If everyone can build software, what decides who survives?

Nune: Welcome to Root Cause. I'm Nunem. This is where we get to the root cause of things. Architectural decisions that age badly, leadership transitions nobody prepares you for, and the hype cycles that keep repeating themselves. If you have ever thought, am I crazy for questioning this, you are in the right place. The market today is flooded with tools, and most of them have AI in it. And we all know most of it won't survive. So that's the question I want to root cause today. If everyone can build now, what actually decides who survives? My guest today named culture, and I want to understand what that word means and how to build one that brings your company success. What stood out to me when we had the discovery call is that he's genuine and passionate about code quality, testing, future of software engineering, and people, most importantly people who work on all these topics. Besides, he promised to recommend a killer book, and I asked him not to tell me so we can reveal it today on the podcast. Let me give him a proper intro. My guest today is John Gallagher. Twenty years in software and not twenty years of the same year repeated. His career started in nineteen ninety eight, when I was nine years old, in embedded systems and programming microcontrollers in assembly. Then iOS, then more than a decade of Ruby on Rails in production, at companies whose systems real people depend on. Today he's a principal engineer at Dynatrace, pouring his knowledge in one of the biggest observability platforms in the industry. Here's the thing he's known for. At one of his previous companies, he lived through the kind of incident we've all had. The app is down, the people you could that could help are asleep in another time zone, and the logs tell you nothing. Instead of accepting that this is normal, he spent the next 18 months changing how the whole team showed the their own system. By his own account, downtime dropped 98%. And then, because all good things come with an invoice, He turned around and cut the observability bill that this new enthusiasm had created. He's also the author of Software Design Simplified, a book for engineers who want to get better at design without wading through theory, and he coaches engineers on exactly these topics that we're covering today. John, welcome to Root Cause and I'm glad that we're doing this on camera this time.
John Gallagher: Thank you Nune pleasure to be here.
Nune: so as I as I sort of introduced in the beginning, wherever I look, there are companies popping up now, as and also in the AISRE space. do you think the AISRE space is now flooded because AI has made it simpler to build a product or has it started even earlier with open telemetry and sort of democratization of telemetry and observability?
John Gallagher: Good question to start out with. Yeah, I I got into observability at quite an interesting time when open telemetry was maybe four years old, three or four years old. Most people hadn't heard of it. And so I've seen almost like the before and after of O-Tel. And before was you have five big companies essentially that rule. the observability space. They hoover up all of the revenue. And if you're using one of these five, you install their proprietary library, maybe some telemetry code that they give you. you push that telemetry code to production. And all of sudden, your data appears in this tool. The tool has a very custom UI, lots and lots of different screens, lots of different ways of looking at that data. And essentially, you're locked into that vendor. And if you want to change vendors, you need to rip out all the very proprietary telemetry for that vendor and then install. Go through the whole dance all over again. Go to a new vendor, figure out a contract with them, install their telemetry code, install their library, and then learn an entirely different user interface with an entirely different set of screens. And in amidst all of this, you need to figure out what traces, logs, and metrics are, which are the three pillars in inverted commas. So logs, traces and metrics and depending on the vendor that you went to you might have synthetics on top of that and then you might have this like maybe user experience monitoring or something and then that special user experience monitoring you need to install more custom telemetry for that which moves you further into the ecosystem of that vendor and so on and so on. And from that, we've gone to a world where you spin up your app, you do a year's development of it, you get it to market hopefully sooner than a year, you get it into customers' hands, you start getting some popularity, you want to add observability, and you install the OpenTelemetry SDK. And then the next question is, which vendor do I choose? I have free choice. you update a couple of environment variables and you're off to the races. And then the theory goes that you can then switch between vendors easily. It's more easy, of course. It's not easy because every vendor has a different user interface. Every vendor has a different way of looking at dashboards, et cetera, et cetera. So the user interface is... now kind of, and the tooling around that vendor is now kind of the lock-in, if you like. So if you've got 200 alerts all set up in one vendor and you switched from one of the vendor, unless you've done the infrastructure as code thing in Terraform, kind of you're starting all over again with the structure. I listened to an interesting interview quite a while back with somebody who said, yes, we switched to another vendor. We were instrumental with O-Tel. But actually, the instrumentation was very, simple, as you'd imagine. But 90 % of our switching costs came from all the other cruft that we'd accumulated with that other vendor. So I don't believe in the just switch it to another vendor with two lines and you're done kind of thing, especially if you're a larger company. But certainly, the choice has never been greater. And in the wake of OpenTelemetry gaining adoption and gaining maturity, We've started to see a whole bunch of startups pop up all over the map in the observability space, which I think is good for everybody. It's good for competition. It's good for customers. It's good for vendors, whether vendors want to admit it or not, because it puts pressure on them to make their products better.
Nune: interesting that you're mentioning big vendors and big players. it was previous the premise that like the small companies, the startups, they have a way to compete with big vendors because they are more agile, that they ship faster, they don't have the bureaucracy of a bigger vendor. And then now the premise is and and need to be careful here because those those promises haven't really materialized yet. That the bigger companies now also can move fast because of the AI acceleration of the development. and so if that is true, then the small companies don't have a chance to to compete with companies like Datadog, New Relic, Dynatrace. what's your take on that?
John Gallagher: On the surface of it, seems plausible, doesn't it? But from within DynTrace, I can tell you that AI is massively changing the velocity at which we move. And so, I see it having a humongous effect inside our company and we're a very big company. When I joined Dynatrace, I said to somebody, wow, it's amazing being in such a big company. It's kind of mind blowing to me because I came from a 12 person company, was my role before Dynatrace, and now I'm at a company with 2000 people. It's absolutely bonkers. The person turned to me and said, engineers. What do mean? said, no, no, 2000 engineers. We have 5,000 employees. OK. Two and a half times the size I thought it was. And still the number is meaningless, right? So if a company like ours is using AI to improve velocity and quality and all this good stuff, then it means there's a lot of other enterprises doing this. And yet, According to Conway, know that the structure of the organisation affects the structure of the artefacts that it produces. So in our case, the structure of our teams will dictate the structure of the product that we create. Not only that, but the number of communication paths... within even a seven, 10, 15 person company is exponentially lower than that of a company with 2000 engineers. me. So, so we're in a world I think where it's a speed up, but it's not the same kind of speed up. And I don't think that I don't really see it that a small company used to be lean and agile and nimble. Agile has all sorts of connotations now. Lean and nimble and quick. And it used to be a massive company moved really slowly. And now a massive company can move really fast. So the playing field is levelled. It doesn't really work like that, least from where I'm sitting. What it means is that... in a big company, you've got lots of communication paths and you have to have bureaucracy in large company. There's the odd large company that manages to get away without it, but it's a fact of life because how do you manage 2,000 engineers? What, just like lay down six markdown documents in a repo and say, have at it, everybody just do whatever you want. It's not going to work. Added to which the customers of a small startup are very, very different to the customers of a huge enterprise. Very often, a sales tool for a big enterprise is the fact that they're a big enterprise, is the fact that they're saying, look, we're not going away anytime soon. You can depend on us to take care of your project for the next 10 years. We're on the New York Stock Exchange. There's a stability. There's a sales tool in that that startup just doesn't have. And so it's kind of apples and oranges. They're just in two different worlds. They're in two different cultures. They make two different types of software. Obviously, a big company can take inspiration from small startups, but ultimately, Dynatrace and lots of other enterprises are not going to make the same kind of software that a small startup would make. And they're not going to move with the same velocity. And then it's like a different quality. not a different quality of software, I mean a different timbre, it's a different style, it's a different... It's a bit like saying, you know, is jazz better than prog rock? Well, prog rock has moved on, so prog rock is better now than jazz. Is jazz more creative than prog rock? maybe. But they're just two different genres, at least that's way I look at it. And those two different cultures will result in two very different... types of software being created. And there's anti-patterns in both kinds of companies. I would say all things being equal, a smaller company makes better software than a larger company. For individual people. For enterprises, arguably it's a bit of a different case, but it's much, much more difficult to make really good, really quality software when you've got a bigger team. Just because there's a lot more opinions, there's a lot more dynamics, there's an exponentially exploding number of communication paths, and this is where culture comes in.
