MarsBased podcast - Life on Mars
MarsBased podcast - Life on Mars is your go-to space for technology, entrepreneurship, and innovation.
Brought to you by MarsBased, a high-end development consultancy specializing in Ruby, JavaScript, and Python. If you are a software engineer, you can expect deep dives into tech, but we go far beyond the code. We share our first-hand expertise on remote work, company culture, management, and digital transformation based on our journey building a successful lifestyle business.
Every two weeks, we sit down with the best founders, investors, experts, and celebrities from across the galaxy to hear their stories.
After years of investing heavily in the startup and tech communities in Barcelona and worldwide, we are bringing that same value to you through this channel.
Follow us to join our journey! 🚀
MarsBased podcast - Life on Mars
I fired a huge-name client to protect my team | Adarsh Pandit
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
Adarsh Pandit (founder of Cylinder Digital, ex-Thoughtbot, ex-McKinsey) met Àlex 12 years ago in San Francisco when MarsBased was just getting off the ground. They catch up on a decade of building software agencies, navigating career shifts, and managing team dynamics.
Adarsh shares the story behind firing a high-profile client who was verbally abusing his team, the panic of dropping a production database on Heroku, and how applying systems thinking changes how you look at code, sales, and business operations. They also touch on how AI and vibe coding are reshaping agency pricing, the return of fixed-bid contracts, and why human trust remains the most critical asset in tech.
🎬 You can watch the video of this episode on the Life on Mars podcast website: https://podcast.marsbased.com/
Adapting Fast Without Losing Yourself
SPEAKER_01When your your company depends on you to sell work, that is not the best use of your time. The world is changing. My local economy is changing. My home is changing. Whatever is changing, you always have to be adapting. And I think you have to be honest with yourself about all of that and to be on your own authentic self. You can't really pretend to be someone else. I dropped a production database. I remember the feeling. I remember your face turns red. I was sweating. I was so ashamed and embarrassed. Former colleagues and other new colleagues who were some of the best designers and developers that I'd ever worked with, they all hated doing sales. They hated it. They didn't want to go out, they didn't want to talk to people, they didn't want to network. I fired a client, which I won't name, but is a very, very big name. And almost immediately I realized that they were verbally abusive to my team. So it needs to stop, or we're done. And it didn't stop. And so I said, we're done.
The Email That Started It All
SPEAKER_01Good morning over there. How are you doing? Good morning, Alex. I'm great. How are you? I'm excited about this. I'm I'm deleted. Yeah, me too. We've been talking about this for a long time, and you've been bugging me to be like, oh, you gotta come on, you gotta come on. And I here I am finally.
SPEAKER_00I totally respect everybody's rhythms and paces, and we've known each other for such a long time that I know that uh we wanted to prepare these to have a good episode. We've known each other since 2014. You actually met me before I was I even incorporated Mars-based.
SPEAKER_01Right before you started. Yeah. How did you find me, by the way? I don't remember how we actually got connected. But I remember at the time I was running the San Francisco Thought Pod consultancy office.
SPEAKER_00We back in the day when, because that was my first foray into going to San Francisco. And I quit my job, and in the process of creating March-based, we said, like, I said, I'm gonna just go to San Francisco for the three months of my ESTA uh um visa. And so we messaged pretty much every company out there doing Ruby on Rails. And out of the 300 companies we wrote an email to you, there were only like five that answered back, and you were the first one. So wow, thank you for that, my teep of the hand. No problem. I felt special for a little while, but now I realize I was just the first one to answer that that you know that friendship or this friendship that has lasted for at least this 12 years. Long time, yeah. Yeah, reproof. Um, and I remember we met at the at the creamery, if I remember correctly, close to third and like the the train station. Yeah, and you were there with I think Yeah. I was a friend. Who else was there? No, I was alone that time. Oh, you were yeah, it was my my first time, and it was I was I was alone a couple of times later on. I came with friends. But uh the thing is I remember that you're like, yeah, I'm I'm like happy to meet you, happy to discuss code and stuff like that, but we don't outsource at ThoughtBot. Like, is there anything else I can do for you? We like we had a good chat, we continued chatting over the years because uh you mentioned that you might go to Barcelona for a conference, kind of like things start rolling from there, and lo and behold, it's 12 years now.
SPEAKER_01So yeah, thank you for this. I remember that conversation because you said, Oh, I'm starting this consultancy and we don't know what we're doing. And what what advice do you have? And I was like, I don't know. I mean, I have I know what I know from my experience. I don't know if that's helpful for you or not, but I'll tell you what I know. And so I wrote some stuff down. I remember we met at the creamery, and then I said, Okay, well, here's what I know, and I just told you everything I knew, and you just scribbled it all down, wrote, row, row, row, road, and you're like, Okay, sounds good. And then you took off. And now look at you all these years later.
SPEAKER_0012 years later, 30 employees now, like you. You're one of the very few people that I've met consistently every year, and you've seen the progression from zero employees to three employees the first year, then second year, like six or seven, you know, going two, three at a time. And uh now we haven't met for the last two years because I haven't been able to come to the US for family reasons, but that might change soon. And uh now that we're 30 employees later, uh, you know, happy to report that we're alive and kicking.
SPEAKER_01So well, no, more than that, I think you guys have done an awesome job. And like I just it's g it gives me great joy to have watched it gives me great joy to watch my friends slowly succeed over the years, right? And go through ups and downs and economy up and down and family changes and you know, technology changes and everything, and and still manage to figure it out over the years. So I don't know. Congratulations on that.
SPEAKER_00Thank you. Um, one of the things that I would really wanted to meet you back in the day was because you're when you're starting out, first-time entrepreneur, um, you literally know jack shit about Everything. And that thing you said that I Verba team said the first time I met you, like, we don't know what we're doing. Uh yeah, it is 100% correct, 100% accurate. We didn't know what we were doing. And we figure out that by learning from companies we really look up to, and we don't look up to many companies, I have to say, but Thoba was one of them. High standards of quality for code, big proponents of Ruby on Rails, uh, remote work, great products, and and all of that. Like super always sharing, um uh caring about community organizing events, all the open source work, yeah. All the open source work, the podcast. So for us, it was like a big and large, uh, by and large, it was a great inspiration for us. So, you know, it was super generous from from your side to share that with us because other people might perceive it as competition. It was like what what should have helped this guy to kind of like uh get some clients, and like maybe five years from now he's gonna be stealing our clients. So uh we don't find a lot of people with that mentality, so so thank you for that as well.
