In this episode of The Godel POD, Joshua Morgan, Client Director at Godel is joined by Kevin Goldsmith, CTO at DistroKid to talk about Kevin’s career so far and how the ever-changing tech climate has shaped a culture of agile working and distributed teams.
Joshua Morgan: Hello, everyone. Welcome Back to the Godel Pod episode 15 we’re at already, it feels crazy to say that my second appearance here. I’m Joshua Morgan. I’m a client director here at Godel Technologies, I’ve been here for almost two years now, which feels, again, crazy to say, but what a whirlwind two years it’s been. Really pleased to be joined by our guest today, Kevin Goldsmith, CTO at DistroKid. Kevin, how are you?
Kevin Goldsmith: I’m doing great. I’m super pleased to be here. It’s great to join. Thanks, Joshua.
Joshua Morgan: Look, it’s great to have you. I think I’ve heard your name referenced a couple of times by people at Godel, but I suppose for the people that are joining us today and don’t necessarily know who you are or what you’re about, it’d be great if you could just do a quick introduction on yourself.
Kevin Goldsmith: Sure. As you said, I am currently the chief technology officer at DistroKid. We’re a digital music distributor, currently the world’s largest. Before I was here, I was the CTO of Anaconda, the data science and Python platform company. I’ve been CTO at a few other companies on Fido in the UK, AVO in the US prior to that. But I first got involved with Godel when I was a VP at Spotify in Stockholm. So that’s when I met Neil, the CEO of Godel and worked with Godel, but also with some of Godel’s customers. Before that, I was at Adobe. I was a director at Adobe for nine years. It was eight years at Microsoft and some other places before that. IBM, Silicon Graphics back in the day.
Joshua Morgan: Perfect. I think you touched really well on your history there. I suppose, more specifically, if you look at the roles that you’ve had in all of those businesses, from Microsoft to Spotify to where you are now, DistroKid, you’ve gone from a number of roles. Starting off as a data engineer, computer scientist, engineer, manager, all the way to being CTOs of these businesses, and even today being almost an advisor and a speaker in certain roles. What’s that journey been like?
Kevin Goldsmith: It seems like it’s a fairly momentous journey until you realize it’s the journey of over 30 years now. So, I started as a software developer. I was working on teams, building stuff, I started doing graphics stuff, and that’s how I ended up originally at Microsoft Research, was part of the graphics group there and was just building things kind of found over the years. I sort of found a passion for the management side of technology I wasn’t particularly good at.
At the beginning, I kind of switched back and forth between being a developer and being a manager. It was really at Adobe where I really sort of figured it out and really found my passion around not just being a member of a team or being like a software architect, but actually kind of working on the people side and growing as a people leader. And that’s really where I started my path. And then once I was on that path, I got to that point where I was leading a large team at Adobe Spotify called I Have a Passion for Music. I’ve worked a lot in streaming media. So it was sort of very natural step. And from there on, after growing at Spotify, the team from like 50 to about 175. I was a very credible CTO type person and then have been just kind of on that journey since and just developing my skills and growing there since then.
Joshua Morgan: It’s really interesting. And you touch on that point about getting to a position where you’re almost respected as a leader in the technical space. Do you think maybe being a smaller part of those bigger businesses, like you said at the Microsoft, the Adobe where you were potentially a smaller role in comparison to CTOs of a business, do you think that helped in regards to your knowledge and learning about being a technology leader?
Kevin Goldsmith: Absolutely. I think having a diversity of backgrounds, having a diversity of environments teaches you, gives you the experience to understand what works and what doesn’t in different contexts. I think being part of a large company, seeing how very large when I joined Microsoft as maybe 18,000 people when I left Microsoft and I took a couple of year break in there to do some startups and then came back to Microsoft, I was at Microsoft eight years, but it was really over a span of really ten years. When I left in 2004 for the last time, it was 120,000 people, something like that.