Nune: Yeah, a lot a lot of times bureaucracy gets a negative color, but that's the only that's a natural outcome of structuring the communication and the processes around the in the com in a larger company. if you overdo it, of course, there then there's a new set of issues you need to solve, but it's not necessarily a bad thing.
John Gallagher: Absolutely, it's kind of... Like, I am very resistant to bureaucracy. I'll just say that now. I hate it. I've got no time for it. But it is a... It's not a good thing or a bad thing, much as my internal emotions rail against it. It is a fact of life in a big company. It's just how the cookie crumbles. It's choose your analogy. It's the nature of the beast. It's not good or bad. It just is. Now there is a book. I've not read it. I believe it's called Mavericks that talks about very big companies doing really unusual stuff and like stepping back with the bureaucracy. But it is just natural that if you have a team of a hundred people There needs to be some structure to that. So it's part of the system and if I really, really don't want to deal with bureaucracy, fine, go and work at a smaller company that's maybe got a three-year runway.
Nune: Right.
John Gallagher: So there's pros and cons, there's trade-offs to both. And obviously there's a lot in between as well, medium-sized companies and so on. But it's the nature of the beast, it's part of the game.
Nune: And again it's it's so much more than coding. A a product and a company is so much more than coding and even if the claims of AI has replaced the coding parts are true, as you said the the customers, the way people work with each other, a lot of a lot of other things are not affected by that and and so there's still the distinction of the bigger companies serving different kind of customers and the smaller companies building a different kind of software.
John Gallagher: Yeah, and there are certain kind of anti-patterns that persist across the large and big companies I've worked in. I should also be upfront and say a lot of my experiences are very myopic. I haven't worked at one of the big tech companies. I haven't worked at FAANG. I have worked at a fairly small number of companies, from very small companies to very big companies, because I want to stay at a company for at least three years in general, three to four years, so that I can... I mean, it takes five years for bad design decisions to come back to bite you, roughly speaking four to five years, and ideally I would like to stay long enough at a company. to have my own design ideas come back to bite me. Nowadays with AI, it's more like four months, right? But yeah, so that's why I've not really worked at many companies and I try and stay loyal. And that means I don't have this massive breadth of experiences and I haven't consulted for thousands of companies or anything like that. yeah, my experiences are very quite narrow, but with that as a caveat. one of the, there's quite a few anti-patterns that I've never really understood that persist, no matter whether it's small company, large company, whatever. And they're all kind of cultural patterns. And I think it's almost like the culture of our sector. And feel free to write in and disagree with me on that and say, you're absolutely wrong. You've just worked at. the wrong companies, whatever, but one of them is a resistance to talking to customers. There just seems to be this kind of, not just from developers, but from companies in general. It's very, very strange that we seem to have kind of a, and maybe it's across all industries actually, but there seems to be this kind of. We know what we should have in the product kind of a mentality. We know the roadmap. We know what we want to deliver. Right. But are you using the product? Is it for you? No. Are you investing like a five figure sum in this thing? No. But we know because we made it. That to me doesn't make a lot of sense. I mean, there's obviously things if you're working on a product, there's things every day. which you think, wish that was better. And they're things that are objectively bad. So one of the things that annoys me is modals interrupting people. Even that has been accepted now as somehow a best practice. I don't know how that happened, but yeah, but back to my point, I guess I'm constantly bemused by this. We know what. the customer wants, we know what they need, how do you know what they want? Well, and you keep on digging, you don't need to ask me any questions before you figure out, right, okay, you just think this is a good thing to have on your roadmap because you like the idea, good. Well, I'm glad that that'll hoover up lots of market share, I'm sure, if you like the idea. And so there's this kind of... And very often companies do a good kind of a sketch of talking to customers. They'll maybe have a trade show and talk to a few customers and then say, right, we'll go off into a cave. And that's the roadmap for the next year or something like that. But that's one of the things that constantly, and startups are a little bit better at this because their entire livelihood depends on customers liking the app. But even in the startup space, seems to be this Yeah, seems to be a resistance. We're going to launch in stealth. That is still a thing. Somebody said that to me the other day and I'm like, wow, has that idea really not died yet? Surely that should be buried. Another one, of course, is user experience. What does that mean? It means drawing pretty pictures now. good. Excellent. Right. So people buy things because of pretty pictures and they do in part. That is true. But they... You know, there's an entire discipline there to do with user interactions. That's what user experience is about. It's about the interactions. It's not about does the button have a three pixel drop shadow or a six pixel drop shadow? I'm not saying that's irrelevant, but that's another anti-pattern. So yeah, these anti-patterns seem to persist over long periods of time.
Nune: that's a very interesting topic you're touching and we have brought this up in the podcast again and again. And one point of view is that it's an ego thing. I mean, a lot of the developers, a lot of people in IT, they are very smart people and they are very proud of what they create and of what they have thought of. And often it becomes, you know, this kind of rock star attitude that I know better, this is my baby, this is my software. and and they they just they're very creative people oftentimes. Founders are and they create this product, this thing, and this is why they don't they don't even if they make it look like they are Getting opinion from a side, I I don't think they want it because they have this idea in their head and they are creating something out of nothing. so that could be an ego thing, for a lot of people.