SPEAKER_01I never subscribed to that, I guess. I maybe to my detriment, but um, you know, especially in the 2010s in the Bay Area, there was so much work that there wasn't really a feeling that, like, oh, either you get this contract or I'll get this contract. And that our survival depends on me beating you or any of that. And it's not sort of baked into who I am. So from the beginning, uh, and you know, maybe we'll talk about this more, but like when I started uh doing sales in in that office, I didn't really know what I was doing either. And part of it is I think you have to collaborate with other people. You have to do what you said, right? You send all these emails out and you reach out to other people. I had this new job managing director, and I thought, okay, I kind of I definitely know consulting. I know about sales, I know about technology, but like, how does this local market work and what are people what works for them and all of that? I started getting in touch with other managing directors of other consultancies in the area. Um, you know, there was Pivotal and some other ones, and we used to get together and have drinks, and I would say, you know, how do you handle this situation? How do you handle that situation? And I think that that humility or you know, that kind of like authentic curiosity is so important uh in anybody's career because other people have information that you don't have, right? And if you keep pretending that you have it, people will see through it. Um and also the other thing is I guess maybe in the same example as as you found with me, people are very willing to help. If you check in with people and say, hey, seems like you're doing something pretty cool, I'd love to talk to you for a little while and learn how you did that. I'm trying to do something similar or different. Um I think that by and large, not everybody, but I think for for the most part, people are generally, if they have the time and energy, interested in helping you out.
SPEAKER_00That is a well, that's a very Californian mentality, right? The paying it forward, the always looking to help others and for helping yourself. That's one of some of the cultural traits from California that had then resonated with me, also being part of startup brand that lives by the same core values. I think that's something that also shaped the the history of Marspace. So, from
Paying It Forward In Consulting
SPEAKER_00the, you know, from from the time that I met you, like not long thereafter, you quit ThoughtBot and you went solo, if I remember correctly. So, yeah, on the topic of what we wanted to cover today, what I wanted to discuss uh with you, because I've seen you transforming your role and your career path many times in these years, right? And um, and you're a very interesting profile. I think that's why we get along well because you live at the intersection of technology, business, and product, right? I think like you do well off the all of these strings, uh, these things, you're very well balanced, not a lot of developers are like that. And so um I'm curious to hear about whether those moves that you did afterthought about were something intentional or not, or how did you how do you plan for these mentality changes, role changes in every stage of your career?
SPEAKER_01So that
Getting Let Go And Going Solo
SPEAKER_01that change was not um something that I had planned. That was uh I was let go as part of restructuring. And initially I was very upset about it um because I had such good connections with my developers and designers and the other managing directors that I worked with. I still am very close with them. Even last week, week before last in uh RubyConf Las Vegas, where this t-shirt, beautiful t-shirt is from, uh, I still connected with lots of people that that I used to work with. Um and I'm very close with. But, you know, it took me a little time to say, all right, well, what is it, what is it that I'm good at and what do I want to do? And I got in touch with my clients and they said, you know, we want to work with you. And so a lot of my clients came with me and I started doing kind of the same sort of work from there. Um it, you know, I think part of I have done a lot of reinvention over the years, and and it has been certainly challenging, but every new phase has been so rewarding and enlightening and fun. And a little closer to uh the things that I enjoy doing, or the things that, you know, frankly, like the economy is ready to to pay for, right? Like if you want to do some job, you should ideally have a market for that. So it's not, you know, some of the some of the changes in my life, and there have been a lot of them, have been planned and some have not. I also in that time I had uh kids. And so, you know, 2015 I had a two-year-old who just turned 13. So um I had a small child, uh, much like you do, and your priorities change immediately. Um during that time, a few years before, I had an offer from Google, and uh I talked through it with them and I was super excited about it. And the team that they had recruited me to, I was very into. And then they said, Well, you have to work all these hours, and you have to ride this bus, you know, an hour, hour and a half each way every day. And I thought about it more and more, and my kid, I don't know if you know this, you you've probably experienced this with your child, but like your kid grows so fast, even at the beginning, that sometimes at the beginning of the day they look a little bit different than they do at the end of the day. And I just felt like I felt like I really wanted to make sure that I had time and space to spend with my family, which like work is always fun and and important and necessary, but I always asked myself, like, what's the point of all this? And the point for me has always been family. And so what I did there is I structured the company that I built to be very flexible, to have low overhead, and to be able to take contracts and um on an ongoing basis, and then if something came up to be able to turn it off. So, for example, in the summer, I don't typically take contracts. I'm with my kids. We're doing stuff, you know, sometimes they have things that they're doing, or we'll go on vacation, or we'll do camps, or go biking, or hiking, or swimming, or what have you. So because, you know, my kids are my kids are now nine and thirteen and they're getting older and they're doing their own thing. And the older they get, the closer they get to leaving the house or you know, having their own interests and not, you know, teenagers, they don't really talk to you very much as time goes on. So I just think like I it's always been important to me to spend as much time with them as I can. And so I structured the company in a way to make that work for me.
SPEAKER_00That must have been some twenty ten years ago, something like that. I remember you went for like you created Cylinder Digital, your own studio, and you went from like, oh, I thought about we don't outsource to all of a sudden, guys, I've got all of these projects, so we can work together now, and we started collaborating more closely with like certain joint projects and and stuff like that, which are we're very grateful for. So I also want to take the time to mention that you've been a big evangelist of Mars-based over there, some of the contractually one over there have been largely thanks to you. So uh my gratitude as well. And um of course, but uh you've been super balanced in the the way like you could you could have devised the way like every other engineer would have done this, which is like, hey, I work on the projects, and uh, whenever I finish the project, then I start looking for the next one. But you've
Why Founders Must Let Go Of Code
SPEAKER_00been you had a different approach whereby you know we have to spend this many amount of hours only developing, but the other amount of hours are doing sales while you are developing. Not everybody understands that. I used to do that when I was a freelancer, which I would work half of the day in the project, the other half was would be like marketing, admin, and sales. Therefore, I never had an like my buffer was always uh was always full, right? Um and you can't do either one of those for you, I don't think. Yeah, exactly. What was your distribution like? How much did you spend?