So seeing how that size or my first job was at IBM, seeing how a large company operates was informative. And it was informative enough for me to know I probably didn’t want to work. That wasn’t the right necessarily environment for me. But seeing these different sized companies, like I left Microsoft, went to a super early stage startup with like 30 people and then did an even earlier stage. I see stage startup with like three people, the three of us that co founded it. So having sort of done all these different sizes, you get to see sort of what works at different scales and you pick up a lot of things like Microsoft does some really did and probably still does some really smart things in certain areas. Adobe does some smart things in certain areas.
I’ve picked up little things I’ve learned from each of these companies. And obviously Spotify, certainly when I was there, was doing just incredible things. One of the reasons why I went to that company, just really innovative things around team, organisation and things like that. So each of these roles I’ve learned something that’s informed kind of how I operate today.
Joshua Morgan: We’ll touch on the Spotify side of things probably later on in the podcast because I think there’s a couple of questions I’m really keen to understand about ways of working around that and almost being a bit of a pioneer about some of the models there. But I think one of the roles you actually haven’t mentioned is you actually did spend a little bit of time here at Godel, sort of three years as an advisor almost during one of the periods for us, which we saw substantial growth, extreme growth across sort of allocations in Eastern Europe. If you could talk a little bit about your time sort of here as an advisor.
Kevin Goldsmith: Originally I think Neil reached out to me for a couple of reasons. One was Godel was going through a period of extreme growth and he was trying to think about how to handle that, how to organise a company, how he was scaling the company. Separately, there were a lot of Godel clients that were extremely interested in the Spotify organisational model. And that led to Godel clients coming to visit Stockholm and me spending a day with them, showing them around Spotify, talking to them about that, just as a way of we were very proud at Spotify, of what we’d done and how we were doing and how we were operating. And just lots of conversations with Neil about kind of how to help Godel grow in a healthy way. So now I do consult. I do have an advisory business on the side. And it was really actually, I think my experience with Godel is what sort of kick started that for me, just really enjoying kind of, oh, okay.
I’ve amassed knowledge over the years, and it’s really great to be able to share it and to help. And that’s something that I’ve done since. But it really started with me working with Godel.
Joshua Morgan: And where you’re at today? DistroKid, CTO, still in the music industry. Touch a little bit on DistroKid. What’s the focus for the business? What sort of things are you facing into? We’re in April 2023. What’s the situation with DistroKid at the moment?
Kevin Goldsmith: So DistroKid today is doing exceptionally well. It’s one of these things that I saw sort of from the I’ve worked, you know, at Spotify. One of the things I was responsible for was ingestion of music. So working with the major labels, working with independent distributors like DistroKid started while I was at Spotify. And I remember us onboarding them into Spotify and on the side, I played in bands and ran a record label for many years. And I was working with one of DistroKid’s competitors. So I kind of understood the space and have done that for a long time. But joining DistroKid has been amazing. One, because it had a business model very different and unique in that it encourages you to actually release music just by the model. And that’s one of the things that has led to District growing so quickly. And then it’s just incredibly artist focused. We are very consumed with how to make it easier for artists to release their music and to help their music sound better, to help them build careers. And for me, it’s incredibly rewarding. I’ve been lucky to have been part of a lot of different things that have helped people. But this is something where you can directly see kind of artists release their music, grow an audience, actually some turn professional on our platform, which is really sweet.
Joshua Morgan: You touched on running a record label. How is that as an experience?
Kevin Goldsmith: That was a fun experience. So that’s something I started in the early 90s before there was streaming music. So it was me as a musician with playing in bands, friends with other bands. I had a good job at Microsoft. I had some money I could spend. I thought this would be one way for me to not only put out my own music, but to put out music of people I liked. So it was a very traditional I put out cassettes and CDs and vinyl records and got traditional distribution globally. I have a palshant for maybe not the most commercial music. I had a naivete that I was going to at least be able to do okay, maybe even someday do that as a full time job. Never did. Never made that kind of money, but did okay. Had a few artists do quite well. But one of the interesting things that was part of that was I still run that label now. It is 100% digital. I don’t generally put out on new stuff, but I still sell the old stuff. But I got to live through the conversion of the very early streaming. Some of the early piracy had problems with that.