John Gallagher: It almost feels, I think that's absolutely bang on, it almost feels a little bit like the bureaucracy topic that we were just talking about. It's almost like that ego has to go with a startup founder. You can't have a founder who's like, well, I'm so humble. I'm not sure my ideas should really be out there.
Nune: Mm.
John Gallagher: Let's take feedback from 100,000 people before I start. You just never get off the ground, right? You have to have that ego. There was some... I forget the exact research, it's maybe been debunked now, but I remember reading about this research. I love a podcast called Radio Lab, and they had this episode on lying. And they said that basically some research was done that all the top athletes in the world are excellent liars to themselves. They are completely delusional. you need that level of delusion to get to those heights. Basically, not one of them is there saying, well, you know, I may win Wimbledon, we'll just have to see. They're like, no, I am going to win Wimbledon. And then they come like 900th or something, you know, and don't even get into Wimbledon. But that level of delusion is a prerequisite. And in a way, that level of ego, that level of, I am going to create something here and it's gonna be... freaking amazing is necessary. I also heard that there are three attributes of every successful company founder. One of which is that they think they're better than everybody else. The second, and they deserve more, they deserve a bigger slice of the pie, they deserve to be paid more, they're smarter, they're more switched on. The second one is that they have a massive chip on their shoulder, something to prove. And then the third is that they have excellent impulse control. And so if you have all of those three, then it kind of makes sense that you're saying, I know better. I know what should go in this product. And yet it's the paradox, isn't it, that you have to have the ego. You have to have that drive. But I believe to make a really good software product, you almost then have to somehow say, right, OK, what do the customers want? Let's focus on that to the exclusivity of. really almost everything else.
Nune: did you mean that when you said that the culture will make difference in the future to defines whether a company is successful or not? Arguably AWS has said this, that they are working backwards from the customer and I think when you have a massive amount of customers it's sort of even easier because you have this feedback loop. i inside the company, hopefully, right? If you have set up the right processes. Do you think that's the main factor of the culture or is there more to this culture?
John Gallagher: a lot more. I wish it was that simple, just focus on the customer and everything else will be fine.
Nune: Yeah.
John Gallagher: I mean, a lot of things flow out of customer obsession. Of course, making a successful company, I'm going to say the quiet part out loud, there's a lot of luck involved, as I've discovered. And that customer focus is, of course, a big part of it. But an obsession with quality. is great, but I mean, I'm saying that it depends for me, it depends if you want to make a successful company or a successful product, a really, really good product or an excellent company. And sometimes the two go hand in hand, but this kind of myth that the best product wins, I think we're at a stage now where we need to really draw a line under that and put that one to bed because it really doesn't work. The best marketed product wins that solves the right customer pain at the right time. So in terms of a company being successful, there's lots of very successful companies that have a very toxic culture. But they get one or two things right. They've got a lot of timing on their side and they... have enough of whatever it is, whatever that X factor is, whether it's customer obsession, whether it's moving in the right circles and making the right contacts, whether it's excellent marketing, that they make it successfully. In terms of what goes into the culture of a company, there are a lot of factors to that. And now with AI, I think... Culture was always really, really important, but now it's just been upgraded by about, I don't know what you want to say, a factor of 10 or something. But certainly orders of magnitude more important because now that we have these tools and now that in theory a product manager can go in and create a prototype of the feature that they want, now then the question is what's to stop them pushing that directly to production? Maybe some companies do that, maybe some companies don't. yeah, what goes into culture back to your question? How people think of each other, how people operate, the mental model for which the company thinks about software is a huge one, I think. I'm trying not to leak details of the book I'm going to recommend, but very soon I'm going to have to start talking about it. So yeah, this idea that another idea that I think is like disproved maybe 20 years ago, but still somehow clings on like a flipping cockroach that just won't die is the idea that a company that creates software is like a factory. and you stamp out features just like a factory. We're making tin cans, we're making footballs, we're making TVs, we're making features. And we just churn out these features and that makes the product better. Please, please, please, can we get rid of this idea? That's another, yeah, so that's another big recipe in the culture is the mindset, how the mental model of what software is. And a lot of companies have worked at just fundamentally don't understand that it doesn't really work like that. I mean, if you have a factory churning out TVs, a TV manufacturer, name me one TV manufacturer in the world that could stop 80 % of the production of their models, like basically make 80 % of their models obsolete overnight and their sales shoot up. or customer satisfaction shoots up. If Samsung now, right now went in and deleted 80 % of their monitor range in their catalog, I mean, I can't even imagine what would happen, but it would be nothing good, right? In software, in a thought experiment, if we went into every piece of software and deleted the 80 % of features that people rarely use, the software would be better, undeniably better. Now, it may not result in commercial success, but it would result in better software. And there's a number of other problems with that mental model. So mental model, how people interact with each other, how we feel about our colleagues, the output of the work that we do every day is a reflection of the culture, how we interact with each other, how we get work done together or individually. a number of other things, but yeah, it's very socio-technical kind of a thing.
Nune: That's interesting. One one remark on the more features. I don't know if you know the the author of the wor book Rework. he had a post l recently on on LinkedIn where he said you do realize that more feature has never been the buying decision-making factor for people. it's it you can build a lot, a lot of features and still not solve the customer needs. So mm that that premise that if we push more code, more features, into the product and suddenly everything will be better is of course false.
John Gallagher: Yeah, Rework's a great book. Jason Freed and David Hanomai Hansen wrote Rework. And
Nune: thanks.
John Gallagher: I've followed both of them for like 15 years or so. I do not agree with them politically. I'll just put that out there right now. But if you just put politics to one side, if you can do. which is very difficult. A lot of the ideas that they come up with, think, are really... Either they have something in them worth discussing or they're really excellent ideas. I remember a story, and I think software could take a lot from this kind of mindset. When they were pitching their first book... to publishing house was their first book rework. I'll have to look it up. Whatever their first book was anyway, they went to the publishing house and they said, right, we're writing the book. Here's how the book is going to be. And they gave all these kind of demands almost like, and they said, the book will start with the first chapter. And so the publisher came back and said, you can't do that. And they said, why not? They said, well, you have to have acknowledgements, right? You have to have like a preface and you have to have a table of contents and you have to have, you know, like all the copyright and everything. it sends signals, you guys were just like, no, we're not doing that. The publishing house was like, but why? He said, when I pick up a book, I want to read the book. I don't want to read a bunch of whiffle waffle that's not the book. The acknowledgments, can go at the end. The copyright can go at the end. And that's always stuck with me. When I use software, just want to use the software. When I, when I, yeah, I don't want to be interrupted. I don't want to preface. I don't want to walk through. I don't want to tour. I don't want some weird widget popping up saying, how can we help? You can help
Nune: Yeah.
