Skip to content
PodSubs
ShowsPeopleGuestsJobsNewsEvents
More
BlogPodCred
Menu
ShowsPeopleGuestsJobsNewsEventsBlogPodCredNetworksSign in
Join nowSign in
  • Shows
  • People
  • Jobs
  • News
  • More
    • Guests
    • Events
    • Blog
    • PodCred
    • Networks
    Join nowSign in

Code with Jason

Code with Jason artwork

333 - Kevin Dias, Author of The Problem-First Method

Nov 9, 202657 minAudio

In this episode, I talk with Kevin Dias about his journey from investment banking to programming in Japan, building Ambiki for speech language pathologists, and tackling feature requests with a problem-first mindset. Links: - Ambiki - Kevin's Personal Site - The Problem-First Method (book) - Kevin on LinkedIn - ProZ.com - Nonsense Monthly

Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.

Transcript

From the show. Tap a time to play from there.

SPEAKER_01Hey, 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.

SPEAKER_01That'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?

SPEAKER_01You 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.

SPEAKER_01The 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.

SPEAKER_01I'm sure that's what you were going for. So just wanted to let you know that you succeeded. Looking forward to more.

SPEAKER_01Something 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 nonsensemonthly.com. That's nonsensemonthly.com. I'll say it one more time. Nonsensemonthly.com. And now without further ado, here is today's episode.

SPEAKER_00Hey Jason, thanks for having me.

SPEAKER_01Thanks for being here. So you are joining us from Japan, is that correct?

SPEAKER_00Yeah, I live in Tochigi, which is about an hour and a half north of Tokyo by train. Uh in the rice paddies, nice quiet area.

SPEAKER_01Okay, and you said you've been there since 2008? Yeah, been here a while now. So what took you to Japan originally?

SPEAKER_00So originally studied abroad for a year in college, really enjoyed it. And then I was in the investment banking world in Charlotte, North Carolina for a couple years. And this was 2006, 2007, but the writing was kind of on the wall of the impending like 2008 financial crises. So and also like I don't know, banking just wasn't a fit for me. I guess. Well, I I was learning a lot, but I didn't feel any particular passion or connection to it.

SPEAKER_00Uh so did a 180, went to teach English in Japan, and it was kind of at that first job where I actually started teaching myself to program because so they had Lotus Notes at the office and it was all in Japanese. But they had these students who had you know been at the school for 10 plus years, and each year they were cycling into this new uh you know foreign English teacher coming in, and it was like they had to like start from zero with these are my goals, these are the textbooks I've studied.

SPEAKER_00So I wanted to create you know some sort of student progress tracking database that you know that the next teacher after me would have something to go off of, you know, for these you know students who had been there for a while. Uh so that was kind of my first intro to like teaching myself uh you know how to build stuff. And actually at that time I used Michael Hartle's. I don't know if you remember this, like he had a kind of tutorial back in the day. It was like I think how to make like a Twitter clone.

SPEAKER_00Yeah. I've I've had Dr. Hartle on the show. I think I might have built that like 10 times, and just each time I would feel a little bit more comfortable and learn a little bit more. Um, and then maybe you know try and take it in my own direction. And yeah, that was kind of my first toe in the water with Ruby on Rails.

SPEAKER_01Okay, wow. Um, and I'm curious, going back a little bit, um, you said that like banking wasn't really your thing or whatever. Any any particular reason why you felt like it didn't speak to you, or or it was just in general, like this is not really my thing.

SPEAKER_00So I was in the asset back securities group. Um, specifically, we were more doing like car loans, not home loans. And I was on the structuring side of the team. So it was actually like I'm core kind of a math logic person. So like we were in Excel, we were using like VBA. I was, I mean, to some extent it was like programming that was more, you know, and so building these like complex financial models was like very uh stimulating and fun and learned a lot, but it was more like at the end of the day, like what value is this bringing to society?

SPEAKER_00Like, how is this helping people?

SPEAKER_01I was struggling with that piece of it, so um, yeah, yeah, I suspect it if it was something like that. Um, because I don't really know anything about finance or banking or anything like that, but there is some fascinating complexity in it. Um, and and while I find that interesting, it's like, well, but like what what actual value is being created? It seems like a lot of um a lot of games um where where you're placing bets and stuff like that, and it's like that's all fun and interesting, but like it's not creating anything new, it's not creating any new value.

SPEAKER_00Yeah. So it was the part that I liked most about the job was actually teaching. So my my second year there, you know, we got to train the new analysts that were coming on. So by my second year, I that kind of clicked with me. I'd like I like helping people, I like teaching. Uh so that also helped prompt the big move of like, well, let me go try something completely different.