Kind of how the entire sort of entire music industry has changed. I’ve gotten to see sort of on the vantage point of an independent label early. I remember paying hundreds of dollars to be on a platform called Liquid Audio, one of the earliest streaming platforms and you had to pay a lot of money just to get the software so you could be allowed to sell your music through their platform with them taking a huge percentage cut of it. Just the way things have changed is pretty outstanding. So it’s been kind of interesting to see it both on the kind of Spotify or DistroKid and was at Microsoft. One of the teams I worked on was Microsoft streaming audio teams. Getting to see it from sort of that side, but then also getting to see it from the artist side and from the label side.
Joshua Morgan: I was going to say it’s almost like there’s a full loop there from the journey at the start of the record label to now where you are as part of a business that is actively supporting artists to get on these streaming platforms and have a bit of a career for themselves. I suppose really interesting to hear sort of the background of where you’ve come from and almost where you currently up to now. We touched on it earlier about your time at Spotify and I think people that work in tech, maybe more tech leadership to be more specific, they hear a lot about this Spotify model of forming and leading teams. Now we know that you’re almost a bit of a pioneer and an advocate for this sort of model. I suppose what would be really useful is if you could just touch on that a little bit, a bit about your time at Spotify, a bit about how you went about forming and leading these teams and maybe the success that came behind that.
Kevin Goldsmith: Sure. So I want to be very clear because there has been confusion. I did not create the Spotify model. Not that you implied it, but people have thought because I’ve talked about it that I did, but I didn’t. I joined the company in spring of 2013. The Spotify white paper that described the original Spotify model was published in, I think, the fall November of 2012. So I joined the company partially because they sent me that white paper. And I was so impressed. I was there through the first several years of scaling it and the adjustments we made and kind of living in it, which was incredible. So me sort of at the time at Adobe, I was working on the problem of I was running a large organization of several teams. The teams had to coordinate. We were building sort of a family of products that all had different technologies that interlocked. And so we were having a very traditional software engineering problem of there’s a team building the core library and there’s a mobile team and there’s desktop team and they’re all trying to end a server team and they’re all trying to coordinate to release a feature in the product sort of simultaneously, which means that you end up with this sort of waterfall Gantt style kind of timeline where you’re trying to if this team is late, it screws up this release and all that kind of stuff.
So I was trying to figure out that I knew that didn’t work right. It wasn’t working well. It was taking us too long to do something. So I was trying to figure out, well, how do we kind of decouple these teams so there isn’t such a strict kind of dependency between them? I didn’t think of it in this terms at that time, but make them more autonomous, how do we kind of break up these dependencies? And I came up with something, some very simple ideas around this. It worked better, wasn’t working great. I started talking to Spotify. Spotify sent me this white paper, and that was what a true innovation they realized, is they had the same realization. They were doing it on even bigger scale at the time. And they’d figured out their team was about six times bigger than mine when I joined. And they’d figured out really how to do it in a truly kind of innovative way. So once I read that, I remember telling my wife, I have to go there. I have to live there. I have to see how this works. This was something I’ve been thinking about, and they’d already kind of made two or three orders of magnitude improvements over what I’d done.
And that was the case. That Spotify model is built around this idea of autonomous teams, cross functional teams, which, especially when they did it, was unusual. It was not the way companies would organize. It’s not that they invented that either. They’re very proud to say, no, we stole ideas from all lots of places. But really that was unusual at the time. So they were building teams that would have not only developers, but have developers and testers front end developers, back end developers, mobile developers, all on the same team. So when they built something, they would build it everywhere, all at once, right? They would build the whole thing soup to nuts and just deliver it. And they also worked on how to up delivery cadence. Back then it was hard to release your iOS app more than Apple wouldn’t really allow you to release it too often because they had to do this heavyweight review. So they were discouraging developers from submitting apps too often. So Spotify had to even figure it out ways to get around that so it could release their applications more frequently. So it was also just an incredibly agile I’ve been an agile software person since 2000, since really the extreme programming book came out.