John Gallagher: by getting out the way and letting me use the software. I've paid for the software. I've just loaded up the software. Can you please let me use the software? And so that, in a way, they have this kind of very powerful style of thinking that is extremely simple. There is a business. The business sells a product to customers. The customers get value from the software. Therefore, they pay us money for it. And that is it. And there's all sorts of other. things now that software companies do that really is not that. And this kind of mindset about how people see software, what people see it as, lots of people see it as a way to get some short term money so that they can cash out and go to a beach. Some people see it as a life calling, some people see it as a mission, some people see it as, well this is just a vehicle to generate wealth. There's lots of different ways now of looking at software. When I started in software, the idea that you would build a $10 million company by selling software was utterly ludicrous. It's difficult to state. It's difficult to really... When I was 13, I was born in 1976. So when I was 13, messing around on my Amiga, doing early 3D ray tracing stuff. The idea that somebody would make thousands of pounds by selling some ray tracing software or hundreds of thousands or millions would have just been, but this is a floppy disk on the front of a magazine that you sell for five pounds. How could you ever in a million years make 10 million? And yet here we are. So this is what's happened is that we've seen that software is actually a really good way of making money. And with that realization comes a whole load of baggage.
Nune: It's very interesting that we're talking about this because I I was born sort of into the world that already had software. I had a computer, I don't know, when I was eight or nine years old, and to me I haven't felt this shift from product that is that you can touch into a product that is just software. So for me software was all also a hard thing to create, let's say. And a lot of people are now comparing what is happening with AI and accelerated development to those years of changing from creation of hard product into the creation of soft product, the software. do you see this analogy yourself? Do you think things are so revolutionary changing as they were back then? And and what are some things that are similar, what are some things that are different from your point of view, because you've seen sort of both
John Gallagher: Yeah, good question. Just I can't resist telling a couple of short anecdotes for those listeners who haven't experienced this. When I was messing around with my Amiga and what I was actually doing some of the time with my Amiga was very, very primitive. Well, I never really did programming, but it was primitive technical stuff. So one example is there was a game called Hired Guns. Hands up all in the audience who remembers Hired Guns. I see three people at the back. Anyway, so there was a game called
Nune: Ha ha ha.
John Gallagher: Hired Guns for the Amiga. And this game said on its covers, on all its recommended system requirements, you must have two megabytes of memory to run this game. That's a deal that was like over five floppy disks. And you put one floppy disk in, there'd be loading screen, and it would then say, insert disk four. And then you would pick your character, and then it would say, right, for this character, we need this little sector from disk six. So you put this six in, and then disk one, and disk two, and then the game loaded. That's how it worked. Because it was too big and too complex. to fit on one floppy disk. And so, of course, to load all of this stuff into memory, needed two megabytes of memory. our Amiga had one and a half megabytes of memory. I grew up in a very poor household. The idea that we could ever own an Amiga was just like traveling to a different universe. It was impossible. And still to this day, I have no idea how my dad did it. An Amiga is... roughly cost about a thousand pounds in today's money, maybe 11 hundred pounds. And my dad was on a cleaner's wage and my mum was not working. So how on earth they could club together and find me the money to do this is still a mystery to me, but they managed to do it. So the fact that we had an Amiga was just like absolute luxury, mind blown. The idea that we would then be able to pay more money to upgrade the memory. was like, no, that's just not going to happen. We'd already got a one megabyte memory card that expanded it to one and a half. So I was like, OK, this is a two megabyte game. I have one and a half megabytes of memory. And a part of me said, you're not going to be able to play this game. I got the game for free on the front of another Amiga format magazine. Again, hands up everybody who remembers that. I see one hand. Anyway, so we had these six discs and I was just obsessed with this game. Absolutely obsessed with it. And I'd seen all the screenshots. I'd read all the reviews. I read the reviews like 10 times, 15 times. I knew all the characters in the game. It got rave reviews across the board and I was obsessed with getting this thing to work. And so I went into the operating system was called Workbench. And so You loaded up Workbench, you put in disk one of hired guns, then you double clicked on the icon and the rest you know. And so I thought, maybe if I can reduce the memory footprint of Workbench, I can get this thing to work. So basically I spent hours and hours and hours removing individual bits of Workbench. So I copied Workbench onto a spare floppy to back it up. And there was maybe like, I don't know, 60 or 70 little files on the floppy disk of Workbench. And I figured out like the structure. So I want to say now, no manual, no manual, no internet, no book, nothing. You have the floppy disk and you have what was on the floppy disk and that was it. And so I went through all the folder structure and was like, well, that looks like it's just documentation and that looks like it's whatever. And I found a lib folder. And I was like, I wonder what lib means. Libraries, obviously. So all these library files. can't remember the extension name of it, it was like .xd state or something. So all these files. And I was like, hang on. Maybe these are the things taking up memory. So I did an experiment where I deleted these things one at a time, loaded work bet, and was like, well, does it still work? it does still work. And I saw the memory come down a little bit. I deleted one of them. Like, workbench does not work at all. So I wish I knew about a binary search back then, but hey-ho. So I went through each one of these files hours and hours and hours and And hiredgun still could not get it to work. And then I just gave up on the idea. And both my brothers were just like, just leave it, John. The game is not going to work. And then one Saturday, there was this bolt from the blue. And I cannot remember what it was, but I realised there's another way of lowering the memory, basically. This really weird esoteric work around that suddenly I was like, if I copy that thing and then remove that other thing and then put that other thing back in, maybe that's the key. I did it and Hyde Guns worked. And the intro music, I'll never forget when the intro music first leased up. And that was the thing that got me hooked on programming. Looking back, both my, I said, to both my brothers, I went through to the, I was in the dining room, they were in the front room, went through the front room and said, you will not believe what I've just got to work. They were like, hired guns. They came through and we played it and true to form, it's an amazing game, absolutely incredible game. And that's where I got my love of keyboard shortcuts from. But that's a far too long anecdote. Maybe you want to cut that down, but that's showing you that back then, And I remember learning 68k assembly language. In order to learn 68k assembly language, you bought one book, it was about this thick, and you read the book, and if it wasn't in the book, that was it. No internet, nobody saying, sir, can you help me? Can you help me with this? Can you? And so when I hear people, like I'm a big fan of people getting support for things, of course. But there's a real spoon feeding now culture that's grown up where if you read a Stack Overflow post and it doesn't just quite exactly meet the thing that you... What am I talking about Stack Overflow? AI answers. If it doesn't meet the exact thing that you quite needed, throw