SPEAKER_01So I would say I would say ThoughtPod, it was almost entirely sales. I I stopped doing development work early on. I still tried to pitch in where I could and do code review or clean stuff up or do some refactoring because I loved that work. I still really love doing that kind of work, but um, it was not the most valuable thing to do. I think part of under you know, all these, and I'll talk a little bit about kind of where I came from in a minute, but like I think part of part of moving through your career is understanding what you're good at, what you enjoy, and what is helpful to other people, right? So even though it was very enjoyable for me to write code all day where you know you get into the flow state or you're pairing with somebody and you really are in sync and you know have a great session and you produce some work that you're really happy with, um, when your your company depends on you to sell work, that is not the best use of your time. And so um, you know, I had to let that go. And and I think at some point, just like you realize like it is a full-time job, especially at so many meetings. Sales is a lot of meetings, it's a lot of talking, which I enjoy. I enjoy chatting with people, as you know. But um the process of talking to new people about their project, understanding what it is at a kind of deeper level, sketching it out, having a back and forth, and saying, I think your project looks like this, I think I would do it this way if I were you, I think it probably costs this much, I think the team looks like that. That takes a lot of time to do correctly. And I I don't think there's room for much else, in my opinion.
SPEAKER_00100%. One of the things that I think like we've been largely successful for, successful, well, don't quote me on that. It's because my profile has been 100% devoted to sales since the get-go, right? And um and Chavin Jordi put it bluntly another way, they said like the company is gonna go well as far, like so far, you don't touch any code in the company. So that probably speaks at length about the amount of trust they had in my technical chops. No, but it meant like it's the other way around. Like it means if Alex needs to go back to the code, that means they're not he's not selling projects, right? So I think like uh you put it uh like you explained very well. Like, if if other people's food on the table depends on you, you have to let go of your so-called like passion and focus on what drives the most value. And I know I would be the most expensive developer in the company, and I'm the most uh most expensive marketer in the company, but the moment I do sales, I know other people could do it. Chavi helps me with that, he's really, really fucking good at doing sales, but I know then the opportunity causes like he's not doing the CTO work, right? So uh somebody's gonna do it. And so um companies like other, like precisely other other dev shops that probably you've met some of them, we've met here in Barcelona and some companies are super talented in the in the technical department. But when you know, and in tailwinds, everything goes fine. Whenever we face some like strong headwinds because the market changes, there's a crisis out there, they were not able to go out there and look for projects because the projects they were not coming their way, right? So yeah, but on the other hand, like companies like us that we have this technical business development um specialist in the in the company that we were able to endure this kind of situation. So that that's why I was saying, like when you went solo, do you remember what was the distribution of time, like 50% sales, 50% um execution, or it was I I right away, I I think that no, I think that so I've been doing consulting since 2006.
SPEAKER_01I started at McKinsey and company doing corporate strategy consulting and operations work. Um, so I've been doing it a long time, and uh it took me a while to understand how it worked, but a consulting company really is very different than a product company. You have to constantly be out talking to people, networking, and selling work because projects end and you have to keep the flow coming, otherwise your company falls apart, right? So, like um, it was definitely 100% right away. And then part of the the, as I said, I wanted to make this work well for my family. My wife has a very demanding job as a physician and uh often is you know, nights, weekends, all of that. So to be able to maintain that flexibility, I couldn't be on the project full-time every day. That wasn't gonna work for my home life. So what I decided, I thought about it quite a bit. And I thought, okay, what is it that I enjoy doing? What am I good at? And what is useful to other people, right? And in the in the developer world, in the technology world, uh, how should I say this gently? Not everybody is great with having uh a conversation with other human beings. Right? Human beings are an API. Exactly. Which is, you know, uh it's great in a lot of ways. And certainly I enjoy my my alone heads down time also. But um, you know, what am I good at? I'm good at uh having a chat. I like you know having a conversation and coffee and lunch and things like that and meeting new people. I'm genuinely interested in them, which is, I think, something that usually comes through. Um what is most valuable is that I would say, especially at a small consulting firm, is really like you have to bring work in. Otherwise, if you have the greatest developers in the world sitting on the bench, what does that even mean? And then the other thing was a recognition that my former colleagues and other new colleagues, who were some of the best designers and developers that I'd ever worked with, they all hated doing sales. They hated it. They didn't want to go out, they didn't want to talk to people, they didn't want to network, they didn't want to, they didn't want to do the invoicing, they didn't want to negotiate rates, they didn't want to have sign contracts, they didn't want to chase down invoices. You know, people wouldn't pay them, and then they would say, Oh, well, what can you do? And I didn't mind doing any of those things, and I thought I was pretty good at a lot of it, and it was valuable. It was valuable to all these other people. So what I did was I would go out and I would sell work, and then I would call my friends and staff the project. Sometimes that friend was you. And I felt use, I have always felt in that way useful, and I felt like I'm doing what I want to do, what I enjoy doing, and what's most valuable for the people around me. And everybody benefits, you know, you create a lot of transparency around that, and and this is the project, and this is what it costs, and this is where I think it's going, and and you develop a high level of trust. Um, and I think I learned a lot of that. Uh, there was a lot of certainly challenging things about working in at McKinsey, and it's definitely changed quite a bit since I was there. But um I I I didn't belong there for sure, but I learned a ton about client service, integrity, um, and how consulting projects work from that time there.
SPEAKER_00I'm excelloid. I can totally relate. I always say that I learn a ton. about how to do things, but I learned twice as much about how not to do things because they were largely the the inspiration for us to start this kind of company, like a company that like just took whatever they did, just turned it upside down. So like we're gonna do exactly the contrary, you know, so like transparent fees, no team, no, no uh what what's the what's the word for that? Like Bait and Switch when you like signed the contract and just like, oh that team I told you about no people for the project. Yeah yeah they're booked. So like here's this uh entry level person is gonna work on this project like you know old technologies like super opaque things and stuff like that. So I totally relate to that. For me like speaking about your your profile right for me it looks like as if you had been preparing for this switch of this change of paradigm that we're experiencing right now in which the the the full stack developer is coming back but the developer cannot be strictly restricted right now to being uh to doing only development. Now they have to know about business they have to know about product they know about like they have to have some sort of interest about distribution as well right so if you're confined to doing like I'm only like a front end view engineer like this market's not for you. And like people at the intersection of technology business and product like you um I think that by and large are sort of like the new profiles that companies will look um look um look for these kind of people because we can build stuff we can promote it we can polish it we can talk to the clients we can talk to the market and stuff like that and we know business so we know even how to price it more or less we have some sort of interest in making it successful not only just you know laying down the code so um I don't know if that's if you share these with me or you have another another opinion topic.
