is the blessing of stupidity. Um, I always say that I don't write good code because I'm smart, I write good code because I'm stupid. Um because it it has to um cater to my feeble brain power. I don't want to have to exert a lot of mental energy to figure out what the code means. But the the machine doesn't have that limitation. But somehow uh open claw has uh you know it looked at this thing and it's like this is this is confusing. It was able to figure that out.
SPEAKER_00It's still gone. Well that's too many tokens.
unknownInteresting.
SPEAKER_03Or something. Yeah. So I I haven't used this open claw thing. I've only heard about it. I haven't tried to look into it at all. Have you messed with this thing at all or no?
SPEAKER_00I uh a little bit. I installed it and started to, and then I then I got carried away. I had another project ongoing at the time, and so I started doing that. I've only got so much time to do the after hours things with running around with kids and stuff. So I've I've installed it, uh, but I've run nothing through it basically.
SPEAKER_03Yeah, yeah. And it frankly, it doesn't seem like the kind of thing that I would be into anyway. Um but I I I'm always very careful not to just dismiss these things out of hand.
SPEAKER_00It it is it is remarkable, like you know, looking into some of what people are doing, it's remarkable, and it's a breakthrough of of sorts. And really that breakthrough is what would happen if I didn't provide as much if I didn't lobotomize the agent, I think is probably the best term way to put it.
SPEAKER_03Yeah, you know what this whole thing is? It's a uh speciation event. So if if anybody's not familiar with that term, it's like when that asteroid that killed the dinosaur slammed into the earth, uh it rapidly changed the climate of the planet. And of course, a lot of organisms died instantly, but then over the ensuing um, I don't know, thousands of of years, whatever, um basically in geological time, it just caused like 90% of the species on the planet to go extinct or some something like that. Um, and then that void uh left by all those species going extinct uh left space for new species to proliferate. And you know, this this tiny little mouse-like mammal, which was like I think the only mammal in existence at that time of that asteroid strike that eventually became us and all the other mammals. Um and so I'm really curious in this speciation event, what are the spe kind of kind of like you said earlier with the horses and all that and the and the loom, um, what species are going to die and what strange new species are are going to appear?
SPEAKER_00Yeah, right. I when you said that I I was thinking of a book that I'm reading at the moment. Well, actually an audio book. Uh it's called, I can't remember who wrote it, I'm sorry, but it's called If Anyone Invents It, We All Die.
SPEAKER_03Oh yeah. Eliza Yadlowski or however you say his name.
SPEAKER_00That's that's ominous.
SPEAKER_03Yeah, not to be mean, but I think that guy's a joke.
SPEAKER_00Well, I it's it is so extreme, but I'll just win.
SPEAKER_03Yeah, he's he's not an intellectually serious person. Um yeah, the thing is uh AI would need a reason like in order to kill everybody and stuff like that. And people have this uh habit of anthropomorphizing AI and uh projecting our human emotions and drives and stuff onto the machine. Yeah, yeah, exactly. And it's such a common thing because you know, humans are our intelligence is unique. And so if something else has this intelligence, even talking with like uh Claude Code and stuff like that, it feels like you're talking to a human. But it it's just playing a trick on us. It doesn't have emotions, it never will have emotions. In order to have it, we would have to program in the emotions, and we'd have to forget to say, oh, by the way, don't kill us all. Yeah. Please.
SPEAKER_02Yeah. Anyway, um Yeah. Yeah, so so what what are you go ahead?
SPEAKER_00Yeah, it just it doesn't replace the human element. Like, you know, the the new ideas are still gonna come from us, and and the the bright ideas and the original ideas are still gonna come from us, and that element of QA, as you said before, the aesthetic approach, the this looks good or this doesn't, that still comes from us. And so, you know, uh going back to the earlier discussion, you know, what what is the place of an agency? Well, it's that. An agency is still how does how do you take, you know, okay, so this person has a bright idea, but a bright idea is not enough of a prompt. You still have to write the tickets, you still have to break that down, you still have to work out how does this thing get done, feed that into the machine, get something back, QA, QA that, and then you have a working application. So sure we can do that in maybe 10% of the time that it used to take, but there's still that element that is irreplaceable. And in the hands of someone without those skills, you would not get the result.