SPEAKER_01Okay. Um so you went to Japan, got involved with Ruby on Rails, and then and then um how did that turn into presumably a programming career and and that kind of stuff coming from teaching English?

SPEAKER_00Yeah, so the next thing I did after that um was I started making an app for my friend who was a translator. So in the translation industry, they have these things called cat tools, computer assisted translation. And at the time, like they had the kind of Photoshop, like you know, power user, like a lot of features, really heavy. And then they had the translators who were just in Microsoft Word, and each translation job they would start with a blank Microsoft Word document.

SPEAKER_00And they didn't really have anything in the middle ground. So I tried building something for my friend that was like a little bit more powerful than just like having to start from scratch in Word each time, but like, you know, didn't have any sort of the feature complexity of the real enterprise cat tools. Um, and so I got that to a pretty good spot. I went to advertise it on this site called pros.com, which was a site for freelance translators, and the sales that was P-R-O-S-E.com.

SPEAKER_00P-R-O-Z.

SPEAKER_01Oh, okay. But presumably pros as in P-R-O-S-E rather than P-R-O-S.

SPEAKER_00Like for professionals, I think. Oh, okay. Professional translators was kind of I think where that name came from. But yeah, PR-O-Z.com. I see, okay. So yeah, that that site was targeted towards like freelance professionals, helping them get jobs, uh, allowing them to ask questions. Um, but they had you know offers on the site that you could advertise to this network of freelance translators. Um, so I was talking to the sales guy, and he's like, Yeah, this is this is rough.

SPEAKER_00Like, you need X, you need Y, you need Z. Um, you know, come back to me. Like, if you I don't want to take your money now because it would be a waste, but if you if you build that stuff, you know, come back to me. So I kind of like stayed up all night and 24 hours later, I like called him up. I'm like, I built that stuff. He's like, No way, like, come on, there's no way. Um and it was that conversation, like, I should put you in touch with our owner. Like, you guys also have a Japan connection because you know his wife is Japanese and he spends some time in Japan.

SPEAKER_00So hit it off with the owner, and he, you know, ended up, I guess, kind of acquiring me. Like, I didn't get you know a lot of money for the app, but like he, you know, got to go, like that became my first development job, uh, and you know, spent five or six years at pros.com and learned a lot.

SPEAKER_01Interesting. Did you ever have any conception before that that you would become a professional programmer?

SPEAKER_00I think by the time like when I first started building that student database uh at the English teaching job, like I loved, I loved it. Um you know, I couldn't put it down from there. You know, just being able to make something come to life uh was super exciting for me. Like having an idea in my head and being able to like build it and make it a reality, I I loved that, and there was kind of no turning turning back after that.

SPEAKER_01Okay, so if I had done this kind of in the proper podcast sequence, I would have asked you right at the beginning, like, tell us about yourself. But if I were to ask you that question, you know, as of today, how how would you answer that?

SPEAKER_00Yeah, so now um I'm building Ambicky, which is a site for speech language pathologists and speech language pathologists' private practices. So it it helps them, you know, do everything from documentation to scheduling, insurance billing. Um and that I feel like has kind of been a melding of like what I worked on in the translation industry, which is some of like the natural language processing and that sort of area, but then also like my teaching career.

SPEAKER_00So when I was teaching in Japan and English as a second language, and I had some a point where I was teaching two and three-year-olds, that is like probably as adjacent as you can get to what a speech language pathologist does in the US.

SPEAKER_01So what does it what does a speech language pathologist pathologist do?

SPEAKER_00Um so we're targeting like speech language pathologists who work with pediatrics, and so this tends to be kids from birth, you know, all through elementary school who you know have some issue with articulation or pronunciation or have some speaking troubles, and the speech language pathologist works with them to help them improve uh their speaking.

SPEAKER_01Interesting. Okay, sometimes I feel like I could use that service for myself, um, which is ironic as a podcast host, I I like can't frickin' talk. Um okay, and and how did you get into building software for speech language pathologists? And what stage are you at with that business?

SPEAKER_00Yeah, back in 2018, uh one of my good friends who I worked with at the investment bank way back in the day, he had purchased a private practice speech language pathology clinic in Tennessee. Um, and he called me up and he's like, you know, we just bought this practice. The therapists are burnt out, like they have low retention rates because they're having to do a ton of documentation every night when they get home from work.

SPEAKER_00They have all these Excel files they have to fill out at the end of the month, like just so they can get paid. Like, would you have any interest in you know coming in and I like Kevin? I know you're good with software, like helping us build something to make these therapists' lives easier. So it sounded like a fun challenge. As I said, it kind of sounded like it mixed like different parts of my career. Um, so did that, and basically, like three years later, looked up and we had built kind of the whole operating system of this private practice.

