Tag: DevOps

Tyler leads the firm’s developer-focused investments and publishes the industry’s most comprehensive market landscape each year. He discusses the developer-led landscape and the 2022 Outlook. The video is above and a transcript of the conversation is below.

Interviewer:                This is Digital Anarchist.

Hey, everyone. Welcome to another Techstrong TV segment. I’ve got a first-time guest on Techstrong with us, but he’s a guy, you know, a gentleman who has his kind of finger on the pulse of what’s happening. I want to introduce you to Tyler Jewell. Hey, Tyler, how are you?

Interviewee:                Ah, good morning, Alan. It’s good to be here.

Interviewer:                Cool. Tyler, you know, I said you have your finger on the pulse, but I purposely didn’t give your whole title, affiliation, background, ’cause I was kind of saving it for you. Why don’t you introduce yourself to the audience.

Interviewee:                Oh, I’m happy to do so. So, Tyler Jewell; I am a managing director at Dell Technologies Capital, which is the VC arm of Dell Technologies. We invest about $300 million a year across disruptive companies in the datacenter and enterprise space. You know, previously, I spent 25 years as a product operator and software executive in developer tools companies, including at Stenset, Red Hat, Oracle, and being a CEO of three different startups that have all gone pretty well. And I’ve been an investor, as an angel and VC, for 15 years, across 15 different companies, again, in the developer tools and developer landscape. And starting in 2009, I built a database around companies that make products for developers, and I published it online in 2019, and it now has 1,100 companies in it.

Interviewer:                That’s fantastic. And what is that database, by the way?

Interviewee:                We call it the Developer-Led Landscape, and, you know, we show you the trends in which segments are happening in the developer tool space, which products are gaining steam, which ones are losing steam, and, you know, general insights about how the market is shifting and evolving.

Interviewer:                I love it. What’s the URL? If you don’t mind [crosstalk].

Interviewee:                Oh, it’s at my Substack, so tylerjewell.substack.com.

Interviewer:                Excellent, and that’s a great resource – people should check it out. So, you know, it was interesting, Tyler, I spoke to a few CEOs, already, this week, and a common thing – and a few VCs, I might add. And a common thing I’m hearing is, are we seeing a change in the market conditions, right before our very eyes? One of the large, you know, public PE firms or – you know, we live in a world, today, where the distinction between VC and PE is increasingly blurring, right? It used to be very clear what was a VC and what’s a PE, but one of the large VC/PE firms, look, they took a big hit, this week, on their stock price, right? What effect is that gonna have? Is that gonna put a chill?

By the same token, Tyler, last week, I interviewed four startup CEOs who all raised $150 million or better, and achieved unicorn status, right. So, it was the best of times, it’s the worst of times, right? [Laughs] I don’t know, look, if I knew, I wouldn’t be sitting here, I’d be off on some, you know, South-Pacific island or something. But, you know, where are we headed? What’s going on? Are we seeing a changing of the landscape? Is it to be expected, an evolution? What’s your take?

Interviewee:                Well, you know, let’s deal with some hard math, to begin with. First, there’s well over 900 companies that have achieved a unicorn status on paper. But when you look at the past year, there was only a handful of companies that were able to get an IPO, in a year, arguably, where the demand for IPOs has been one of the hottest in a long time. And, you know, and then when you look at the global PE market, again, there’s only about appetite for maybe a dozen PE acquisitions in that unicorn range. So, you know, if you’ve got a hot market and 30 companies can get liquidity, really, either through a strategic buyer, an IPO process, or private equity, and you’ve got 900 companies at that valuation range, you’ve got a 20- to 30-year backlog for all those companies to get their liquidity. Or, you know, you’re gonna have, potentially, a correction if those companies need to raise some more cash.