SPEAKER_03Definitely. And I imagine there's at least some amount of guidance that you're giving clients. Like sometimes people come to me and they're like, hey, I'm thinking of building this product for the certain market. Like, how do I approach? Usually it's like a low-level technical question. It's like, how do I build this thing? And and I'm like, well, what's your overall plan? And stuff, and asking them questions and just guiding them in all these ways that aren't technical. And I'm sure that there's some amount of that too, where you just want to be talking to a person, not a machine.
When AI Adds No Value
SPEAKER_00It uh interesting example. We have a particular client who um who came to us and said, you know, jumping on the hype train, said I want to turn this particular feature, I want to, I want to develop this feature and make it AI powered. You know, the the client's words, not mine. Make it AI powered. And and then there was a discussion that ensued, obviously, what do you want to do? What do you want to get done? What is the objective of this particular feature? So the actual feature was properly scoped. It was just a it was basically just a form that did some basic checks through an API, did some basic checks to make sure the person had you know appropriate was a real person, a real business, credit, all the rest of it, and then allowed them to shop on the B2B website, basically. That's the short of it. And they wanted that AI powered. Truth of it is, there was no need to involve AI in that. Like there was no advantage. In fact, an AI tool would have just cost more and slowed the process down. It wouldn't have been a smart tool to use. Just calling out to the right APIs was the right choice. We built that in, you know, less than a week, I think it was maybe a week. Built that and it increased the conversion of new clients by 30%. That was worth a massive amount to that business. But we were able to achieve that in a short time by just understanding what that client wanted, and it was an AI.
SPEAKER_03Yeah, interesting. Um, that gives me a brilliant idea for a new product. Um and I'm not trying to poke fun at your client by this or anything like that. Um, but the product could be called the AI fairy. And what it does is it comes and it sprinkles some AI onto whatever you want. So if you have a program that doesn't really call for AI, but you want to you want to be able to truthfully say that there's AI in it, uh huh, you just invoke the AI fairy and it puts some AI in it somehow in some inconsequential way, and then you can say, Yeah, it has AI.
SPEAKER_00Yeah, and you can put a badge on it, AI powered.
SPEAKER_03Exactly. Yeah, it's so interesting. Um there's when AI was a little bit newer, and this is surely still happening, so many products, maybe you noticed this too, they would just have like some there's this popular like magic wand icon thing where it's like here, do this thing, but do it with AI. And you can get like some summary of something you don't really need a summary of, or or something like that. They just did it just to say they could just so they could say they're doing something with AI.
SPEAKER_00Uh-huh. Yeah. Yeah, just expense some tokens for no no benefit.
Staying Ready Through Relationships And Learning
SPEAKER_03Yeah, so okay. Um, I don't know about you, but I um we're we're in this weird new world. Um and I am not without concerns, you know. Um we are barreling toward this future that nobody knows anything about. It's like we're driving a hundred miles an hour in the dense fog, you know? And what's what's out there ahead of us in the fog? What are we, you know, is is it just straight road for the next five miles, or are we gonna drive off a cliff? What what's gonna happen? Nobody knows, but we're we're barreling toward it no matter what. And so my question to myself is like, how can I position myself? What should I do so that I I I don't end up uh so so I don't get caught flat footed at some point in the future. And to put it a different way, if some big change happens, what might I regret not having done over the last several years leading up? Uh and and and I try to think about that. And and I'll give one one answer to my own question, which I know will resonate with you just because we know each other and I I know the kind of person you are. Um one thing I will never regret, and I'll always be happy that I did, is build relationships.
SPEAKER_00Yeah.
SPEAKER_03Because no matter what personal contact. Exactly. No matter what technological change happens, I think that's a very safe investment.