SPEAKER_01I I
Systems Thinking As A Career Advantage
SPEAKER_01look at that as systems thinking so to me I've been thinking in the last few years like okay what is the thread that sort of ties together everything that I've been interested in and passionate about and the way that my brain works. And it's always been systems thinking. So out of college I studied chemistry and then I went to graduate school in Chicago and uh I studied physics and chemistry there and uh what I studied was how systems within the cell work. So I studied proteins and how they fold up um which is like a system itself but then there's also other systems that impinge on it. And I found that very interesting. But as a career it's very challenging here in the United States to be an academic. So that's how I ended up looking around for a lot of different jobs and then I ended up at at McKinse after graduate school. When I got there I it was a completely new thing to understand business and marketing and sales and finance and all these things that I thought were very opaque and complicated and difficult. And but I looked at them and I think maybe the way that McKinsey does training is they look at it as a system. Like okay your company is a system and you have income and expenses and here's how things come in and here's how they go and if you push on this side of the organization then this happens and how did this organization affect other organizations? So I've always thought of those things through that lens. And then even as I became a developer I taught myself in 2011 out of uh books out of the pickaxe the big thick pickaxe book and another book called Agile Development and Ruby on Rails and what was the third one? Chris Pines Learned to program. Those three books basically taught me how to write Rails but I again I looked at it as a system. Okay you have a system you have a front end and you have a login and where does that go and how does that touch the database and how does that touch um the server and how do all these systems fit together and so I think that in the I think that it for developers nowadays in the age of AI as you say where the skills that were valued for a very long time are now changing certainly in value. And what you're expected to do is generally higher level work. Higher meaning like zoomed out not smarter or anything like that. It's just you your focus is a bit bigger. I think that those I think that developers have to use their already uh probably develop system thinking skills and look at things in a systemic way. Instead of looking just at the square of the program right these files and and this server and um these responses and these logs like you have to look a little bit outside of that and say like okay how do we how does our customer find our site how do um marketing funnels work how do our costs actually work how do we keep our server costs down how do we make sure that we're making a product that's good and I think if you look at things in from a systems point of view I think things tend to make a little bit more sense.
SPEAKER_00Actually how would you and then that that that explanation What do you think about that? Let me ask you like uh your your perspective on that that's good because it buys me a little like five more minutes to kind of like prefer the next question because you raise a lot of very interesting points I just want to make sure that we go through each one of them.
AI Changes The Shape Of Projects
SPEAKER_00Um for us we're seeing we're seeing a transformation I think like contrary to or we have a contrarian point of view to what the market says right the market says that we are doomed that we're gonna be fucked Skynet's coming developers will have to kind of like take janitor jobs or something like that because we're out of the market and stuff like that. And but we are seeing that since January this year we had more and better deal flow than ever. Right we signed three large contracts which usually a large contract for us is over a hundred grand per year and and that's something that we did like once per year for that this year we've signed already three and we've got like three or four more in the pipelines like what is going on like something's coming so for starters I think that we're seeing that people are being able to test their business hypothesis in the company then they outsource to make that production ready. So yeah yeah for us but I was gonna ask how are the how are the projects different than they were before yeah before like up until end of 2025 it was like somebody the big corporate would come to us and say like oh look I I raised like 30 to 50 grand to to build an MVP and these are the specs and you guys build it so we gave it back and maybe they didn't launch it maybe they launched it but that was not what they needed and stuff like that. So the project might or might not float or take off right now what they do is like they vive code it internally they test it with all of their production data then they go to their boss and instead of saying like I need 30 grand or 50 grand for an MVP they said like I build this look at this pretty amazing B BI tool and it's working but it's not production ready. And they said how much do you know it's like I don't know I'm gonna ask probably 150 grand 200 something like that. And they're like yeah fine because this tool will have an impact in the company of millions of dollars per year right or tens of millions of dollars.
SPEAKER_01So they come to us they're like look we got this you guys make it production ready how much is it going to take like six months eight months something so uh I don't know it can go up to like a quarter million or something like that for the kind of project it's like a risk risk mitigation tool in a way right like we used to build those MVPs way back in the day but they were expensive to build and we tried to push people to do them as cheaply as possible and say look you just need something to test and you need something to look at and to use your tests and to show people and to see if it works.
SPEAKER_00And now that's very cheap to do which is great in a way they're doing part of the job because they're validating it themselves and they can spend on the iteration because it's like they can iterate for free as many times as they want. Whereas right before when they outsourced this MVP if that MVP didn't do exactly what they wanted they didn't have more budget to build another one right so they had one shot. Whereas now they have infinite shots exactly so I think that that changed. And for us what we're seeing is like the transformation of the profile as well we're just hiring AI native developers right now which might not be the most junior people they're not the super senior people either they are people who are genuinely interested in working in this way they have to know how to work without AI but they have to have fully embraced working with plot code with RPI so some of these methodologies because otherwise they don't have a place in the market right so they have to be super positive super embracing super open-minded and that this is going to be the way to work from now on like you can't go back to manual. I mean I know that in you in the US you have been driving automatic since like the dawn of time but uh at least here we switch over to automatic transmission like uh some years ago and like going back to manual for me is like wow that's from like the Stone Age or something like that. So it's the same for development nowadays and the same for the project managers. They're not project managers anymore they're product managers because they have more understanding of the business and they have more understanding of the code whereas pure project managers back in the days like oh give me a time a project and a budget and I have to figure it out right but they didn't go into the like the trenches of like the code and the psychology behind the UIUX and stuff like that. So I think the profiles are more complete everybody knows a little bit more of everything. They're more like renaissance profiles but I don't know if we'll have to go back to the old ways. What do you think?