Now, the flipside of this is that there’s a tremendous amount of cash on the sidelines. Most of these VC funds, private equity funds, have recently raised significant amounts of money; their goal is to deploy that money. They oftentimes have deadlines by which they can deploy that money, and so, there is, you know, a valuation and sensitivity that comes from those companies that, you know, need to deploy the money into what they perceive to be high-quality companies or early-disruptive companies. Now, what we do at Delta Capital, you know, we think of ourselves as early-stage venture capitalists, you know, seed through Series B is our first check-in, and, you know, for us, it really matters that is it gonna be a high-quality team, is it gonna be an emerging and important market, is the product heavily differentiated.

And in that situation, I think we’ll be a little bit less – we are more valuation-sensitive, meaning, you know, we’re not gonna be as insensitive as some of the others who have a large amount of money to deploy. ‘Cause we don’t have any minimums, you know, so we have the ability to look for quality deals and to, you know, pay what is appropriate given the market conditions that are out there.

Interviewer:                Yep, agreed, you know, I don’t disagree with, frankly, anything you said. You know, but when one hears that number of 900 unicorns in a market of 30, you really get to see, you know, it takes me back to when I first started doing startups in the venture world, right, where, look, nine out of ten venture-backed startups, or eight-and-a-half out of ten, you know, didn’t succeed. They were lucky to get some of their money back, right? [Crosstalk]

Interviewee:                Well, I mean, venture is definitely a boom-and-bust cycle, you know, when you’ve looked at the history of it over the past, you know, it’s been 35 years that we’ve now got a venture capital out there. And taking it back down, you know, to more narrow, what happened in the developer landscape, you know, we were speculating on the potential that we’ve hit peak dev ops as a market condition here. And one of the datapoints that we cited is, last year, in the 1,100 companies that we track, which just, you know, make products for developers, they had, collectively raised $16 billion, last year, across all those companies. And at the same time, that same cohort only had $35 billion in liquidity.

Now, you could sit there and go, “Okay, well, $35 billion out is better than 19, you know, the $16 billion in,” but really, that number should’ve been 5 times larger, if these things were in balance. You know, that $16 billion maybe purchased, you know, 10 to 15 percent of those companies. And in that $35 billion in liquidity, that was only a handful of IPOs, and the rest was, you know, like, 23 of that 35 was from IPOs, and the rest were from acquisitions. And a lot of those IPOs have been underwater, now, at the beginning of this year, right? So that $35 billion number might actually be a lot closer to 17, at this point in time.

So, you know, the math just doesn’t add up; something’s gonna have to give at some point in time. You know, that’s certainly one factor. The other factor, you know, facing the developer landscape in particular, when you _____ _____ _____ companies, is that it’s just gotten really complicated to build software, over the past two decades. And, you know, it’s a confluence of a number of factors that are out there. One, you know, the wage growth for software engineers hasn’t grown as fast as the demand for software. It’s pretty remarkable, you hear about all these software engineers who are supposedly making a lot of money, but when you look at the Bureau of Labor Statistics, you know, from the federal government, on what engineers make, it’s been a pretty steady four or five percent linear increase, for the past two decades, on that, and it really hasn’t changed much.

And as a result of that, there’s been a pretty linear increase in the number of engineers available to work. So, programs aren’t spitting out more talent, companies generally aren’t hiring more talent. So there’s a disconnect between the demand for software, which has been, arguably, on a logarithmic exponential curve, and the supply of people who can do it, right, which is [crosstalk] on that front. So the intersection of this is, “Hey, let’s have a bunch of tools. Let’s have 1,100 tool company out there, to try to help with the productivity of all this.” And as a result of that, we’ve created an ecosystem and an industry that’s really complicated. You know, agile processes are complicated, microservice architectures are complicated, cloud, cybersecurity, you know, huge surface area of attack. And then you’ve got a talent population with only linear wage growth, and this is gonna be an issue, here, this is not sustainable.

Interviewer:                So I gotta play devil’s advocate, if you don’t mind.

Interviewee:                Be the devil, that’s okay, I’ll be the angel.