SPEAKER_00Um, and it was at that point that we're like, you know, this is working really well for this practice. We helped it grow from 70 therapists up to 180 therapists. Um we could probably sell this, you know, more broadly across the US as a software as a service app. So that was when we kind of spun Ambicky out of that the umbrella of that one singular private practice and you know started selling that uh to other private practices across the US.

SPEAKER_00And so Ambicie, we've been at it now, I think, for four four-ish years.

SPEAKER_01Oh wow, okay. Um, how long did it take? I I mean, only share what you're comfortable sharing, obviously, but like how long did it I'll ask a few questions and you can take your pick of what you feel like sharing. Um, I'm curious about things like how long did it take to get your first customer? Um, where are you at now, like how many customers, or how much revenue, however, you want to, if you can give us a sense for just where things are at now, just that general trajectory, take your pick of any of that.

SPEAKER_00Yeah, so actually the first couple years were really rough because so we had built, we had spent three years building it, a bespoke system for that one practice. Um, and we had grown to a place where like they no longer wanted to pay for our salaries like and their practice. They wanted us to spin us out, but also we didn't have any software ready that was built from a foundation that it could, you know, that now support any practice. So I knew like from spin out, we're gonna have at least two years of building before we have the full kind of EMR as like electronic medical records software ready to go.

SPEAKER_00But also, like now we had funding and investors, and we couldn't just say, like, hey, we're not doing anything for two years. So we had to have some sort of bridge there. So we started off at first um with what we did have, which was like the resource library and the materials that therapists would use in their sessions, and also the teletherapy platform. So we had that stuff, but the first two years were rough because like we knew we wanted to be here and we knew we were gonna be there eventually, but our marketing positioning had to be here and had to like keep evolving every three or six months.

SPEAKER_00So I guess technically we we had some you know individual therapist users in those first couple years who were you know paying us for the resources, but it probably wasn't until year two or so where we had our first like private practice real customer that was using us as a practice management system. So, where was the gap?

SPEAKER_01Naively, I would think, oh, one practice is using this, clearly, other practices can just use it as is. Like, uh what's the problem?

SPEAKER_00So we learned a lot in those first three years. So part of it was we wanted to take everything we had learned and build from a solid foundation, but also like part of it was like we had built this very bespoke system that fit this one practice like a glove, but not all private practices are the same and they all have slightly different workflows. Um and then like we had to, you know, think of like, all right, so now we need to have the concept of organizations, and you know, potentially could a user belong to multiple organizations over time if they quit this one job and go to this other practice?

SPEAKER_00You know, could they be working for multiple organizations? So there were different concepts in the application itself that we had to build in that didn't exist when it was just building it for this one practice.

SPEAKER_01Did you end up rewriting the project from scratch or modifying the existing product?

SPEAKER_00Yeah, we we re rewrote from scratch um a brand new brand new Rails app.

SPEAKER_01Yeah. Wow, that's yeah, I I can understand that move. It's like if you have an organism that has evolved over millions of years to very specifically fit a certain ecosystem, a certain environment, um it's hard for that species to evolve into something completely different. Um probably that species just has to go extinct and a different species uh evolves into this uh new in environment that it needs to fit.

SPEAKER_01Um it's it's hard to it once you have a complex system, it's hard to run it in reverse and like make it more um repurposable to different use cases and stuff like that. Um so uh th this is interesting. Uh y you and I have some certain parallels that I've been noting as we've been talking. Um I've told the story on the podcast before, but I built a custom EMR myself uh some years ago.

SPEAKER_01A doctor contacted me out of the blue. It it in my case it was not speech uh speech language pathologists, it was uh an ophthalmologist. Um it was an ophthalmology uh clinic run by a husband and wife. The wife was the ophthalmologist, the husband was uh a pediatrician, but in the clin in in the ophthalmology clinic he just did the business stuff.

SPEAKER_01Anyway, he's the guy who contacted me, and they had uh they had like nine different systems cobbled together, and it was just a lot of overhead, a lot of like unnecessary overhead. So we built a new system from scratch using Rails, um, and it fit that practice like a glove. And there was never any intention. To make it work for anybody else. In fact, there was the opposite.

SPEAKER_01There was the intention to just keep it for ourselves. And we didn't want anybody else to use it because it was like that business's competitive advantage. And so we didn't go that direction with it, but it's kind of interesting because it's not every day you meet somebody who's built a custom EMR, but sounds like we both have.

SPEAKER_00Yeah, it's really cool. How did you did you end up with any? I think I feel like with EMRs, there tends to get to a lot of complex business logic. How did you handle that in the application? Or did you find that? Or you know, did you have to did you use any special techniques when you were building it that you're like, well, in most other kind of plain apps, I haven't had this, but in this EMR, like I experienced this thing?