That’s the only way it worked. And I kind of struggled with it at larger companies. I’ve done it at Microsoft. I kind of allowed that. I got that convinced them to allow me to do it at Microsoft, which was very antithetical to the way Microsoft worked. Adobe did it better, but again, it was kind of big company, agile. Spotify really was agile from its heart and so that was another important piece of it. So Spotify organized around these kinds of small independent teams working at their own, each of them working about as fast as they possibly could go. And then had structures around them that supported them so that they could continue to scale it across a wide variety of areas. But it’s also structure let you scale very well because you weren’t having to have as you added a new team, it didn’t have to figure out how to coordinate with 80 other teams, it just could do what it needed to do. And so that was sort of the innovation of that Spotify model. Now, as Spotify has continued to grow, I also want to be clear, as Spotify has continued to grow, they’ve continued to modify it.
So I left Spotify about seven years ago and it’s several factors of magnitude larger as a company and that’s continued to change. And so there’s things that I know that they don’t do anymore because it doesn’t work at that level of scale. But I can tell you that that model scaled incredibly well from there was about 300 developers when I joined, about 800 when I left. And we added one new kind of innovation onto that, which was we got to the point where we couldn’t scale horizontally anymore and we added another level, but that was about it. And that model, I think, lasted, I think to another order of magnitude even after I left. The funny thing is, when I would talk to folks at Godel, or talk to Godel customers, I would tell them a lot of them were very interested in copying the Spotify model. And some Godel customers had already started to do that. And I said absolutely do not copy the Spotify model. Not that it didn’t work, but that it worked for Spotify. It wouldn’t necessarily work for other companies because Spotify had also a very unique culture and a very Swedish culture that was very team focused.
It was very supportive of autonomy. Most companies don’t have that idea of autonomy is very interesting, but they don’t actually know how to do that. And so I was very vocal about telling people, these are the cool ideas that we do. You can steal these ideas like we stole ideas from other places, but do them in your way, don’t try and copy them in our way. And lots of companies actually a lot in the UK for some reason had read this and were trying to copy it, and we’re struggling because it didn’t really match with their company culture. What I’ll say is sort of since those days, there are now lots of companies in the UK and elsewhere that have kind of copied that Spotify model. More cargo culture, that thing is probably the biggest one. But then they put their own twists on it and did their own thing. So they kind of did effectively figure out a lot of companies did effectively figure out how to steal some of those ideas and then kind of put them into the context that worked for their company. So I went from just telling people, please don’t, to seeing just more and more examples of companies utilising those ideas and putting their own twists on them and figuring out how to make them work. So now I’m a little bit less I’m a little more aware what you’re doing and be careful about what you’re doing. But I’ve now seen it work in some companies, actually quite effectively.
Joshua Morgan: I think it’s really interesting and you touch on the point of it not working everywhere. It’s very much a case of a business has got to take a model that can be successful somewhere and adapt that to its own circumstances that isn’t necessarily just a lower level in terms of how they sort of structure their teams. Actually goes all the way to the top. It goes from a management all the way down. So I suppose from your perspective, you’re more saying it’s more of a learn and adapt a model rather than take what someone’s got and just include it in this business and everything.
Kevin Goldsmith: Probably not in every industry, but the industry I know is the software industry. We love trying to like, oh, this worked for this technology worked here, we should use this technology, or this model worked here, we should use this model. We really like doing that and it generally doesn’t work right. You can do something that Google does, and it works great for Google and doesn’t work for you at all because Google has Google problems and you have problems. I don’t have Google problems, I have distributed problems. But I think the more mature companies or the more mature leaders say, oh, well, that’s an interesting idea that Google did this. Why did they do that? How did they get to this thing? Is there parts of this that make sense for us or make sense for our company? Let’s steal that part of this. This is sort of an interesting way of doing that. Let’s integrate that idea, or let’s take the spirit of this and figure out how we can do something in that spirit that makes sense for us. Every company has its own unique culture. There’s the things that people at the company believe. There’s the things that people at the company do. And you have to honour that, because if you try and do something without really kind of thinking through, well, what’s this going to work like for us? Yeah, it almost never works because you just live in such a different context.