SPEAKER_01Yeah it's
AI Pricing And The Next Reinvention
SPEAKER_01hard it's hard to know especially because first of all I think that the acceleration of all of this uh really has taken you know a tick up and certainly in the last six months uh nine months um and the other thing that I think is very uncertain which is important for us to be aware of and I think it's very healthy to have uh people who know how to do things without AI is that the pricing is going to be very dynamic, right? Like this is an extremely expensive asset intensive, energy intensive compute operation. And it's been very heavily subsidized for a long time. That's not you know some hot take that I'm providing that's just basic financial analysis. And at some point the price has to go up um you have to start paying for all those services. So what is it you my question is like what does it look like when that happens? When the price so the price of developing the code uh through AI is maybe here right now and doing it by human is here. At some point how do these prices you know dynamically change over time and how does that affect how our work is done? That to me is the fascinating part of all of what's happening right now.
SPEAKER_00There is a part of it for me will be like a like the boiling frog situation in which like it will slowly go up and because there will be no much there's the price sensitivity for AI services is like we there's still a lot of available room for price increases right but a few months back everybody was on either the free plan or the 20 bucks a month plan. Now everybody in the company is 200 uh the 200 bucks a month but I'm pretty sure that in six months from now everybody will be on a plan of a thousand a thousand or two thousand bucks a month and it will still be worth it because we're extracting so much value out of it right so we can pay it yeah not everybody else can pay it in the industry especially outside of development roles but for a developer if I can extract 10 15 grand per developer per month and then I'm paying 2000 I think it's worth it right then you have to subtract you know I'll factor in the salary uh structural uh costs and stuff like that but I think there's still a lot of uh elbow room for price increase however um I don't know if that will be also like the the cloud that you know which is accepted that everything has to be on AI like everything has to be on the cloud and we stopped conceiving scenarios where things are not in the cloud. How many yeah what is what percentage of on-prem projects have you worked on in the last 15 years?
SPEAKER_01If anything I maybe one so yeah if anything what I did was move things off on-prem right like yeah we have a you know we have this thing with the server in the closet down the hallway and we it's it could flood and we need to to move it off there. You know I think it's I think in ways it's similar um to the cloud transformation but in a lot of ways I feel like it's very different because I think the transition to like having your servers in your office versus having somebody manage your servers is very straightforward and linear and the costs are very clear, right? The benefit is very clear, the costs are very clear. And so that transition from you know AWS and Azure and all those cloud transformations feels very straightforward. Whereas like this transformation the costs feel very unclear and the dynamics of it feel like they're changing on a day-to-day basis, which is why I think it's such an interesting time. And I think it is a time for everybody to be at least prepared for a reinvention. And I'm thinking I was thinking this morning you know about what it means to be able to be flexible and be able to reinvent yourself. And I thought a couple things one I think you have to really be honest and authentic with yourself. You have to say okay here's here's where the situation is here's things that I'm good at as I said right here's things that I'm good at here's things that I enjoy doing here's things that are valuable and the world is changing my local economy is changing my home is changing whatever's changing you always have to be adapting and I think you have to be honest with yourself about all of that and to be on your own authentic self. You can't really pretend to be someone else um and I the other piece of this I think is to take a beginner's mind which I think we talked about this a few months ago on the phone I think that you guys have taken a really good approach to this to to look at it as a beginner's mind. Okay what can uh AI do for us? Why don't we all learn it, right? I know that you asked everybody in the company to start experimenting with it and then to start talking about what you learned and what works and what doesn't work right like that uh apprentice mind that beginner's mind the learner's mind I think is very important and to to always kind of be a perpetual student. And I think the third thing that's important is to talk to experts is that when you know when you're starting in something new or you're trying to move into a new area you have to go and talk to people much like you sent 500 emails out right when starting a consultancy right no I I think that's that's the way to do it right you go and you talk to people and you say well how did you do it my dad used to say um you know there's all this dad wisdom right which I you realize that some of it you pass on to your kids he would always say whenever you make a big decision you should ask five friends and that always stuck with me and I watched him do it and over the years I've done some similar version of that which is like if you don't know someone else does and instead of beating your head against the wall doing all the primary research going to figure it all out in your own if you talk to five people you'll get a picture of what reality is from these people who already know and have spent time getting smart on this subject right or becoming an expert on some subject.
Fixed Bid Versus Time And Materials
SPEAKER_00And I think those are the to me the important pieces of how to make sure that you're flexible and dynamic and you're able to reinvent yourself there's this other uh point that is very interesting and I wanted to get your opinion on this because Thoughtbot were super opinionated on never taking fixed speed projects right um we really drew inspiration from that in Spain traditionally has been a country of fixed with projects we were perhaps some of the first companies to um to normalize the uh time and materials or retainer kind of contract was not popular here last maybe so 12 years ago nobody would accept that 10 years ago so so five years ago sort of became the thing right and so for us it was good because it was a great way to capture the value minus the friction of you know the like discussion dragging for too long definitely said this I said that like um like the the constant iterations on oh that was out of scope and and and this kind of paperwork that amounted to like a lot of like you know stupid gibberish didn't bring any value and and eventually you would never work again for that client right whereas with type of materials we saw less friction but now with the advent of this a AI gentic development however you want to call it I think the tables have turned again because for instance a if I give you a developer like I assign a developer to a project for 40 hours a week and now we know we have calculated that we produce 3x that so I'm giving you the value of 120 hours a week at the cost of 40 we're not capturing this delta right and so in a sense and I know that it's not exactly percent right or or precise we feel like time and materials is not worth it anymore because I'm giving you way way more value than what I'm actually uh getting back in terms of money. Oh that's interesting yeah with the fixed bit projects whereas before they were like a surefire way of losing money right now it's like oh we signed for this 150k project we know we can do it for 75 and I'm gonna give you a little bit extra I'm gonna be working until 110 but I will have that margin of 40k that will be my margin like a margin of like 30% whatever you you want to set it um to so we're going back to fixed feed projects because I think like design would be either much more interesting. However time and materials and retainer they give you the stability of working for big contracts and knowing that your team will be assigned to a client for a long time. Right? So healthy makes up both of them. So what's your take nowadays?