SPEAKER_01Yeah. Um, so I learned a lot on this project because I think it was my first kind of like industrial strength uh greenfield project. Like I had done plenty of serious projects before and plenty of greenfield projects before, but this was my first big serious greenfield project. Um and you know, by this time, this was like 2018, I had been pro I've been programming since the 90s, so like I was already quite experienced by this time, and so uh I was able to I was able to bring to bear a lot of knowledge and experience right from the start.

SPEAKER_01But the thing is, I was unencumbered by the baggage of other people's work, and so I was able to just go pedal to the metal and go as fast as I wanted because it was um it was all clean. And you know, every greenfield project starts like that. It's easy to paint oneself into the into a corner quickly, which I've done a whole bunch of times. Um but now, you know, by by this time I was quite familiar with test-driven development and aggressive refactoring and stuff like that.

SPEAKER_01Um and so I was able to keep things um I was able to keep things well organized and easy to work with all throughout. That doesn't mean that every area of of it was perfect, but on the whole I thought it was quite good. Um and so, you know, were there any certain techniques that I used to manage all this complexity and stuff like that? Nothing other than the principles that uh that apply to all areas of programming.

SPEAKER_01If if I can try to briefly share some of the most um important ones, um test-driven development is is a huge one, and there's so much to that, like when I say test-driven development, the meaning that's behind that in my head is probably different from what most people hear when they say that, or hear when it's probably different from what people think when they hear that.

SPEAKER_01Um, but but long story short, um it's a process of Kent Beck has this great post called Canon TDD, um, and it describes a process of listing your specifications, picking exactly one, writing a test that describes that specification, uh letting it fail, and then writing just enough code to make that test pass, and then moving on to the next one and repeating the loop.

SPEAKER_01Um, so that is is one of the most important ideas in programming. Um another one of the most important ideas in programming to me is abstraction. Um, so like I think of it uh like if if if anybody's read the book uh Zen and the art of Zen and the Art of Motorcycle Maintenance, he talks about the analytical knife and he gives uh an example of a motorcycle. You can slice it up into you know the electrical system, the combustion system, the power delivery system from the engine to the uh drivetrain to the wheels or whatever.

SPEAKER_01Or you can slice it up in a different way, like here's all the rubber components, here's all the metal components, blah blah blah. And I think a big part of a programmer's task is to look at the world and apply the analytical analytical knife and decide how to slice up the world, um, how how to conceptualize things in a way that is truthful and useful, which might sound like duh, but you look at a lot of programs and their code is not really truthful.

SPEAKER_01It's it's it's not an honest uh reflection of reality, and it's not as useful as it could be, because the the concepts that are baked into the program are kind of at odds with with the user's domain, and it's kind of a mismatch and it's awkward to work with. So, anyway, I could talk for a long time about these different principles that I applied, but there's a little bit.

SPEAKER_00And maybe one thing you're getting at there, I mean, tell me if you've felt this or not. Like, I've found in my career that if you get the data model correct, everything will flow very nicely after that. But if you don't have like the data model in a good place, then like maintenance and adding a new feature and everything else becomes a mess down the road that I've spent I've tried to spend a lot of time up front making sure that I'm modeling it in the right way with the right naming and the right tables and the right relationships.

SPEAKER_00And if that is in a good place, like everything else seems to flow like magic. But if that if you get that subtly wrong, like it can lead to a lot of issues.

SPEAKER_01Yeah, I think that comes really close to hitting the nail right on the head. Um, I I'd put it maybe just a little bit differently. Um, because the data model, yes, like totally, and that's the way that I thought of it for a long time. Um, but now I maybe maybe it's just a terminology difference, but I I I talk about the conceptual model behind the the app and even the conceptual model behind the product.

SPEAKER_01Um and and then the data model I think of as being uh downstream of the conceptual model behind, I'll use the word system, the conceptual model behind the system, because uh you could think of the system as being like the most upstream thing, and then like the the product is downstream of the system and the program is downstream of the system, but the root of everything is the system, and the system has a conceptual model behind it, and this EMR was something that helped me realize that.

SPEAKER_01Um because like the quality of the conceptual model sets the ceiling on the quality of the code. Um we had some certain concepts in in the program. Um, you know, it dealt with four areas scheduling, billing, administrative, and clinical. And most of that was pretty sound conceptually, but we had some conceptualizations in the clinical area that I felt were a bit muddy.