the toys out of the problem. this is a terrible answer. What's the AI doing? I hear people who are swearing at their AIs all the time. And I'm like, are you kidding me? I remember a time when I had to read a page of text and there was something missing between paragraph two and three and I had to figure out what that missing thing was. There was nobody to help me. There was no computer club. There was no online community. yeah. So from that world where there was shifting around, I remember shifting around individual bits and bytes. in PIC microcontrollers to this world now where we ask the AI to create an application and it creates an application, a very basic application. In React, which I will not go into my thoughts on React, but I'm not a fan, from that to this, from everything is physical to, no, you can sell things that are not physical, not touchable. to you can now sell things in software and now anybody can make software, it's a different type of revolution. Whether it will be as much of a revolution, I think personally the jury is out. There's people comparing it to the industrial revolution and there's definitely patterns that we can learn from from the industrial revolution. I watched another amazing conversation with the creator of Pi with Giorgio on the Pragmatic Engineer. Giorgio, I've mispronounced his name, sorry. I never know how to pronounce his name, but the Pragmatic Engineer podcast. And he was saying about how in the industrial revolution, every improvement in the bottleneck pushed the bottleneck upstream. and he was talking about really interesting examples with weaving. So there's a lot we can learn from the Industrial Revolution. Is it going to be that much of a change? I'm not sure. There is some people, there's an amazing post, a blog post by Charity Majors as well. So it's a name drop all these people, but this is really, really good stuff where she talks about... the AI skeptics and AI enthusiasts. And you're kind of in one camp or the other. The AI skeptics say it uses too much world resources, uses too much energy, blah, blah, blah, blah, blah. The enthusiasts say it's the next industrial revolution. This changes everything. I just read a post from Ed Zitron saying AI does not change everything. It's a toy at best, essentially. So there's all of these different perspectives. And I'm neither a skeptic nor an enthusiast. try and integrate both these perspectives into my worldview. So is it as big of a change? It's certainly seismic for our industry, for sure. One kind of cooling factor, I think, is that the AI tokens are being subsidized by 20, 30 times. The local models are getting pretty good, but they're still not an opus. So maybe it will all go local. If it doesn't go local, and we're still depending on paying for inference, we're going to have a big economic shock, I believe, in the next, who knows, two, three years. A company cannot keep losing 13 million pounds a day, can it? I mean, that's just absurd. So that is something that we need to kind of build into the equation of, it going to be as revolutionary? It could be now we're in a golden age of building stuff with AI and everybody can do everything because all the tokens are so cheap, it doesn't matter. And then the moment the tokens rise significantly, it then starts to become, it will, like we'll never go back obviously, but it could be more niche. So this is a long way of saying, I don't know. there are a lot of things that have changed forever. And so in that sense, you can never go back to a day where software didn't exist. And we can never go back to a day when AI didn't write the software, at least in part, for you. So that bell can't be unrung. So in a way, it is a seismic shift that can't be undone. So in that way, it's very, very similar. Another way in which it's similar is that all priors have been thrown away. I am constantly amazed at my brain's ability to not get with the program. It constantly is moving towards the AI to just speed up the old patterns. Example, the other day I was in, we're using GitHub to build our software and the team I'm in at Dantrace. And the other day I was in GitHub issues. And I found myself starting to manually rearrange the board. I'm like, what are you doing? Literally, what are you doing? You have the... How? What? So the most challenging thing for me, in a way, I'm kind of jealous of people who are growing up in this new AI world. And also I'm heartbroken for them as well because there's lots of headwinds they face. But one of the headwinds I face that they do not face... is I have to relearn everything. All my priors have gone out the window. I believe our friend Kent Beck said it best that out of all the skills that I have, 99.999 % of them have just dropped to zero in terms of their value. And there's a tiny, tiny subsection of my skills that have just increased by 1,000 times in terms of value. And so... The question is, what is that thing that I have that is more valuable than ever? I have a guess at that. But the most difficult thing for me is throwing away all my priors and rethinking everything from the ground up. Right now, I'm building software really as I always have done with the AI as an assistant. I mean, it's writing the code for me, of course. But I say, of course, two years ago, If I'd have come on a podcast and say, course it's writing the code, people would have been like, are you absolutely bonkers? Is there something wrong with you? Nobody does that. So we've got used to this new normal. Claude code has only been out for a year. If you can believe that, I still struggle to wrap my head around that. Probably a bit longer than a year now, but certainly not two years. So the parallels are all our assumptions. not all of our assumptions, but a lot of our assumptions being wiped clean. When the internet came along, when you could sell software, all of a sudden the idea of having a factory, don't need a factory. Okay, so we need a machine, right? Nope, don't need a machine. And there's lots of people who couldn't get their heads around that, which I believe was where the feature factory thing came in. So a lot of priors have been erased. That means therefore by definition there's going to be a lot of innovation because we have to make things up. And so that's another thing. When people started building websites, people were doing all sorts of weird and wacky stuff, experimenting. And out of that experimenting came Ajax and their Web 2.0 and the rest is, as they say, history.
Nune: I I loved your anecdote from your childhood, because even though I was I was born in nineteen eighty-nine, we we also as kids we hacked a lot, right? We were p little hackers, all of us, figuring things out, experimenting, there were not a lot of books, there was not a lot of information in the internet. And this is what I always think of when people say, Well, the human brain works like an LLM, it needs a lot of data. And then it learned something. But contrary to that, you used to do this one thing, you tried to optimize this one thing, and I bet a lot of what made you engineer then co came from this one experiment. You learned the patience, you learned how to learn, you learned how to experiment, how to verify, etc. etcetera etc. So there are a set of fundamental things. that a person needs to know and and this is what I think you also mean when you say that you are maybe a bit sad for this new generation that they don't go through that. and I and I want what I want to get Add is what are those fundamental things that a person needs to learn, in your opinion, independently of whether they are in the internet boom era or AI boom era, or in you know Plato's and Aristotle's era? Because I I also spent this last year or two, even though I I speak a lot against the hype. You as as as you said, the charity said, you need to learn these things to be able to argue about these things, right? With the knowledge. And so you go into you try things out, you you try every new LLM that's coming out. and eventually you circle back into yes, things have changed, but they haven't really changed. You still need to critical thinking, you still need debugging skills. And I can't help but wonder, should I have spent this year on some more fundamental thing? I don't know, learning resilience engineering or learning residuality theory or or things like that instead of tinkering with the tools. So the question to you, what do you think did we waste this year? A and B, what are the fundamental things we still need to learn absolutely?