SPEAKER_01Well I think okay so just like some definitional terms um you know so time and materials is like okay we we had this many people working on this project for this week or that month and this is what it costs and here's the bill for that right and um fixed bid I would have this conversation with clients all the time and I would explain to them that fixed bid meant fixed cost fixed scope and that the fixed scope of it was the thing that I tried to hammer into them like that can't work. Like the reason why is just like any project even like for a plumbing project you open up the wall to replace the pipe and you're gonna find something in there that's either more work or less or what have you like you go through a discovery process and your scope will change. Like that's the important piece I think that is uh worth agreeing on with your client and saying okay the scope is going to change um the amount of people we have is going to be pretty fixed for now, right? And we're gonna have some defined costs. But the we have to let the scope vary and there's a system around how the scope is going to vary. We're gonna have user testing we're gonna have uh sprint planning meetings and things like that and we're gonna adjust and we're gonna meet with our customers and say oh we built this thing but now we need to build that thing and so forth. So that's really where the risk has always been hidden. And I think it's always been fair my customers have always come to me and said well look we have this budget right we have 100K for this project. I understand that you don't want to do a fixed bid project but like this is as much money as we have to spend. So the cost is fixed. And I think that I've heard uh consultants call this the iron triangle that one side has to move. Yeah so the the cost has to stay fixed because that's how budgeting works at large companies right it takes a there's a budgeting cycle and you say okay we have this much money for this project and we're gonna you're allocated for it and it can't really go up uh without trouble really so you say okay we have this much money we have this much of a team and then this is how much can we get done in the middle of that? And what doesn't surprise me about what you're experiencing is that the scope risk is I would say probably lower than it used to be because of the thing that you said earlier, right? That people are able to vibe code and build prototypes much cheapy much more cheaply and then test those. So which then cuts a lot of that scope risk right I wonder if you've been seeing that um as much churn and change in this overall scope of a project from the beginning to the end you know 10 years ago versus today. Like I would bet that more often now people because they've done this prototyping on their own very cheaply they're able to say look I know what I want to build. And if you know pretty detailed what you actually want to build estimating the cost is a lot easier. Estimating and then and then nailing that project and getting it on time and under budget seems to me like it would be easier. Do you find that to be the case?
SPEAKER_00Yeah. However they tend to overbuild That's one of the problems that we're seeing with these kind of projects. It's like because if the code the cost of creation of code has gone, that's plummeted to virtually zero. So they just built like nonsensical things, some of them like humongous mastodontic platforms, and with things are replicated in in three different screens just because it's cheap and like they didn't bother to check it's like oh no, yeah, we'll fix that later, right? Because but you get the idea. And there's like there's so many things that when you actually cost it up, they're like, oh, that's actually very expensive. So like, yeah, because the platform is really fucking big, right? Right. So um there's a job at I don't know what the word like the exact word for that is like chiseling away when you can like with the hammer, yeah, you take away the the the the the uh the the the the rock from like a satchel and stuff like that. So I'm saying like yeah, we have to trim it down, we have to to scope it much better because there's too much fluff over here, and you don't need all of that. And uh so you know the scope creep, um it's still there, it's just in another disguise, it's in another flavor, so to speak.
SPEAKER_02Yeah, but they deal with it.
SPEAKER_01Yeah, it is interesting. Um, yeah, it it is scope. My my uh weapon against scope creep has always been let's show it to a customer. And that always gets everybody grounded and based in reality. So I will say, okay, that's fine. Why don't we prototype this and show it to three or four or five customers and see how they respond to it? And then right away, it's not my opinion or your opinion. Like I think this is right or you think this is right, or we disagree, or we agree. It's like the person paying you money for this likes this feature and doesn't like that feature. You can't really argue with that, right? Like, and that's ultimately what you're doing. You're building something for your customer or your user. And if you're not helping them do what they want to get done, if you're not creating a good experience for them, you're not helping them out with your software, you know, there's a lot of other choices out there.
What Still Matters Beyond The Tools
SPEAKER_00How are you keeping up with all of these chains? Let's uh let's switch gears for a moment here and uh talk a little bit more about you and your the way you're transforming yourself. We're referring to this a little bit earlier in the conversation, but nowadays, with the last six, nine, twelve months of something like that, how are you keeping up with all the advances in different methodologies like testing fables and opuses and that codex versus cloth and stuff like that? How are you keeping up with all of that?
SPEAKER_01So, to be honest, I'm I'm not very detailed in it. And I and I I because of the decision years ago, that as much as I again I enjoy writing code, that it is not the thing that distinguishes me from other people in the industry. I have not invested as much in making sure that I'm the greatest developer out there. I certainly have downloaded and played around with these things and been impressed with uh certainly like what they're able to do. I also find sometimes some disappointment. Like there are, you know, it is a kind of 90% of the time product, I would say, 95 in my experience. Uh I give it a lot of different complex tasks. Uh I have my my wife is using it for you know developing PowerPoints and things like that. I think she's on um the Microsoft product. And uh my mom, even, you know, at 75 years old, she took an AI class and she's been, you know, she had a um air conditioner break, you know, uh heating air conditioning break in the house, got a bunch of quotes and fed them into AI and like got some analysis back on like, okay, this price is good and this is a good vendor and things like that. So like there's certainly utility out there, and I think it's important to uh experiment with it and be able to be conversant in it. But I think that the thing that will not go away is your ability to talk to somebody, your ability to understand what their business goals are, your ability to say, all right, well, let's look at how we can achieve those goals, scope the projects, um, you know, and make sure the other thing is like kind of team management, staffing, the human element of it is not going to go away. Like you can't AI your way out of um, you know, making sure that the right person is on the project, or after you put the right person on the project, something has changed and they need to expand their role or get more people or they need to leave the project. All of that, I think I find very interesting. And I look at it as like, well, we all have to be at work. And uh I look at it as like, how do I care for the people on my team and make sure that they're having a good experience, having good projects that are valuable, that are reasonable or not and have some balance, things like that. And so I I honestly have not been looking very deeply into it. Um, because I don't think I think that while the technical end of it is going to change, I think the rest of it is not. I've had it uh I've had AI generate some project proposals for me. And I think what I realize is like they're good first drafts, maybe, but I go through them and I'm like, this doesn't sound like me. Um it doesn't include the things that I want it to include. And sometimes it saves me time in terms of saying, like, okay, just get something started and then I'll edit it from there. Um but then after that, it's you know, it's sort of a mixed bag. I don't know.