SPEAKER_00Anyway, I was just curious if you think about No, I've I've thought about this exact same question, Jason. It's it's a question I've asked myself. And the reason is because probably about a year ago, on reflecting on all of this, I felt a little bit at sea. You know, I felt like you know, that car barreling down the road in the fog, knowing there's probably a brick wall there somewhere. I felt like that. And I felt like there was nowhere to turn. Like, yeah, and and it was just me being uh an effect of all of these things happening in the environment and not being causative about it myself. So I had to switch my mindset. So I did a couple of things. One is exactly as you said, I spoke to as many people as I possibly could. And that gave me the confidence that uh it wasn't just me. Everybody was trying to figure this out. Um and you know, the people I spoke to, people like yourself, community leaders, senior engineers, CTOs, founders, these are the people I spoke to because I wanted to know what they were doing about this and how they felt about it. The other thing I did is I got myself as educated as possible, picked up every book on the subject I could find, and I'm still doing that. Read or listen to as much as I possibly could. Podcasts, articles, Twitter comments, books, audiobooks. By getting yourself more educated, you're grounding yourself in reality. As you said, there is uh there is a a tendency for people to misunderstand what their machines are doing and start to think there there's a human in there or there's a soul in there and there isn't. There is a machine processing ones and zeros. That doesn't lessen any of the magic that it's doing by any stretch of the imagination, but it's also not a real person in there. You know?
SPEAKER_03I want to emphasize something you said. Uh you said something like uh the more educated you are, the more grounded in reality you are, or something like that. I just want to emphasize that because that mirrors the way I think a lot, and I think that is just so important. The more accurate of a mental model of reality you have, uh the more successful you can be in interacting with reality.
Artists Versus Developers On AI
SPEAKER_00Here's a question I've thought about recently, uh uh which is directly along this line. Two completely different careers. One is artists, like um people who paint and draw and make movies and things like that, those sorts of artists, and software developers both have um both have the potential of career-changing things happening because of AI. The artist community has almost universally, not completely, but almost universally rejected or kept on the fringes of anything to do with AI art as not art, as not, you know, like they're rejecting that completely. Developers, on the other hand, are embracing the changes. They're going, bring it on, let's go, let's see what happens. Like it's a very different response to AI. And I think the difference is you've got one subject which is uh it's not self-taught because there's a lot of technical things in the field of art, but the actual output is not. The actual output is you know a lot of creativity. Developers, on the other hand, study almost continuously because there's always been changes in the industry. There's always been, you know, there's a new feature to learn, there's a new upgrade to this, you know, you know, Rails is being upgraded, there's new features, or there's security concerns to take into account, or there's a new application to learn that will do a particular thing. And you know, like there's always been changes in this industry. And so developers are more grounded in continually learning. Maybe that's right, maybe I'm completely wrong, and you'll get a bunch of comments saying that guy's an idiot, that's also possible. I could be completely wrong.
SPEAKER_03That is a fascinating question. Um, and I confess that I I didn't digest what you said as as thoroughly as I should have, because it it just set my mind in motion. Like there this difference between how artists have reacted reacted and how developers have. Um and I think I'm I'm gonna make a statement that I'm not sure I actually believe, but like art is valuable precisely because it doesn't need to exist. Um I I I think art is a statement. Like when you make a piece of art, whether it's a song or a painting or whatever, you're making a statement. It's it's a form of self-expression, and and that statement that you make might resonate with other people. You might make a statement and somebody decides something to them. Right. Somebody receives that statement and they say, That resonates with me. Like today I was listening to some song and I was thinking about the people who made that statement. And thinking about the fact that there's we are of the same mind. Whatever whatever pattern was swirling around in their mind at that time, that pattern is now in my mind, and we are like sharing this uh sharing this thought. The the same pattern is existing in both of our minds at once. Right. Um, and that is something that is like to automate that is not even it's not that it's good or bad, it's just that it's not even a meaningful idea.
SPEAKER_00I I don't think it's entirely possible. I don't think it's actually possible without a real human there doing that.