John Gallagher: So I would say any learning of any kind is not a waste, even if it's not getting you towards your goal. Now it may be suboptimal, it may be not what you wish you'd spent your time doing, whatever. But if we're... And I speak this, say this as somebody who recently has spent a lot of time watching crap on my phone. If we're just watching crap on our phones and just consuming stuff versus learning, give me learning any day of the week, right? So you have been learning and you have been trying things out and this is the era of trying things out. So I would not say your time has been a waste at all. Definitely not. It's not for me to judge. how you've spent your time. That's not really any of my business. But because also by learning the specific tools, there will be maybe consciously or unconsciously cross-cutting things that you've learned. I'm not a big fan of going out and learning all the different tools and keep swapping between them all the time. That does have some value, but it's diminishing gains, obviously. But if you went and learned Claude and then learned Pi and then learned, you know, if that's what you mean by specific tools, learning all those specific tools, I think is an interesting learning experiment because you will have learned, I would have thought that all of those tools, there's all these differences, but there's a strong set of commonalities. And if you then think about those commonalities, that then gives you potentially the next stage of your learning journey. In terms of your question about the fundamentals, another idea that comes from 37signals, I'm really not a fanboy, I have to say, but they also have some incredibly smart ideas. It's difficult to separate the wheat from the chaff. Very early on when they created Basecamp, they got Jeff Bezos as a mentor. Now I'm sure if you try and get Jeff Bezos now as a mentor, good luck. If you've got $100 million in the bank and you can give him a tenth of that, maybe you might buy a minute of his time or something. But back then Amazon was a lot smaller. Anyway, so they got Jeff Bezos as a mentor, which again now seems utterly... absurd, but it wasn't that absurd back then. Yeah. So they got Jeff Bezos as a mentor and he said to them, they said to him, what's your advice? What should we do? We've got this feature, we've got this customer. And one of the pieces of advice he gave to them was invest in things that don't change. said, what do you mean? And he said, well, the customer is never going to want a slower bit of software, are they? They're never going to want support that is a bit worse. There are evergreen things that the customer will... always want. So just think no matter what technology, no matter what all the other things, no matter what the market says, what are the things that never change? And so very early on I started thinking about what are the things that don't change in software? And so I really didn't get that deep into the weeds with learning a specific technology, learning a specific framework, learning a specific data flow pipeline technology, blah, blah, blah, blah, blah. I more focused on testing, design, feedback loops are a big thing for me. and learning how to updating my knowledge, iterating. Critical thinking, as you've said, working against my own biases, which, hand on heart, I'm terrible at. It's really a huge weakness of mine. I succumb to group think, which is why I think the JavaScript folks have really, really lost the plot. That's a different thing. We might want to take that out. But yeah, so critical thinking design, both in terms of user interface design and code design and what even is design, getting a clear idea on that. Interaction with the end user, no matter whether AI is creating the product or even if AI is talking to our customers at one point, could be. It's not beyond the realms of possibility. We still need that feedback loop from customers. It's feedback loops all the way down. Great talk from Dave Thomas. saying Agile is dead. Who can link it in the show notes? And he said really agility was never like it's now capital A Agile, but it was a verb. It was agility. He said all agility is, is you take a step in one direction, you evaluate the step and then you improve and you take a different step. It's a feedback loop. The difficult thing is you have to have that feedback loop at every single level of the organization. At every, it's fractal. Every single level, everything that you do, from the smallest test all the way up to the biggest feature or the biggest commercial decision that company has to make, its feedback loops all the way down. So I got really very good at taking a small step, evaluating, then figuring out what I learned. So that is the ultimate skill. And I'm still surprised when I see engineers debugging. because it's clear to me a lot of engineers still can't really do that in a loop. Not really. So many times, this was before AI, right? But it's still relevant with AI. You still need to evaluate what the AI is giving back to you and say, is this right? Is this a good path? Or if you don't do that, you need to give the AI tools to improve its own workflow, which means you need to evaluate what the work. There's no getting away from validating what you've done. verifying the AI is doing what it should be doing, whether that's validating what software is doing, whatever. Well, so many times I'd pair with an engineer when AI was not a thing and I'd say, why is this not working? Let's figure it out. And immediately they would jump to a conclusion. Maybe, it's definitely that. It's definitely a race condition in Redis. Okay. And so I'd say, what makes you think that? Well, it just is, isn't it? They'd say something like that. It's obvious.
Nune: I guess I guessed right.
John Gallagher: I guessed right. Yeah, no, no, no, it's a good guess. Yeah, yeah, and I would like, right, cool. So let's explore that. Let's try and collect some evidence for and against that. So what if we ran this experiment? I wouldn't say it as an experiment. I'd be just like, well, let's run this command. And sometimes it would literally say to me, there's no point. I know it's a race condition. But how do you know it? And so we'd run this little experiment, and it would. give credence to their argument, they go, see, it's a race condition, let's fix Redis. Stop, stop, stop. And then you'd run another little experiment. You could see the impatience. It's like, just let's do some small loops. And then we do one loop, yes, it's Redis. Don't do another loop, yes, it's Redis. Number three,
Nune: Yeah.
John Gallagher: no, no, it's not, that doesn't support your argument. And by loop number three, they'd be locked in. No, it's definitely a race condition in Redis. It'd be like, loop number four. it's nothing to do with Redis. And maybe that's the point where they would say, yeah. And then you get to see how, back to your ego point, how do they handle being wrong? Some bluster and BS, some say, no, I was absolutely wrong. Some people still cling to the fiction that it's a race condition in Redis. So yeah, that's the point where you get to see. how many times somebody's been wrong and how open they are to being wrong. Another really interesting mindset shift that I've heard about, I wish I did more often is, okay, yeah, sure, I think I'm right. How could I be wrong? Make up all the reasons why you might be wrong and then, you know, disprove them or examine them at least.
Nune: The the more you talk, the more I again reflect how much mathematics and science has taught me. And like because it's all about experiments and all about proving yourself wrong. I mean y you take a theorem and then you walk around it, you try to prove it, you fail, you know your peer is gonna Ask you how you proved it and why you did the steps that you did, and basically everything that Barry said about walking the problem. But for people who didn't get science education or didn't maybe also have this experimentation as kids, is it too late to learn critical thinking? And how can a person learn a book trying out hands-on? What's your recommendation? How to be learnt that?
John Gallagher: Wow, I am not best placed to answer that question.
Nune: From your personal experience?
John Gallagher: I look back at all the things that caused me to try and get hired guns to work, there were certain ingredients that are still relevant today. So the first one is obsession. Pick something that you're freaking obsessed about. If it's just like, how do I get the flipping dynamic operator in Kubernetes to work on a three node cluster when I want four nodes? I don't know, maybe there's somebody who's obsessed about that, there probably is, but choose something that you're really passionate about, that you really want to get to work. It might be a small software idea, it's never been a better time to start building software, even if you're not a software developer. certainly don't read a book. That's the first thing.
Nune: Yeah.