Joshua Morgan: And you sit here now moving away from Spotify, as you said seven years ago, and you’ve been into a number of businesses, and as you said before, gone in as an advisor, a CTO, different roles. How much have you taken from that initial model and made it your own? Because you’ve spoken perfectly about there, about other businesses need to do it. But I’m sure that you’ve done exactly that same thing. You’ve moulded that exactly how you want it to look. What have you taken away and what have you added to those models?
Kevin Goldsmith: The idea I think there’s a lot of I learned at Spotify. For me, it was almost sort of an MBA in organisational culture and organisational design in quite a lot of things. And so there’s ideas. The company I went to after Spotify was called AVO. It’s a legal marketplace company. And there they were kind of interested in Spotify. But I was very clear, we’re not going to do this Spotify. I’m not coming here to turn you into Spotify. But what I found was the ideas around autonomous teams. And we were also different. We were growing company, but at a different level of scale. So we took this autonomous teams idea, which is something that is one of the things that I absolutely carry with me and look for as a way of scaling teams. So in scaling companies, I think it’s an incredibly effective way. But what that looks like in different teams is different. At AVO, we created something called journey teams that were built around user journeys and organized that way and didn’t have these tribes, because we weren’t tribes. At Spotify were anywhere between 50 and 150 people. The entire engineering team was 150 people at AVO.
Kevin Goldsmith: So what we did was created smaller teams, smaller teams that were more dynamic to service the needs of the business, were organized around user journey. So it worked slightly differently, but it had some of those same kernels of how we organized from there to Amphito. So I was at Amphito in London and we did something again similar. We moved towards more autonomous team model because same thing scaling company. But the way it worked there was very different the way we’re doing things at DistroKid. Similar idea moving towards a more autonomous team model, but in a way that’s kind of very uniquely us because we’re a very different kind of company than any of those companies. So there’s these core ideas around. For me I think it’s this idea of how do we decouple hard dependencies between teams and there’s lots of different ways to do that. That’s one of the things that I carry with me to every company. There’s lots of things around Agility and how do you build kind of scaling agile, which is different, organization is tied into that, but they’re really different things like how do you scale agile processes across larger groups, things like that that I carry with me and they’ll look different in every company. But there’s some core things that underlying those things that are informed to me.
Joshua Morgan: One thing I really like that you touched on there is about big scaling businesses. We’re seeing so many companies, not just in the UK I’m sure it’s happened in the US as well that have scaled so significantly over the course of the last two years. Maybe off the back of a bit of success during COVID and now layoffs on top of layoffs and these businesses are almost panicking at that point and scaling back. One thing I was keen to ask you is when you’re going to these businesses that are looking to scale pretty significantly over a short period of time, how would you make a success out of that? How do you get a business to a position where it isn’t in panic mode at this point? Now I know that’s a question that is easier idolised rather than answered but what’s the success behind onboarding and scaling up these businesses?
Kevin Goldsmith: I haven’t had to do it this time around but certainly I shut down entire company in the.com explosion of 2000 actually two companies. So I do have experience with it. The trick of it is it’s hard having been inside multiple scale ups, right? I can say one of the kind of companies that usually try and hire me are companies that are scaling because I have a lot of experience scaling and growing companies. The mistakes I often see or companies I advise because I advise a lot of companies as well. The mistakes I often see is this kind of forgetting how this whole thing works, right? Our industry, like most industries, is cyclic. We’ve seen this now happen several times of boom and bust, boom and bust. But especially we tend to be kind of a young industry and we have lots of young kind of first time founders who don’t have a lot of experience in this. And what happens, we have a lot of money, we have free money kind of endlessly available. We have a strong tech especially has a strong belief, especially because of venture capital or growth PE towards this like grow big fast, right?