SPEAKER_03Yeah, because it's like if I were listening to an AI-generated song, maybe an AI could make a good song. I I haven't heard one yet, but I wouldn't rule out that it's possible. Um In fact, I'm sure it's possible. But a huge part of enjoying music is again resonating with the statement that someone made, or looking at a painting, you can be like, wow, that thing that that person was feeling. Like, I get it. I this resonates with me too. And then there's an interesting spectrum from like purely artistic to purely commercial, and then there's everything in between. Like, what about a billboard? Like people don't make billboards to make an artistic statement. Maybe more than 0%, but it's certainly not 100% artistic statement, it's maybe 3% artistic statement. So right, there's a scale of that, like there's a you know, yeah, yeah, and if there's like an AI-generated billboard, like I don't care about that. Like, fine. Right. Not to be cold, but like the artists who make billboards, like I don't really care that much if they can't do that anymore. There's a lot of programmers who can't do that anymore, so they'll have to, just like everybody who's been bulldozed by technology, they'll have to go do something else. Um But yeah, I I think that that first kind of art that will always exist.
SPEAKER_00Yes, it will always exist and absolutely should, and there's no question about that. Um, to what degree those artists embrace or or is there a new genre of art? I was looking back on some images that I created with DALI, if you remember that, back in 23, maybe, I think was the year, 20, yeah, something like that. When DALI first came on the scene. And that they were a very interesting, like they were strange, they were odd, but they had Had a certain style to them. You can't mistake what they are.
SPEAKER_03Right. Yeah, it's it's the differences between reality and the image that make it interesting. Like I I don't know about you, but I hate, hate, hate, hate drawings and paintings that are made from photographs. Because it's like then you're not seeing an artist's eye view of the world. You're just seeing a copy of a photo.
SPEAKER_00Right, photo-realistic painting. It's what what's the value of that?
SPEAKER_03Yeah. Yeah, exactly. It's like that's just the photo, just but somebody making an imperfect uh representation.
SPEAKER_00Yeah.
SPEAKER_03That is interesting. And that's why I find like uh Dowie generated images in a way a lot more interesting than these more recent ones that are like better.
SPEAKER_00Yeah. Yeah. They they have they are flaws and all. And you can't mistake what they are, but I think it's a I think it's an art style unto itself that existed for actually only a very brief period of time. Probably that is. But maybe somewhere in the future people will look at those and go, there is something about that.
SPEAKER_03Right. It's like there has to be a distortion between the reality and the artifact. It has to go through some distorting process.
unknownYeah.
SPEAKER_03I think.
SPEAKER_00Yeah. That's yeah, yeah, that's right. I mean, I'm yeah, I think we need an artist to weigh in on this.
SPEAKER_03Yeah, that would be that would be great. Um yeah, yeah, so it makes total sense to me why artists would reject AI so fervently.
SPEAKER_00Yeah, and and and developers are more uh well, they're pragmatic, right? Most of the people.
SPEAKER_03Well it benefits us. Yeah. It's doing the it's having, well, in a way. It's it's having the exact opposite effect in a way. Instead of destroying our well, I don't know. I disagree with myself as I say it out loud. I was gonna say instead of destroying our craft, it's boosting us. But I think a lot of people feel the exact opposite way.
SPEAKER_00Right. Time is gonna tell on that. It it's interesting. I I I'm with you. Like I I'm I'm seeing different sides of this argument, and I'm not sure where I land.
SPEAKER_03Yeah, it's like imagine there are this this just happens to be a good analogy, I think. Imagine there are artists who paint portraits for a living, and now you don't have to do that, you can um you can take a photograph of the person and you get paid a little less, but you can still do basically the same job, and it's so much easier. Instead of having to make a painting, which you're not making for artistic purpose, you're just trying to capture the the person's uh uh likeness or whatever. Now you can take a photograph. A lot of people would be unhappy about that. They've just my craft got ripped out of my hands, now I have to do this like soulless thing, but then a lot of other people would be thrilled about that. So like, wow, instead of taking weeks to do one single portrait, I can do like 10 portraits in a day or whatever. And it's the it's the same phenomenon affecting people in the same way, but different people feeling vastly differently about it.