John Gallagher: You don't need to really read a book. Reading a book about critical thinking is like reading a book on tennis. Just go out, get a racket and play some freaking tennis. And within the first 10 minutes, you'll understand a lot more about tennis than you ever did reading a book. Now, of course, books are great. But in this case, there are some simple ingredients. One is some kind of drive, some kind of obsession. Second of all is a small... feedback loop. And so there has to be some way of you checking whether what you've done is heading in a direction that you want. We can actually do, we can actually translate this kind of critical thinking in some ways to things that are outside science and mathematics as well. It's maybe not quite critical thinking, it's more creativity. But if you're really not into building software, anything technical, you can still apply the same ideas of feedback, maybe not critical thinking, certainly feedback, if you're making a painting. So instead of pulling out your paint set and saying, I've got a specific idea of this thing I want, and I'm going to paint it, and then I'm going to stand back and look at the painting and evaluate it. Maybe you want to do that, but maybe it's a good idea to start painting. I went to a painting class many years ago and I just started painting and the lady looked at me and she was like, are you okay? I'm like, yeah, fine, fine. I'm very repressed emotionally. She's like, you really don't look like you're fine. I'm like, no, I'm not. I'm really annoyed. She said, why? like, well, because this freaking painting isn't coming out as I want it to already. I'm only like... I've only spent 20 minutes on this. She said, you really don't look happy. I'm I'm not. She's
Nune: Yeah.
John Gallagher: like, is this going to continue? Do you think that's going to happen for the rest of the session? I'm like, yeah. She said, right, so you're here for fun, right? I'm like, yes. But this isn't fun. She said, well, why don't you change the painting that you're doing? So what could I do instead? I was painting an object. She said, well, what do you think would make you happy? I was like, actually, if I just focus on the brush strokes. She's like, there you go. Started doing that, and I made a painting that I never thought I would make. It's basically just like one massive gradient with loads and loads and loads of overlapping bits of watercolour. And I got it on my wall. I really love it. And I had a great time making it and that painting is a reminder to me. If you're hating what you're doing, that can be a feedback loop. Noticing your own emotions can be a feedback loop. But you need to catch yourself. There's no point in doing what I did the other night when I started a game of tennis. I started getting angry. I got angrier and angrier and angrier. And after an hour, I came off the court ready to smash something. My friend said, are you okay? I'm okay. Well, then last night we went to play together. And he saw me getting angry again. He was like, what could you do differently? I was like, OK. He said, why don't you try laughing at your own shots? OK. I'll try that. And it was a little experiment. And immediately, I started playing better. So it doesn't need to be in mathematics. It doesn't need to be in programming. It doesn't need to be in software. You can do that feedback. Now, in terms of critical thinking, you can't really think critically about a brushstroke that you've just done you either like or you don't like it. There's an element of critical thinking in that. One thing that I think might be quite fun to develop one's critical thinking is take something that you're absolutely obsessed with, that you love, that you think is the best thing since sliced bread, and find out all the ways in which it's bad. Go online, find people who hate it. If there's anybody who loves React, who wants to be disabused of that notion, I'm here to volunteer for that. How anybody who thinks Kubernetes is a best things since sliced bread come to me. I'll, yeah, I'll give you some counter counter arguments to that. But, yeah, find counter arguments to your position, read them, ingest them. the other thing, I think I've touched on it there, being aware of one's own emotions. I follow a guy called Joe Hudson on YouTube and he says emotions are the reason we do absolutely everything. Some people disagree with me on this, but the reason we have all this group think is because somebody somewhere wanted an easy answer to something. Somebody was like, this is cool. And they got excited. The thing that they grouped think about fit their use case at that time. They tried it. When they actually tried it, if they were really honest with themselves, it kind of did 80%, but not the other 20%. the emotion takes over. my god this is this is so amazing. And then they started waxing lyrical about it. They were maybe very effusive, they were maybe good at persuading other people. Other people were like why is it great? And they took five percent of the thing that they liked about it, wax lyrical and people are like that sounds really good. And then they started waxing lyrical about the five percent and then somebody else and by the time you get to the stage right now browser, you can't click a link in a browser and expect it to open a new link, a new page. What is going on? So the point is this group think is really pernicious and pervasive, but it all comes from emotions. If we can develop a habit of catching our emotions and asking ourselves, what am I feeling? I'm feeling a lot of joy. Cool. Okay. You're feeling joy and then re-examine. the facts of it and try and decouple the facts from the feelings, then I believe we can become better critical thinkers. But yeah, small feedback loops, becoming aware of our emotions, choosing a problem that you're obsessed about, and very, very, very small steps. Try and take a step, break it down into smaller steps. Take that step, break it down, before you do anything. And for goodness sake, don't plan anything out. plan the next
Nune: Yeah.
John Gallagher: step, maybe the next two steps, and try and just, instead of saying, we are going over here, try and let go of that, we're going over here, and try more, like embracing the idea, we're exploring the territory. We're gonna take a step, I believe we're gonna get over there, the second step will be a bit further away from there, and then you'll be going a very different direction. Another person to read up on is G. Paul Hill. Many, many more smaller steps. I can't remember the full acronym, but he has an excellent series on how many, many, many small steps gets you to your destination much faster because one massive step there seems as the crow flies, it will be faster. But actually the... smaller steps are a more effective way of getting there because if you take one massive step very often you find that you can't iterate from there because you're just in this kind of a while I've just written a thousand lines of code and it doesn't work so what didn't work I guess I have to go right back to the beginning now and all those thousand lines and all the tokens to create that were a complete waste
Nune: I'm absolutely enjoying this conversation and I have a thousand more questions, but we have to wrap up and come to my favorite section, which is book recommendations, and it's the time for your promise, so please go ahead.
John Gallagher: The Psychology of Software Teams by Cat Hicks.
Nune: All right.
John Gallagher: So caveat, I have not read the full book yet.
Nune: Mm.