SPEAKER_00For me, it's sort of like um I I see different kind of value, and and you'll probably you'll probably understand this, and yeah, yeah, I I don't know if you you will agree with this, but everybody has got the best developer and the wall in their computer. But the engineering part resides in he who writes the prompt, right? And being able to ask the right question, challenge the AI, saying, like, you suggested this, but uh, we're actually gonna go this other way. And like cherry-picking certain things here and there, I think that that mix, that chef-like approach of saying, no, we're gonna do it my own way, with my pace, my decisions, my calls. I think that's what will the this taste, right? This expertise, this know-how, it's still not lost, it's still not condensed into the AI. And so I think like you having created um so cylinder, that that boutique consultancy, or with us, we also have like the boutique uh uh etiquette, if you will. Um, we will retain our client base because they yeah, they precisely they trust our way of doing things, the our wisdom, our accumulated background. And uh, we've seen this in other projects. We know this is the right thing to do, right? So I don't know what's your take on this and whether you you agree with this or not.
SPEAKER_01Yeah, I think so. I think you can only ask AI so many questions before, you know, and and you might see this from your clients, where people are coming to you with leads and they're saying, Well, the AI said this. What about you? Do you agree or do you not agree? Like sometimes it tells uh it says things that are that don't make any sense, right? Um and and sometimes it does, and sometimes it's directionally correct. But I think ultimately um it is it is incumbent upon us to ensure that the human-to-human connection is and stays the most valuable part of all organizational work, right? Be it companies, nonprofits, uh neighborhoods, friendships, all of it. Like that's never gonna go away. And I think in many ways, that will become more and more valuable as this other stuff becomes more uh voluminous and cheap, right? Cheap in the cost sense. Um and so I think that's really I think that the human piece of this is is gonna only become more and more important. I think the other thing that's worth mentioning, and I know that we don't really talk about it too much, is the moral element of this is like pretty gray, right? Like the the data center stuff, all of the energy usage, the environmental impact, all the training data is not always, you know, uh acquired in the most moral way. And like so to me, I just I view I view a lot of that as very uncertain. Uh it creates a lot of uncertainty. And I think one of the things when people say like, oh, AI is here and now, uh the future is gonna look like this. Maybe, maybe not. I don't know. I mean, I I think that what's the future looks like is is more unclear now than it have ever has been in the future. So I personally feel very reluctant to say, okay, well, we need to rebuild our our lives around this because this is the way it's gonna be forever. Because it could be we could get model collapse, right? But the training data that they're that they're training on is more AI stuff, it just sort of starts to circle the drain a little bit, right? Um so I don't know. I guess I I find myself on the skeptical end, I guess, of all of this. Um I do use it, um, but I I I am not unaware of the systems piece of it, right? That uh how do how do all of these things affect the systems around us?
SPEAKER_00The the non-technical ones, right? Going back to the beginning of the conversation where you were mentioning that you it's really interesting for like the the systems part of uh of engineering. Um,
Biggest Mistakes And Best Decisions
SPEAKER_00a couple questions to wrap it up because we're running out of time. We're actually run out of time already, but uh I'm pretty sure we can we can uh squeeze in these two quick questions. Signature first one signature question of the podcast. What has been your most expensive technical fuck up? You have to own it up, it's gotta be yours, and it cannot be like I hired the wrong person because like everybody gives me that bullshit.
SPEAKER_01Um I dropped a database once. Um early on. I I dropped it completely, and um we lost this was Heroku, and I I definitely was learning how to use the CLI, and I thought I had dropped the local or the uh the staging database, and I did not. I dropped the production database. Um that was pretty, you know. I I remember I remember the feeling. I remember your face turns red. I was sweating, I was so ashamed and embarrassed, and uh I was so upset, and I just was like, I need to hide this. But then I realized you can't hide this, right? Like immediately everybody knows that you've dropped the production database. Uh luckily we had the you know a replica, so you didn't lose that much. But and and the other luck that I had was I've always had a good, I've always had good luck in surrounding myself with good teammates and mentors and teachers. And so I remember a senior developer was like, everybody does this. Everybody does this once, and um, it's okay, and now you know, right? And here's ways in which you can prevent that from happening. Um, so yeah, I I don't know that that's a very exciting answer, but it definitely was. I mean, right when you said that, I was back in that moment where where it happened, and I was just I remember sweating profusely right away.
SPEAKER_00Do you ever drop a database? Uh I I always struggle to think about like a I th I think I don't have like a big major fuck-up. I think I've got like a consistent record of medium-sized fuck-ups in the in the projects I've been working in. But like not a big one. Like Yeah, consistency is key, you know, like uh that that that brings your that that builds your curriculum. But uh I the I don't recall any like big one. Maybe I was never in charge of like too big of a thing. Um or the projects where I was working at the like were 30 people, so like not it, they didn't depend on a single individual. So luckily for me, it was like that. Then at Mars Visa, I think like the most like the most technical fuck-up. Like I have a lot, like a lot of business fuck-ups for sure. CCing the right people, sending the wrong contract uh to the to the wrong person, stuff like that, but like technical fuck-ups because I've been too away from the code for such a long time that I don't even remember because I Marsbase is 12 years old. I haven't been coding for this amount of time. Like now I've gotten back to coding, so to speak, with uh with Claude, uh Claude. Um, but before that, let me ask you my previous company. I hadn't been coding that much, so I was more like doing architecture stuff. So yeah, go ahead.
SPEAKER_01Let me ask you the more optimistic end of that, right? Like, so what is the one kind of single isolated thing that you've ever decided or done that has been the smartest thing that you've ever done instead of the dumbest fuck up? Like, what is the one moment where you made a choice where you were like, I'm not sure about this, but it turned out to be one of the smartest things you've ever done.