So that’s what companies do. They take in as much money as they can, they grow as quickly as they can with the intent of being the winner. And sometimes it works. That was Spotify. Spotify, that works great for Spotify, that worked for Google, that worked for Amazon. There’s plenty of companies that’s worked for, but way more it hasn’t. And so when I’m in a company, usually what I try and do is, at least from my perspective, working with the CFO. I make sure we try and we may be growing quickly, but I’m always trying to make sure that our growth tracks, a growth in team tracks against growth in the company, growth in revenue. And so that way we don’t get too far ahead. It doesn’t mean that when times get bad, sometimes you maybe have gotten a little bit too far ahead of your own revenue growth or you’ve been growing right at your revenue growth and revenue cuts because of economic situation. So it doesn’t mean you can completely avoid it. But what I’ve really seen go poorly is companies grow so fast. Like, oh, we have all this money. We got to hire a bunch of people so we can have the people to make our revenue grow, and then our revenue will grow, and then we’ll hire more people, or we’ll just raise more money, and then that game works until it doesn’t. And when it doesn’t, it tends to fail fairly spectacularly.
Joshua Morgan: And on that onboarding piece and scaling up, how’s the shift been with this hybrid world we’re now part of COVID has shifted technology even more than we actually thought was possible. And look, technology changes every day. But now that we’re in a hybrid world, everything seems to shift overnight. Do you think that’s become more difficult in this hybrid world in terms of scaling teams and bringing people on?
Kevin Goldsmith: Prior to COVID, I’d led distributed teams, right? So it’s Spotify, for example. Or adobe. My team was in Seattle, my team was in San Francisco, my team was in Minnesota. I had individuals in Oregon and Virginia and all over the place. So we were coming. That was pretty good with distributed teams, but people were generally working out of offices. And at Spotify my team was in Boston and New York and Stockholm and Gothenburg. So I was kind of used to that model to a fully distributed people sitting in their houses team really was until COVID, like most of us, right? So that was a very new experience for me. What I’ll tell you, I learned there’s things I love about it. And since then District is fully distributed, we have no offices. Anaconda, when I joined it was fully distributed. We had offices, had gotten rid of them. The thing I really appreciate about this is I hire people everywhere, wherever they can. There is such an amazing amount of talent. I’ve been lucky because I’ve managed people in Israel and Portugal and all over the place.
I know there’s talent all over the globe, very good developers and I love being able to hire folks wherever they are.
My CEO is in San Francisco, the COO at DistroKid is in Portland where we are all over the place. I had to turn down jobs when I left Spotify because I wasn’t in the right location or wasn’t willing to move. Right. So I love for me that I personally can do this, but I also like that when I’m looking for talent, I can find it wherever it is and there’s great talent everywhere. That part I very much appreciate. The onboarding is harder, it is just like realistically it is harder, it is not the same. And we learned lots of stuff over COVID to make it so that we play games together over zoom or we do other things or whatever. It is different and we just have to acknowledge it’s different. You’re not going to have the same kind of kind of cohesive teams. You have to work a little bit harder to make teams, teams. That said it’s not impossible, but you just have to acknowledge it’s front the benefits you get from being fully distributed and having all this access to that talent. And essentially now since COVID I’ve been in companies that essentially operate 24 hours.
Not because we are putting a team in Singapore and we’re putting a team in India. It’s because we have people in every time zone. So I don’t have to set up on call schedules where people have to wake up in the middle of night. There’s somebody that’s always up and working. The bad part is we have to do more asynchronously we have to assume that somebody can’t attend a meeting, things like that. We’re now starting to do more where we bring teams together for little periods of time, kind of build up that camaraderie before they all go back to home and then kind of work from there and just trying to figure out the right cadence of bringing those teams together. That is something that I did with larger distributed teams, but that’s something that companies I know that were fully remote used to do even pre COVID.
Joshua Morgan: Yeah, I think it’s interesting as well, because the context of Godel as a business and looking where our engineering talent is and the conversations we have with businesses based in the UK, we’ve probably noticed it as well, that the businesses are a lot more receptive now to working with a software provider that has the talent based in Eastern Europe. And it probably comes really hand in hand with what you said there about opening up your talent pool, understanding that the best people aren’t just based around the corner to where your. Offices is and understanding that if you open your horizons as a business, not necessarily just with a partner like Godel, but also, as you said, with those talent pools or resources that sit as part of the internal teams, you’re opening your growth and sort of streams as where you can take the business to another level. Really?