John Gallagher: Generally, I don't recommend books without reading them. But this one is extremely special. Cat is one of the best pieces of advice I got way back in the day, apart from invest in things that don't change, is instead of trying to be world class at one thing, try and be like the top 5 % in two different things and combine it. because you will make yourself a category of one. The chances of you being unique are far higher and therefore it's a lot easier to sell your skills, your services, whatever. Cat Hicks is, as far as I'm aware right now, a category of one taken to the thousandth level. And the most bonkers thing is she's talking about something where there should be an entire discipline set up around it. I would have thought by now we would have a thousand companies talking about this and consulting on this and going into, if not 10,000, 100,000. And yet it seems to me like Cat is the only person doing this work, which is kind of mind blowing. the work she's doing is peer reviewed, evidence-based research on the psychology of successful. software teams. How is successful software made? from Cat, we're going to learn what are the thought processes, is the socio, what's a socio-technical side of successful teams. And she talks in the book, and I'm very, very skeptical about this, but if I trust anybody with authority about it, it's her. She talks about how to evolve. culture of the organization using scientifically proven measures basically. Which to me, I think evolving culture is, I want to say impossible, it's not impossible, it's just the hardest problem. Designing software is very hard, but designing a culture, embedding a culture, evolving a culture is unbelievably hard. and she's talking about science-backed ways of doing that. Evidence, certainly the evidence for evaluating the culture of your organization, evaluating your culture in a very scientific way, basically. Whether the book talks about how to evolve that or not, I think she gives some hints at how to do that. It's not just a step-by-step guide, number one, do this, two. Obviously, it's not going to be that simple, but a fascinating subject. somebody who's really done their homework in a big way. I was introduced to her by a post on LinkedIn. And from pretty much the first post on LinkedIn, I was hooked. I was like, wow, this person is talking about things on a very, different level of rigor to an AI-generated post about how culture just evolves. Or there's some wishy-washy, hand-wavy nonsense that I could write, anybody could write, Cat is the exact opposite of that. And I know that Charity Mages is a big fan of her and she's starting to get traction. So, psychology of software teams, I'm about halfway through it. I'm going to finish it off. She talks about a brains in jars mental model, which is extremely destructive. This idea that you have a... that a software team is made of an individual set of brains in jars. have four people in your software team, you have four brains in jars and they're just all stacked on a shelf and you give them work to do and they're geniuses and they churn out the software and all is good with the world. It's back to the feature factory kind of mentality to some degree. So yeah.
Nune: Maybe I can get her on the podcast one day. I hope.
John Gallagher: Holy... Yes. Holy Majole. I really hope you can't hold and listen to that episode.
Nune: we have also a tradition of every previous guest passing on the question to the next one. and Sometimes they're technical, sometimes not. And you get a question about pets actually. if I don't know, do you have pets? Let's start with that.
John Gallagher: This is a really bad question for me. I do not have pets and I'm going to leave it at that because I can't wait to hear the question.
Nune: The question is if you could own any pet, real or imaginary, what would it be? Why are you laughing? Ha ha
John Gallagher: I hate all pets
Nune: Yeah.
John Gallagher: with a passion. That's why I'm laughing. I hate them all equally. Iguanas, reptiles, snakes, rats, cats, dogs.
Nune: Do you have allergies maybe?
John Gallagher: I have this theory, right? And the theory goes, it's a very unpopular theory. I will answer your question, by the way. It's a very unpopular theory.
Nune: Yeah.
John Gallagher: Theory goes like this. Animals belong in forests and people belong in houses. That is it. Let the animals go into the wilderness, let the people live in the houses. We're all good. Put a wild animal in my house that's been trained and been acclimatised and been bred and inbred and whatever. We've got a problem.
Nune: Mm.
John Gallagher: I don't want urine all over my floors from some animal. But thanks for offering.
Nune: Mm-hmm.
John Gallagher: It's a great offer. I really appreciate it. But I'd rather not go down that route.
Nune: Yeah.
John Gallagher: So yeah. That said, only animal that is kind of... Sorry, I should also say I'm allergic to pet hair.
Nune: Mm-hmm.
John Gallagher: So that has been a big factor in not really being around animals and not being around pets and not having any pets. The evidence is against my position. Talk about critical thinking. All the scientific evidence is absolutely against... hating pets. Having a pet reduces loneliness, it improves our nervous systems. The scientific evidence is absolutely clear and yet, very much like all the JavaScript folks out there, I refuse to accept the reality. If I could have a pet, the only pet that even comes close, first of all, dogs versus cats, dogs. every single day of the week, cats just, I mean, I don't really get either, but dogs I can, so like a hypoallergenic, really chilled out, really cool dog that isn't ugly. So like, I saw a golden Labrador just the other day. I'm like, okay. Coming the other way on
Nune: Maybe.
John Gallagher: the street, you know, I'm like, okay. And if you go to a pub and, like an Irish pub or a British pub, and it's like one of those really sleepy kind of ones with like oak on the walls and like, you know, old pictures and just this nice, not like football blaring from a loudspeaker or anything, just this nice cozy thing. And there's a dog lying on the floor and it's not doing very much. And I don't have to care for the dog either. I don't have to walk it. I don't have to feed it. don't have to, like the dog is autonomous. It's an autonomous
Nune: Yeah.
John Gallagher: dog. that is very, very chilled out, is hyper-allergenic, that has no
Nune: It's all loops.
John Gallagher: bodily functions, that would be my pick for a pet.
Nune: Okay, virtual pets or robotic pets, would that be a thing you can ever say yes to?
John Gallagher: No, not robotic pets. It would have to look like a real dog. Right, so sure if it's robotic,
Nune: All right.
John Gallagher: fine, but don't let me know that it's robotic. I want the chilled out dog in the pub without any of the robotics, without any of the whirring, without any of the...
Nune: Mm-hmm.
John Gallagher: Any of that. None of that, thank you. Just the like, the chilled out bit. and not having to clean up any bodily effluents and also lasts forever as well. And just like, just wanders around and doesn't interrupt with my Zoom calls and doesn't like, but like, I can just pet it and it's nice and it cools my nervous system down. So yeah.
Nune: I love how these questions find their recipients. So now the pressure is on you for a question for the next guest. Can be technical, can be non technical, I mean okay.
John Gallagher: Confession, I haven't thought about this. Let me just think about this. I'm definitely not gonna ask anything about freaking pets. Forget that. Gosh, I need to find the person who made that question.
Nune: You'll you'll see in the podcast.
John Gallagher: I'm going to... My mock question for the next person is, do you think house prices are too high or are they the cheapest
Nune: Yeah.
John Gallagher: they've ever been? Yeah, that's my extremely uninteresting question. Let me think of a good one. What is the most powerful emotion that you feel and why?
Nune: Hm. Regularly or maybe once you have felt it and that drives you.
John Gallagher: No, just the most powerful. So like when you look back on your life,
Nune: Hm.
John Gallagher: it might be a specific feeling that you had at music festival 12 years ago, right? It might be a feeling tied to a moment or it might be a feeling like personally, as an example, my answer to that would be the feeling that I get when the bass drops. and the melody ascends when the music is already great and I'm dancing at the front of the club and I'm giving it my all and when the bass drops and then some kind of melody starts ascending out of that and that feeling of just awe and joy and all the emotional circuits in my brain lighting up like a Christmas tree. So it doesn't have to need to be like maybe the guest can't put it into words. It's not just like happiness. It might be when their child is born or it might be a special moment they've evoked a very specific emotional resonance or it might just be I like walking in the sunshine it gives me joy and that emotion so yeah whatever works.
Nune: Yeah, thanks a lot. That's a that's a great question. And thanks a lot for joining. I really, really enjoyed it and I know a lot of people will do as well.
John Gallagher: Likewise, thanks very much for having me, Nune.