SPEAKER_00I have one. Probably, probably uh going hard and and taking the leap and believing in Marspace, like at the beginning, we're like, you know, we incorporate the company in Spain, you gotta chip in 3,000 euros, right? So roughly 3,000 dollars. Um, when you create the company and you don't pay incorporation costs, like it's a small sum, then here to the notary, here to the lawyer, and stuff like that. But we said, this is the amount of money we're gonna invest in the company. So by the end of this 3,000 euros, so you know, maybe we have like six months of uh of uh runway. I said if we don't have any project, we'll just shut it down and go back to finding another job. Or actually, I had quit my job, Jordi and Chavi hadn't. And so I said, like, look, I cannot do this while holding another job. So I decided to say, like, fuck it. I'm gonna go full-time, I'm gonna, I'm gonna be in investing in this. And I decided to take the leap and uh went full-time, and I even told them, like, look, uh, let me just earn a little bit of salary from this, just pay my living expenses, and that's it, and I will give you the projects, and it worked. So I I, you know, I'm not saying I'm not taking the credit, but I'm saying that because I really believed in this, the first project I was able to sign was like 20 grand, which at the time for us is like it was a salary of a person in Barcelona for a year. 20 grand, you know, and then the second one was 40, and then the third one was 90k, and so it worked. So for me, probably that was one of the riskiest but most successful decisions I've ever made. How about you?
SPEAKER_01That's awesome. Mine was uh I fired a client, I had a client, um, which I won't name, but is a very, very big name. And I was very excited to close this client. I worked hard, they were local. I met the founder, it's like a kind of single-ish founder, main one main person kind of company. And um we outlined the project, and they were used to having developers in India. And um then, you know, we had I'd scoped the project and signed the contract and everything. And almost immediately I realized that they were verbally abusive to my team. And, you know, these are grown-ups who are like can handle it, uh, but didn't like it, obviously. And they would start telling me, hey, here's some comments that I'm noticing, and I had this call, and it was really unprofessional, and all of that. And so I mean I gave the client a warning and I said, This isn't gonna fly. Um, this isn't how we do business, and this isn't professional, and this isn't worth it, right? Uh, so it needs to stop, or we're done. And it didn't stop, and so I said, We're done. And boy, was he mad. It's like super mad. They're not used to that, yeah. No. Um, but what I realized is like the people that are I had on that team that I'd staffed and also so the project with, they weren't ready to make that choice. Um, because they said, well, this is a huge name, and there's money and it's a project and all this stuff. Like, why are we throwing all this money away? And then over the years, we've we've talked about it more and more, and they were like, that's one of the best decisions I could have made, we could have made as a group, and I think it also set the tone to say, you know, we can be choosy about who we work with, and I think it is important to be selective about who you surround yourself with, including your clients.
SPEAKER_00Um, so yeah, that's a very good learning. And I have to say that I share that we've done that two or three times twice with like a really big client, uh so with two different uh clients, but like big corporations, and I think that's one of the top reasons why our employees don't leave the company. It's because they know we we do that. Like if there's ever a client that is a bully, or that they are super aggressive, or they're like super pushy and stuff like that, and unreasonably pushy, not just saying like, oh, we gotta make this deadline. It's like um trying to call your region like outside of office hours, writing you like private emails and stuff like private messages on Slack. And like super.
SPEAKER_01I think there's a difference, there's a limit. There's a difference between assertive, assertive, and aggressive.
SPEAKER_00Correct. Correct. And the couple of times that where that has happened, we have dropped the client. And so our employees, they know that we can do that, that we have done that in the past. If that ever happens again, we'll do it again. Like, no matter how big the company is. Uh, and in these two cases, they were really big corporations, and we're like, fuck it, it's not worth it. Uh, it's more expensive to lose talent. Like, clients have so many out there. And normally we've got like a one or two clients lined up, so we can drop someone and at the drop of a hat, just you know, onboard the next client. So I know that's right. The optionality of having these kind of super opinionated and boutique companies, not everybody can have this privilege, but we do. So there's that.
SPEAKER_01Well, I I think it's self-referential, like I think it's self-reinforcing that if you have that level of um opinion where you say, no, we want to do this kind of work with these kind of people and we want to communicate in this way, I think that you attract more of those kinds of clients. And the cost, I think the estimated cost of replacing an employee can be as much as two to three times their annual salary. So just from a business decision perspective, I think it makes sense to do that.
SPEAKER_00Perfect.
Intros, Email, And Staying Connected
SPEAKER_00Last question, Adarsh, how can we help you and Cylinder or your the other projects you are involved with? Because I know you're involved in three or four other things. So let us know. I'm rolling the red carpet for you, my friend.
SPEAKER_01Well, I appreciate that. This is very helpful. Thank you. Um, I just, you know, I just have always appreciated our friendship and our conversation. So um just keep in touch. You know, let's get on the phone uh once a month, every other month, and chat and and catch up. I love hearing about uh how your business has evolved over the years. I know that initially this started off where you were learning from me, but I've learned a lot from you as well. We've learned from each other over the years. And I think that that's you know, that's the beauty of all of this work and and life in a lot of ways. So um, yeah, that's certainly one piece of it. And then um I've introduced you to people and clients and things like that, but I always love introductions. When people say, people who know me well will say, you know what, you need to meet someone. And they introduce me to someone really phenomenal who I can, not just for business reasons or professional reasons, but just, you know, somebody who's just an interesting, funny, amusing, smart, insightful person. That's always like the most valuable thing to me, is to have those people in my life. How can I help you?
SPEAKER_00Oh, just getting on the pod every now and then as well. Just having you some sort of like uh resident uh fellow podcaster or podcast co-host. Like it's uh I we're loving these kind of conversations, they're super mentally stimulating. And and of course, just uh if we ever have the opportunity of working together again, uh really much love it. So uh staying touched, maybe sending some people your way. Um and how about the other people, our community? How can they help you? How can they reach out and and send stuff your way?
SPEAKER_01Um, I guess I, you know, I enjoy getting email. Uh I think email is the simplest way to interact. Um, I uh my email is hello at cylinder.org. And that's my public inbox. And I like getting random emails from people. I read them all, and uh I have a clean inbox. Unlike many people, I don't have tens of thousands of emails that are unread. I actually keep it pretty clean. Um in terms of socials, I'm active on uh Ruby Social. Um and yeah, I'd love to, I'm always interested in meeting new folks.
SPEAKER_00With that, we can wrap it up. So thank you very much for being here. Love talking to you and uh have a great week. Thanks. Thanks for having me.