SPEAKER_00Yeah, absolutely, absolutely. It reminds me of a uh someone I spoke to once. My father was an artist, um, and he he he was very good, but he knew a lot of very, very good artists, like you know, top-tier artists. And um he was trying to sell, he had a business at the time, he was trying to sell part of it, and I I went to visit this particular person who was interested in buying this part of the business, it was picture framing. And she was a very, very good artist, very well known, and produced the sorts of works that you would see in large corporate foyers and things like that, massive, massive canvases. And you know, I was talking to her and she said, Oh, you know, do you want to see some of my work? And I was like, Yeah, absolutely. So she took me upstairs to where she had her workspace and the the place just had uh hundreds and hundreds of canvases, these massive things. And she was showing me them. They were in different styles, like some were like you know, a mixture of different things, um, which sort of broke my idea that an artist has a style. She didn't. She had you know a a range of different things. And she was showing me and what I was looking at one particular piece of work that was like very, very large. And I said to her, How long did this take you thinking she would tell me months? She said, Oh, it's about a day. I was like, that took you a day. She was selling this piece of art for something in the range of thirty to fifty thousand dollars. Wow. So and it took her a day to paint. And she went, she started to justify it to me. She started to go, you know, she started to explain why why she could charge that. And I said to her, You don't have to. You've been an artist for 40, 50 years, probably. You have honed your art for 50 years. If you can produce something that someone's willing to pay you$50,000 for in a day, it's because you've spent 50 that 50 years creating that. You don't have to justify it. So I I didn't I'm not sure why I told you that story, but it was an interesting story.
Value Creation Trust And Pricing
SPEAKER_03No, no. I I I think that's instructive. Um, I asked somebody recently why do we get paid? Yeah. Why why do we get a paycheck at our job? And I'm curious, let me let me ask you, you know, why why do you give a paycheck to your developers?
SPEAKER_00Yeah, well, uh, because they produce something of value. They produce something of value and there's an exchange in it. So they give it to a client, the client pays something back, that's the exchange.
SPEAKER_03Exactly. The this person I asked, he he said something uh regarding time and effort and stuff like that, and and expertise and that. Expertise. And it's like those are all factors, but that's not why we get paid. We we get paid just like you said, for the value that we create. And it's interesting how many people uh have uh a mistaken idea of of the real reason why we get paid. You know, so somebody doesn't care. It's interesting that that artist wanted to justify uh the price of that painting. Because it's like I don't care if it took a day or a month or five minutes if the painting speaks to me and if I would rather have the painting than my fifty thousand dollars, then I I come out on top. I'm I'm better off, yeah.
SPEAKER_00Yep. Yeah, absolutely, absolutely, and art, uh perhaps unlike software, goes up in value.
SPEAKER_03Oh, art, yeah, yeah, yeah. Right. Yeah, art is or I mean uh software is much more ephemeral. It rides.
SPEAKER_00Ephemeral, it is. And and maybe more so now Yeah.
SPEAKER_03Yeah, I think that the more technology advances, uh the the the more the rate of advancement increases, the shorter the shelf life of any particular thing.
SPEAKER_00Yeah, so it's a a question then of what what is the true value of that piece of software? There's a few things. One would be in the goods and services that it promotes. So um uh if you think of say Spotify, the value is not the application itself, it's the it's the collection of music and podcasts. So the application is a gateway to those things. It has to work, it has to perform that duty flawlessly. But the value is the the goods and services behind that. Um or it could be access to something that no one else has access to or provides access to something that's difficult, like maybe um uh I can't think of a particular example. I don't know, maybe if there's a CRM that has a difficult API and you know how to access that in an inexpensive way or in a way that produces something that no one else can do. Those sorts of things provide value. Um, these are some of the things I think. Or if you can provide something in the same way.
SPEAKER_03By the way I can I cannot I cannot imagine why you possibly use that example. Uh inside joke, dear listener. Um okay, yeah, yeah, yeah. Um yeah, Spotify is interesting because it's kind of the the value is in it's a middleman. It's there's there's this pool of value, there's all the world's music, and you can get it. You can go to a record store and buy some music. Yeah, yeah. But Spotify did this middleman work of uh, you know, I don't know, making these deals with the music companies and stuff like that, and conveying the music from where it originated to your phone and doing it in a way that, like you said, is reliable and doesn't make you want to blow your brains out when you use it. You know, it's it's like not a perfect UI and stuff like that, but surely there were other competitors that couldn't pull it off.