SPEAKER_01Um and my my boss, this doctor, was like maybe the smartest guy I've ever known. Uh, we had one of the best working relationships I've ever had with anybody. Um, we just really like we we gelled really well. Um but w one of the one of the few weaknesses there was that we never could like quite get on the same page regarding the conceptual model behind this clinical area, um, certain parts of it.

SPEAKER_01And I I realized at one point, like I'm really unhappy with a lot of the code in this area, and it's um hurting our productivity. Um and so I wanna I want to clean up this code, but I can't clean up the code because cleaning up the code would require getting some better clarity on on how the things that we have to deal with in this area are conceptualized.

SPEAKER_01But I couldn't do that on my own. I I didn't it wouldn't have been right for me to just go in and start changing all that. It would have to have been a collaboration between at least myself and my boss, if not myself, my boss, and other people. And and so that's that's when I realized, like, oh, like I I can't just make the code as good as I want to make the code. Like, that's downstream of the conceptual model.

SPEAKER_00That makes sense. Did you ever feel any, or did you ever ever have any struggles with scoping? I feel like with EMRs, it's one of those things that is never ending, expanding. It's like, what is that? Maybe like Jamie Sabinsky quote of like every program attempts to expand until it can, you know, read email or whatever. Like one thing with EMRs is like it touches on so much that like I don't know, how did you scope that project or how did you keep it in a way that it didn't end up into the never expanding of trying to do everything?

SPEAKER_01Oh, I mean, we didn't try to do that. We we of course tried to be judicious about what we chose to take on, um, because your time can be spent in better and worse ways. So we tried to be smart about you know, for if we're gonna invest a dollar, let's shoot for the thing that's gonna give us a hundred dollars back instead of two dollars back. Um but we never said like we're gonna limit this product to a certain size, it was just uh however much value we can add to this app, let's let's add it.

SPEAKER_01We're never done. There's and there was always something screaming for priority, you know. We my my feeling is that like my feeling, my experience is that every software product is always in a state of being like way behind. Um behind relative to like everybody's wishes, like it it it always feels like it's painfully behind where you want it to be, and and this was no exception.

SPEAKER_01So yeah, we didn't attempt to to limit scope, we were always trying to fulfill the latest needs.

SPEAKER_00Very cool. Yeah, not every day that I meet someone else that worked on uh EMR. So that's it.

SPEAKER_01Yeah, how how about you? Um, you know, the managing all the complexity and the scope and stuff like that. How did you approach that?

SPEAKER_00Yeah, um, well, on one hand, that's one thing that I love about the job and excites me is that uh I feel like 10 years from now, there's still gonna be, you know, things to work on and things to build. Um, but how do we prioritize uh yeah, that's a great question. Um so that was something that you know I I added to the book and something that we tried to do a little bit differently, especially in our last few planning cycles. So we've gotten a lot of inspiration from 37 signals and like their shape up methodology.

SPEAKER_00But one thing that I found was happening was like there was this pressure to have a roadmap. And a roadmap tends to be like a list of features. Um, and if you think about like a map, a map can be very detailed, it has you know streets and buildings. Um but part of my frustration was like, you know, how did we get there? How did we get to this list? And like, did it just come from uh a customer support ticket?

SPEAKER_00Did it just come from sales saying, like, hey, the competitor launched this?

SPEAKER_01Um, I felt like wait, sorry, before you go further, you I I think you said there's pressure um to have a roadmap.

SPEAKER_00Um, pressure from where pressure from our like investor advisory board sometimes, of like, hey, you know, what are you guys working on? Sometimes pressure from customers asking for like, you know, what's the roadmap? Um, but I would say more so I was speaking from like uh the investor side of things.

SPEAKER_01Yeah, okay. Okay. Um, yeah, they wanted you to have a roadmap, and maybe it's a so for me, when people want me to like uh share what's coming down the pike, like far into the future, I'm just like, like, I don't know, like I can give you a very general prophecy of what might happen, um, but that's very much subject subject to change and low resolution.

SPEAKER_01Is is that how you feel about it? Or how yeah?

SPEAKER_00Definitely, but even like uh even in the short term, I felt like even if we're just planning the next like 60 or 90 days or eight-week cycle or whatever, I still felt like we were kind of jumping to solutions and kind of skipped some of the discovery and the problem step. Um so what we decided to do a little bit different was this. I first just kind of sent out a survey to the team of everybody on the team, like, tell me what you think are your in your mind your top three problems that we need to solve over the next quarter.

SPEAKER_00Um, and got some like really good things. Uh, I cleaned up that list, I you know, merged any that were duplicates, I kind of took ones that were trying to like stuff like solutions into it and like just pulled out that and just focused on the problem. And then I did a second survey, which took those and I asked like a smaller group, like more like a leadership team, of like, hey, rank these problems of like on two scales of like one, you know, how big of the customer base does this impact, in your opinion?