Kevin Goldsmith: Absolutely. I think earlier in the industry, there were good universities in the UK, good technical universities in the US, maybe in India, but it wasn’t really understood. We didn’t believe that some of the other countries were producing talent at that same level. That’s definitely no longer the case. There’s amazing talent, and the Internet absolutely helps with that as well. There’s amazing talent in Poland. There’s amazing talent just all over now. It’s really different.
Joshua Morgan: I think conscious of time, we’ve got around sort of ten minutes left or so. Moving on to what’s next, then? As we’ve touched on tech progressing and new advancements, we see everything, we see things in the news every day. I think the biggest thing at the moment is AI and Chat GPT. I think I was keen to pick your brains on that, given the industry that you’re based in. I was reading an article the other day about Spotify’s new AI model, which provides you with music that you want to listen to every day. And if you tell it that you’re not interested in this, then it’ll provide you with something different or it’ll change mood swings. I was interested to see a DistroKid looking into anything with Chat GPT. Is it something that you’re actually sort of looking into what’s on the radar in terms of that?
Kevin Goldsmith: I’ve also been lucky to be part of the industry as kind of machine learning and AI have kind of grown and become more accessible. And I’ve worked at machine learning companies, I’ve worked at AI companies, so I’m very interested and I’m very aware of kind of that. And we are absolutely paying a lot of attention to this. We are absolutely thinking about how we can use AI and machine learning to help improve experience for our customers, help make them more prolific, help make what they do better, help them grow their careers. We’re absolutely looking at that a lot. So, the head of product at DistroKid is somebody I worked with at Spotify and was one of the people Spotify had been doing machine learning generated recommendations for a while. But if you really think of kind of the thing that kind of kick started the company around this space, it was really if you think of Discover Weekly or Release Radar, those kind of now your daily mixes, that came from this product manager that’s the head of our product, Matt Ogle. So, he was really the one that pioneered that stuff. So, both of us come out of that world and we’re both really interested in figuring out how to use it effectively.
I think it’s interesting not just for what we can do for our customers, because we were going to figure that out anyway. The new large language models, especially Chat GPT, are interesting for me as a CTO. Not just in how we can use it in the products, but also how we can use it to actually improve and make our teams more efficient. So that I’ve actually been spending quite a bit of time with as well. So not only just like co-pilot and that kind of code generation techniques, but what else can we do to improve our customer service, to improve our SEO, all those kinds of things. I don’t want to overstate it because new cool technologies come out all the time. Large language models are like it’s a massive step forward in usability and accessibility of machine learning, but they themselves are just cool things. Cool models that were trained on a lot of language. They’re not revolutionary stuff, they’re an evolutionary step of what array existed. The real revolution was making it just so accessible for normal people that they now get to see the potential here. But that said, the difference between what I would have had to have done to do something to help the customer service team, for example, kind of understand the different conversations or surface kind of patterns in customer service. Chats, it’s so much easier now, right, with the other side of it, of we have to be very careful. OpenAI will use that data for training unless we do some stuff. So, it shows a lot more possibility, but it’s definitely going to change how we do our business. We’re just figuring out the right ways to use it right now.
Joshua Morgan: Yeah. And on that, are you concerned about any areas of this? Reason I ask is I saw a really interesting video of the Day of and it’s very specific. Someone had managed with an AI model to create a song that was, I believe it was a Justin Bieber song, that they then managed to use an AI model and they changed the voice to the Weekend within the space of about five minutes. Now, look, that’s harmless, it’s a bit of fun, but do you think little things like that is just showing a bit of a sign of where it could go in terms of a concern? Are there concerns within the business where there’s a bit of risk around this?
Kevin Goldsmith: Any business that allows people to make money has a fraud problem, right? Streaming has a fraud problem at Spotify. When I was at Spotify, one of the things we were constantly doing is finding people who had uploaded copyrighted music as their own or people who had taken were flooding Spotify with music whose copyrighted expired and trying to get clicks, essentially like you do on Google where you’ll try and do click traps to make money. So, we have to deal with that. As a distributor working with Spotify and apple Music and Deezer and all these other companies because we’re essentially the first line of defence. What we want to make sure is that we are not enabling fraud at scale. So that is an area that we do data science on now. That’s one of the things that we were already doing data science when I joined the company, is looking for fraudsters. It’s very similar to when I was in Fintech. Some of the things we had to do there, same kind of problem. There’s money to be made, so people are going to try and figure out how to cheat the system to make money.