SPEAKER_00Torrent Torrents was the was the competitor. Oh, yeah, yeah, yeah. Yeah, like that was that was the competition, and that showed that there was a market to download and listen to music easily. That was the competition. As shoddy as that was, but that was the competition. And that's the market they replaced by doing it better, even at a cost, they did it better.
SPEAKER_03Right, because the way I look at it, uh paying twenty bucks a month or whatever for Spotify, that is cheaper than spending all these hours searching uh the Pirate Bay or whatever and illegally downloading music like I did for many years.
SPEAKER_00Right, right. I guess that then there's there's another thing that is that provides value, and that would be trust. I was thinking about um about Shopify, actually, just you know, while we were talking. What does Shopify provide that no one else has done before? What nothing really, like nothing I couldn't do with WordPress. Not that I'd want to, but you know, you could set up a commerce store on WordPress and have it running and sell things, and certainly many people did. Why do people use Shopify? Because they trust it, they know it's going to work, they know it's not going to steal their money, they know that it's not gonna fall over and not be there tomorrow morning. There's trust, confidence.
SPEAKER_03Yeah, that's a very good point. That's something I've thought about um a lot as I'm working to get Saturn CI off the ground. I don't yet have that. It's it's similar to the network effect thing, you know. The more people you have in a uh social network or something, for example, the more valuable it is. Um there is this flywheel where it's gonna be more attractive to people the more people are already using it. Nothing attracts a crowd like a crowd kind of thing. But the first few people who use it are taking a risk because they're and I myself am very hesitant to use products where I'm like the first guy to use it, because like I don't know if this is gonna be around in a year or whatever. Um, so that's very hard to get that flywheel started and get those first intrepid souls who will use your product and and live through the the roughness of the early version of it, so you can get through that period uh and then get your later adopters and stuff like that.
SPEAKER_00Yeah, because it could be the best product on the planet, and you know, if we're talking about seven, so it probably is. But is that enough to m make people use it? Like it's not yet. Like they have to go, oh, this is I trust who developed it. I trust that it's gonna be there in a year's time. You know, that that's an important consideration.
SPEAKER_03Right. And I think what you just said is is a big part. I trust who developed it. Um, I think you will probably appreciate this as somebody who works in sales. Uh the I think the early people who use it, I think about a hundred percent of them are gonna be people I know personally.
SPEAKER_00Absolutely. Uh, but that and that's who we use, right? We go to our immediate network and we say, I want you to use this.
SPEAKER_03Exactly. Yeah. Yeah. Yeah. Um that's all we're doing. And have you gone through that process uh in in all of your all of your work of trying to bring a new product to market? Or is that something that you might do in the future?
Getting First Customers For New Products
SPEAKER_00No, no, that that is that is something I've definitely gone through the the teething on. So um several years ago within Re Interactive, we developed a product called Store Connect. So Store Connect is an e-commerce product. Um I think you might be spoken talking to Michael. I think he might be scheduled to talk with you soon. Anyway, Michael is the founder of Reinteractive, so built that built that product, RD'd it, and it's uh it's an e-commerce product that sits on top of Salesforce. So it allows Salesforce users to sell using data within their Salesforce org. Um and it it was a very, very good product. Like we spent a lot of years developing it, in particular through those COVID years when you know there was less development work around the developers concentrated on getting this done. So a lot of time was spent on doing it. Um and then once it was built, it was a great product. But then how do you get people to start using it? And I had this experience of trying to sell it to people uh and trying to get those first few customers, which was tough. It was hard work because people didn't know it, they didn't trust it. It was still an incomplete, like there was still, you know, people would say, Oh, does it do you know blah? And I'd go, yes, absolutely, in six months we'll have that built. And then I'd go off and tell the developers, hey, you've got to start building this. They hated me. Um But this is this is how it went. You know, I remember the you know, the first large sale that I did, and you know, half of half of what I sold was not available. I had to build it. It was incredible. But they did, and it, you know, the the the fortunate thing there is we had some people that I knew quite well, and other people internally knew quite well at Salesforce and were willing to say, let's give this a go. Set up a couple of clients initially, you know, got a couple of people onto it, they were happy, and it sprung, you know, further clients, and then eventually it broke off into its own business and is that now today and doing incredibly well and growing. So um, yes, definitely had experience with that.