SPEAKER_00Like, are you hearing this from like one customer or is this gonna help you know 90 customers? And two, like, you know, how much is this gonna move the meter if we solve this thing?

SPEAKER_01Um you you mentioned jumping to solutions, and this might be a decent um decent uh uh opportunity for me to ask what what I've been eager to ask, and tell me if you want to push this off until later so you can continue what you were saying. But um your book, um uh the the problem first method, um when I saw it, I was positively filled with glee because I've seen so much not just past tense all the time.

SPEAKER_01I still see people just jumping to solutions. They like it it's like if if if doctors behave like programmers and even engineering managers, a patient would come in and they'd be like, hey, my arm hurts. And you'd be like, Oh, well, I got Tylenol. Here, have some Tylenol. The the that'll uh that's what we should do. It's like whoa, whoa, whoa, like wait, like we do not understand what the problem actually is.

SPEAKER_01But you know, unfortunately, that's seen as like delaying and uh seeking perfection before we do anything and stuff like that, which frustrates frustrates me to no end because I think it's just frankly really dumb. But uh anyway, um I've I'm I'm curious about uh what led you to this book, the uh the the problem first method, because it sounds like you really understand this thing of jumping to solutions and not jumping to solutions.

SPEAKER_00Yeah, and I think there's two things that are making it like even more difficult in today's world. One is that I've observed that versus five or ten years ago, customers are much more sophisticated today. Like five or ten years ago, like they weren't using web apps on a daily basis necessarily. They didn't have it, it wasn't part of their daily life. So they would come to you with like this vague, like, can it do this? And if they and if it could, they're like, oh my god, technology is so amazing! Like, I don't believe that.

SPEAKER_00And you had the opportunity to do more discovery and to ask them questions, like because it was still like a rough idea in their mind. They were coming to you with this like pain point, but they couldn't necessarily articulate it. So it was a great, easy way for you to jump in and get more data there. But now I find customers are much more sophisticated, like everybody's using web apps on a daily basis, they've seen you know great apps, so they're not coming to you with this like ambiguous pain anymore.

SPEAKER_00They're coming to you and saying, like, I want to export to Excel button here, or I want SMS messaging. And so that's that's a great observation jump and just say, like, yeah, we'll do that. Instead of like taking a pause and like really trying to dig in like why they're asking for it. It you know, at the end of the day, like that might be the thing that you implement, but it might not be. Like, it might be that you find that their problem is a little bit different, and if you solve it in this different way, it might eliminate this whole other class of uh you know issues they were having.

SPEAKER_01And you know what my favorite move is, is when we encounter an apparent encounter an apparent problem, gain an understanding of it, and our decision based on all of our weighing of options and stuff like that is just no, we're not gonna do anything here. Uh, move on to the next thing. Because it's like a puppy. Um when you when you solve a problem by addition, um you're you're adopting a puppy, and now you have to feed and water that puppy, and you're gonna be uh, I don't want to say stuck with the dog.

SPEAKER_01Um it's you know, you have this nice companion, but but um you you're gonna be stuck with this freaking dog for the next 15 years. Um, and so you better be really careful about what you choose to take on. That might seem to contradict what I said earlier about the the EMR that we were building, and the scope never stopped growing, but we were careful about, you know, we were always adding things, but we weren't cavalier about what we were adding.

SPEAKER_01Um, we only added things that were going to be worth not just the initial price of building it, but the ongoing cost of ownership.

SPEAKER_00Yeah. And I think the other thing that is working against teams these days of who want to be like problem first, who think like, oh yeah, I should definitely do that, is AI. I think AI in people's minds, like it like it makes the cost of building, they think the cost of building is free, or like it costs less than it did before. And so like it's so much easier to be like, hey, I'm gonna do a spike on this, or just like let me, I can, oh yeah, I can add that you know, export to Excel button like that.

SPEAKER_00And I think that is dangerous because like it it it's easier to. Very quickly go in the wrong direction, or you know, as you just said, like, you know, add that puppy to the household that now you have to take care of for the next 10 or 15 years.

SPEAKER_01Yeah, both pre-AI and post-AI, people are it's really easy for people to think about the upfront cost of adding a feature, but they don't think about the cost that stretches into the indefinite future of owning that feature. Um, and something that I wish people would think about a bit more with regard to AI is what programming principles have changed and what principles are the same.

SPEAKER_01Um and really that's not even the right question because none by definition almost uh principles don't change, uh principles are just true. And so what are those principles that were true before and are still true now? And what's different? Like what's different is that a lot of things are cheaper, but not everything is cheaper. And just because we have this incredible technology doesn't mean that two plus two isn't four anymore.