So, this new technology absolutely a problem for that on one level because the law here is very open because it’s so new. So, if I sold a weekend’s voice or Bieber’s voice or whoever’s voice, put it on one of my tracks, but the music was an original composition. Okay, is that copyright infringement? That’s right now unsettled law. So that’s something we’re being very aware of and being very careful about. To your point, a lot of it today is more in doing fun stuff and making viral videos. But realistically that’s not going to last. It’s not that many more steps. I think the other thing we’re also very aware of is kind of AI or machine generated music and kind of flooding services. This is already a challenge because there are people that are generating 60 minutes of white noise and uploading it, and people listen to it and the major labels don’t like it, but people are listening to it. So, this is something that we kind of as somebody that’s supporting the artists, but also somebody that has to work with the streaming providers. We’re sort of having to negotiate this a little bit.
As machine learning gets more involved in music making, they’ll just continue to be a thing. So that’s certainly something we think a lot about, but I can’t say definitively kind of what we’re doing around that. But we are doing things around it.
Joshua Morgan: Yeah, of course. Look, it’s really interesting and I could sit and listen all day about these new models and almost your past and where you’ve come from. I think we’ve got a couple of minutes left. I think one question I’d be really keen to leave you on is sort of a leading question. If you could give one piece of advice to technology leaders out there that are looking to potentially end up in your sort of position. They’re starting off small and they want to get to these high heights, what would that one piece of advice be?
Kevin Goldsmith: Be interested, be curious. Try and learn as much as you can from as many places as you can, but then take the time to kind of synthesize it into what it means to you or what pieces of it resonate with you and always challenge yourself. I think the one thing that sort of me earlier career to me later career is me figuring that out and talking to lots more companies, trying to understand why they do what they do, trying to figure out what I thought about it and really kind of parse it and understand it for myself. That really is the difference of kind of what let me be a really effective kind of architect and maybe small team manager to being able to run larger organizations.
Joshua Morgan: Really interesting. Kevin, it’s been great to chat to you today. Thanks for joining us on the Godel Pod. I’ll leave you with one thing. If you’ve got anything interesting you want to shout about now, anything interesting going on in your life or anything you want to give a shout out to, now is the perfect time to do so.
Kevin Goldsmith: As you mentioned earlier, since I joined district kid and I’ve been a little bit quiet, but because I’ve been busy. But I do speak at conferences, I do write. The easiest place to find sort of either talks I’m giving at different conferences I speak around the world or sort of my latest kind of thoughts generally around kind of management leadership are on my website. Kevingoldsmith.com, perfect.
Joshua Morgan: Thanks for joining us, Kevin. And thanks everyone for listening. Keep your eyes out for the next episode of The Goddard Pod.
Kevin Goldsmith: Thank you.
[Outro] Joshua Morgan: Huge thank you to all the listeners out there who’ve listened to this instalment of The Godel POD. If you like what you hear and would like to know when we’re releasing even more episodes, please just subscribe to the Godel page on pod bean or Spotify.
Godel Named One of The Fastest Growing Larger Companies at the 2023 Northern Tech Awards
Godel Technologies has been recognised as one of the top 100 fastest-growing companies in the North at this year’s Northern Tech Awards. Organised by GP Bullhound, the Northern Tech Awards, which this year celebrated its 10th edition, shines a light upon the leading technology businesses headquartered in the North of England and Scotland. This year’s award ceremony… Continue reading The Godel POD: Finding the Stream with Kevin Goldsmith
Godel Celebrates 100 People in Vilnius
At the end of 2022, Godel celebrated its second anniversary in Vilnius, Godel’s first Lithuanian location. Over the past few years, both the team and the Godel brand there have grown, establishing the capital as a key Godel location. The team, which started off with just over 20 people has now grown to 100 Godel… Continue reading The Godel POD: Finding the Stream with Kevin Goldsmith