RubyConf 2026 And Mini Eiffel Tower
SPEAKER_03Yeah, it's it's quite the journey. Um, and like you said, it is so difficult but so rewarding when it hits. Yeah, um one more thing I want to touch on real quick before we go. Um, we spoke about this, but I don't remember for absolute certain your your what you said. Um, are you making it to RubyConf this year?
SPEAKER_00Uh look, I I absolutely want to. Yes, I'm having conversations with people all over the place that are that are reminding me of it. And the gala I'm super interested in. So I I absolutely want to make it over for that.
SPEAKER_03Yeah, well, I'll just uh do a quick pitch for the listeners' sake in case you haven't heard about RubyConf 2026 yet. Um it's taking place July 14th through 16th in Las Vegas. Um we have a keynote speaker who I'm not announcing yet, but he's a very big name in the Ruby community and in the programming world in general. He's the author of a very famous book, which there is a very good chance you have on your bookshelf, dear reader or dear listener.
SPEAKER_00Maybe one that I have on my desk right now.
SPEAKER_03It could be. Um Yeah, and and myself and uh my friend Freedom Dumlau are co-chairing RubyConf this year. Um and we're gonna try to make it weird. We're gonna try to have a weird conference, obviously in a good way. Um I I was just texting with Freedom earlier, and we said, hey, we should encourage people to dress eccentrically and and wear like Vegas themed clothes. So if you want to show up to RubyConf and dress like an Elvis or something like that, wouldn't it be great to show up at you're you're a speaker at RubyConf, you look out into the audience and it's just half elf. Yeah.
SPEAKER_00Um and and here's another thing You can easily walk around Las Vegas looking like that. No one would even blink an eye.
SPEAKER_03Exactly. Um and and and one one thing I want to do, and this is where I could maybe use your help, dear listener, is um okay. Paris has the Eiffel Tower. Las Vegas, if you don't know, has a mini Eiffel Tower. It's one-third the size of the original Eiffel Tower. Well, I would like Rubikonf to have a mini, mini Eiffel Tower. So a smaller version of the Las Vegas smaller version of the Eiffel Tower. Uh Lego makes an Eiffel Tower set. I looked it up at$630. And it's pretty big. It's like, I don't know, four or five feet tall, something crazy like that. Um, and I am now accepting an unofficial Eiffel Tower, mini, mini Eiffel Tower uh sponsor. You don't get anything in return. This is not an official RubyConf Ruby Central thing. This is just me personally. You can be our mini, mini Eiffel Tower sponsor. It costs$630. With that money goes directly to buying a Lego Eiffel Tower, and we'll put it somewhere on display in RubyConf. You know, it'll start off.
SPEAKER_00Do you get the attendees to build it? Like set it up so that people can of course put a block in here and you can put a block in there.
SPEAKER_03Of course. We just you know set the box on a table or something like that, and the attendees go nuts over the course of the conference.
SPEAKER_00Yeah. I I think you just want about a hundred attendees.
Wrap Up And Links
SPEAKER_03I sure hope so. Um so as of this recording, tickets aren't on sale yet, but they will be very soon. And if you want to be our mini, mini Eiffel Tower sponsor, my email address is jason at codewithjason.com, and we can set this up. Um, Errol, I really hope that I can see you at RubyConf. And if not, I'm sure that we'll see each other at some conference soon. Um before we sign off, any links or anything like that that you want to share with people?
SPEAKER_00Oh, look, I think just reinteractive.com is the best.
SPEAKER_03All right. Well, Errol, thanks so much for coming on the show.
SPEAKER_00Awesome. Thanks, Jason. Talk to you soon.
SPEAKER_03Okay, up I just hit the keystroke. That is, okay, one second, stop.