SPEAKER_01Like there's certain immute immutable truths of nature that are are never gonna change no matter what. And it seems like people uh a lot of people haven't really realized that yet.

SPEAKER_00Yeah. The other thing I've observed is that sometimes it can be somewhat of an ego thing of, you know, if one example from Ambicky was, you know, the the sales team came to us and like, can you can we add discount codes? And so you know, my first instinct as a developer is gonna be thinking, like, oh yeah, the Stripe API, and we can do this and we can do this. Um you're already imagining like what that feature is gonna look like.

SPEAKER_00Um but then I caught myself and I'm like, all right, well, this is a discount code for uh this was for a convention that we were going to. Like, how many realistically, like how many customers are we gonna have go buy our booth? Um, and we're like, all right, we did the math. It's like, okay, it's gonna be like max like 100 or 200. And like at the end of the day, it was like, we can just manually do this discount for them because, like, again, like these were bigger customers, they're gonna be having a demo anyways, they're gonna be talking to sales.

SPEAKER_00So if I had just gone with my like initial thought of like, oh yeah, I'm I'm a developer, I can build that, I know the Stripe API. I would have spent you know a couple weeks building out this discount code thing that we just really didn't need, and we could have just you know had sales do it manually and enter their subscription.

SPEAKER_01I I love that. I love that you thought all the way through that because that sounds like a case where you'd be you know metaphorically paying ten dollars to get a return of one dollar because if you know you have a hundred people coming by your booth, how many people are gonna be prospects out of that? Like one or maybe five if you're lucky, and you're gonna build this whole new feature for not even five customers, but just five prospective customers.

SPEAKER_01That seems like a very high price to pay for what you'd get in return. And it sounds like you realize that.

SPEAKER_00But yeah, I feel like there's a lot of examples out there where you know, as a developer, your first instinct is gonna want to be like, oh, I can do that, like I know how to do that, like I could use this gem, or I could do, you know. Um, so it can be tough to like kind of put the brakes on that and be like, oh wait, well, does it make sense to to implement this? And you know, what what is the bigger picture problem here?

SPEAKER_01Exactly. Um something there there's one other comment um I wanted to make, and then since we're getting close to time, this this time has flown by, but um, I wanted to ask you if there's anything else that you wanted to touch on that we haven't yet. But um you were talking about gathering feedback from customers and stuff like that, which is obviously good. Um, and I'll I'll make a comment that is probably preaching to the choir for you, but something that I used to tell my consulting clients is like don't just collect feedback and ideas and problems and stuff like that from your customers, but like understand their world and then you decide.

SPEAKER_01Like there is a way to do it where you're basically just outsourcing the design of your product to customers, which might work out well, but it also may very well not work out well. Um or you can go and gain a deep understanding of your customers' world and the job that they need to do. So far as I used to advise all my clients like go physically to your customers and like sit with them and watch them work and understand their world, and then you go back and design your product, don't let your customers design your product for you.

SPEAKER_01Um, and again, it sounds like it's probably preaching to the choir in your case, but that's like it it kind of goes against it, doesn't go against, but it's it's not quite the same thing as what we're often taught because we're we're taught so much, like listen to your customers, talk to your customers, uh blah blah blah. But it's like that's that's good, but I but to me it's not quite far enough. You you still have to go understand their world.

SPEAKER_00Yeah, 100%. Maybe there's two things I would add on to that. So one thing that I preach to my team often is I feel like there's this spectrum of on one side you have the vision-driven companies, maybe like an Apple, and on the other side would be uh, I don't know, maybe like a Yahoo and kind of like the the tail is wagging the dog sort of thing, where I think if you are just a hundred percent like listening to customers, you're just gonna be putting on a bunch of feature after feature, and five years down the road, you're going to end up with this product that has no clear vision that is it feels awkward to use.

SPEAKER_00So I think you need to find that balance where you have this very vision-led product, but also you're humble enough to like take feedback along the way and make sure, as you said, like you're you're visiting your customers, you're talking to them, you're you're understanding that. But the thing, the way that you work in their feedback fits into this strong vision that you have for the product.

SPEAKER_01Yeah, yeah, a metaphor is hatching in my mind where somebody has a pet dog, different pet dog metaphor, but they're like, you know what? Can we make uh my dog have wings? And then it would be great if um the dog could also uh swim underwater for extended periods of time, um, and like all these different things, and it's like at the end of it, you just have this horrific uh monstrosity with all these different animal parts, and it's like, what is this thing?

SPEAKER_01It's certainly not a dog anymore. Um, and great examples, Apple and Yahoo, because with everything uh Apple did during Steve Jobs, it all their products had like an essence to them, you know. Um, whereas Yahoo, it's like, what is the essence of Yahoo? Nobody knows what it is. I like they didn't even know what Yahoo was, you know. I don't even know if Yahoo still it exists.

SPEAKER_01Um it's it's deeply irrelevant at this point as a um as a company. Um but yeah, uh in in the bit of time we have left, is there anything that you wanted to touch on that we haven't yet?

SPEAKER_00Maybe just the one other piece I would add to that conversation is I I've seen I do think it's super important to talk to customers, and I've seen so many teams going in the opposite direction where they want to outsource their support to AI and chatbots. And we've kind of taken the opposite thesis where like we think it's super important to have a human customer service team because like that is just a goldmine of like being able to like not only have that customer relationship to reduce churn, but just so much you can learn from those conversations, from those support tickets.

SPEAKER_00That I don't know, I don't know your thoughts on that, but I've just seen so many teams like rushing to see like how fast they can you know cut that and have uh you know the AI chatb handle everything. Um, but we've seen so much value in the opposite direction of that.

SPEAKER_01Yeah, I haven't been close to that kind of thing personally, but I can certainly believe it. That reminds me of another thing that makes me very sad, which is an over eag an over-eagerness to use metrics and telemetry and and that kind of stuff to try to make inferences about user behavior where it's like, hey, the guy is like right over there, like just go talk to the guy, and you'll get like uh incomparably rich understandings from that, whereas not only are the metrics missing uh a lot of essential color, but they will actively mislead you into believing things that aren't true.

SPEAKER_01Um, I've I've seen this happen where people waste so much time and money like astronomical amounts based on a misconception of reality by looking at at metrics instead of having like a one-hour call with one person that would totally change their whole view of everything. Yeah, yeah. Um okay.

SPEAKER_01Well, I wish we had several more hours because there's a lot of different topics that I think we could talk about for a long time. But but sadly we're yeah, it has. Um we're getting close to the end here. Um any links you want to share and and that kind of thing before we go.

SPEAKER_00Sure, you can find me on LinkedIn. My personal site is kevin sdiaz.com. The book is at the problem is at problemfirstmethod.com. Um yeah, come say hi.

SPEAKER_01Okay. We'll put all that stuff in the show notes. And Kevin, thanks so much for coming on the show.

SPEAKER_00Yeah, and Jason, if you uh ever want to come to Japan for I know they have the the kaige on rails this year. Dave Thomas, who you had on the show, is gonna be speaking there. Uh, look me up and would love to meet you in person and show you around.

SPEAKER_01Absolutely. Uh, we'll make it as soon as we can.

SPEAKER_00Thanks, Jason.

Open the transcript file

Hosts and Guests

Nobody is credited on this episode yet.

Credit the people on this episode

Be the first to rate this episode

More episodes

All episodes
  1. 338 - Code Review, Testing, and Staying Sharp in the AI Era with Joel Hawksley

    Nov 18, 20261 hr 5 minAudio

    In this episode I talk with Joel Hawksley, Staff Software Engineer at GitHub and creator of ViewComponent, about whether AI is making us dumber, why blocking code reviews struggle under a flood of AI-generated code, and...

    Credit the people on this episode

  2. 337 - User Testing, Paper Prototypes, and Escaping Hourly Billing with John Athayde

    Nov 17, 20261 hr 18 minAudio

    In this episode I talk with John Athayde, designer-developer and co-author of The Rails View, about what architecture school taught him about software design. John tells the story of a user test at LivingSocial that...

    Credit the people on this episode

  3. 336 - Riccardo Carlesso, Co-Organizer of Rubycon Italy

    Nov 13, 202646 minAudio

    In this episode I talk with Riccardo Carlesso, Google developer advocate and co-founder of Rubycon, a Ruby conference held in Rimini, Italy in May 2026. Riccardo explains how a joke about PyCon became the conference's...

    Credit the people on this episode

  4. 335 - Blastoff Rails: Starting a Ruby Conference from Scratch with Travis Dockter

    Nov 11, 202654 minAudio

    In this episode I talk with Travis Dockter, organizer of Blastoff Rails in Albuquerque. When Sin City Ruby and RailsConf both ended, Travis decided to start his own conference. We discuss how cheaply you can run a...

    Credit the people on this episode

  5. 334 - Merge Queues and Building Mergify with Julien Danjou

    Nov 10, 202655 minAudio

    In this episode I talk with Julien Danjou, CEO of Mergify, about merge queues: why PRs that pass CI separately can still break main together, and how testing changes in a queue keeps main green. We discuss how Mergify...

    Credit the people on this episode