Interviewer:                [Laughs] Right? It’s a good role. So, number one, I think things have grown more complex, and I think, you know, that is due to an evolution, you know, the continuing development of this whole software development universe we live in, things tend – you know, as you build more things on top of more things, I don’t know if they necessarily get simpler. I think they do get more complex. I also think that the needs of the – the requirements put on the applications that we’re developing, the requirements for what we need this software to do have grown exponentially. And so, as a result, that’s part of that complexity, too. But certainly, developing an application on a microservices architecture deployed via Kubernetes and the cloud is, you know, when we talk about things that are hyper-scalable, right, there’s sort of the mainframe stuff, which is very hyper-scalable, right, probably the most, one can argue, you know, penny for penny, dollar for dollar.

And then, you know, the scalability offered by these architectures, you know, and you can get into the thing that it’s only the unicorn, you know – I mean the real unicorns: Google, Facebook, Twitter – that need that really hyperscale, right, there’s a handful. But clearly, these are the architectures that do that. And in terms of the amount of developers, I mean, look, unless GitHub is blowing smoke up everyone’s butt, you know, they – something, I think it’s almost, like, 50 million GitHub accounts, today, that developers, if they were a separate country, would be the fastest-growing country on earth. You know, not just US, obviously, you know, I’m talking about global. And then, in terms of compensation, look, I’m not a big believer that the Great Resignation is due to the money being paid to software developers, or cybersecurity people, or dev ops folks, or what have you.

But I know enough of those people to know they’re making good livings, Tyler. You know, my friends who are CTOs and VP of engineerings are paying good software developers a quarter-of-a-million dollars, and more, and more. And not just those guys in Silicon Valley. New York, Boston, Austin, down here in Miami – you know, we’re in Boca, but –

Interviewee:                There is, you know – you know, first of all, I think GitHub has something like 78 million repos now, right, [crosstalk] probably even claiming 50 million accounts. You know, but there is a massive disconnect between the data that Microsoft and GitHub gets to report and what the BLS statistics, you know, share, right. Which is, you know, the indication of people who are professionally paid, you know, who give themselves some sort of software engineering title.

Interviewer:                But maybe the proliferation of titles says something – I know, Tyler, in our mailing list, you know, we try to mail based on title, sometimes. If I showed you the titles people have, it’s crazy.

Interviewee:                Yeah, you know, look, there could be some title proliferation, but the BLS doesn’t let proliferation, you know, proliferate. And also, it doesn’t reflect international trends, as well. But just in general, when, you know, there’s a lot of struggles and there’s a big gap that exists between people who are professionally paid as talent workers to create software, you know, versus an even larger population of people who aren’t, you know, who have software development skills. And they may contribute to opensource and they may contribute to GitHub, or they may even contribute to some project, you know, but they’re not being compensated for the work that they’re doing, and, you know, they’re doing it for some other motivations, right. So that can explain a huge portion of the disconnect.

But generally, you know, the software that’s produced to satisfy the demand of consumers, right, or businesses on that, it’s usually gonna be a team, that’s under compensation. Because sustainability of that software, maintainability of that, holding against the SLAs, you can’t depend upon people who are not under contract with you to achieve that. You know, and generally the case, yeah, I think that, you know, a lot of software has gotten so complicated, and the media has definitely portrayed a lot of people who have got, you know, outsized compensation, you know, because maybe they have a unique type of software engineering skillset, or they’ve, you know, risen to some sort of CTO or VP of management position. And I think those stories are gonna continue to persist.

But when you look at the data in aggregate, you know, for people who identify as some sort of quality assurance engineer or a software engineer or a cyber _____ professional, the aggregate data across the United States hasn’t, you know, been on the same curve as what the storylines would suggest on that. And –

Interviewer:                Yeah, but, you know, to be first-world, you know, first-world problems, let’s be clear, for my audience out there, our audience is developers and engineers and so forth. But when you go down the block to your neighbors and you say, “Woe is me, I’m only making a buck-and-a-half a year, you know, as a cybersecurity admin,” or something like that, and the guy next to you is doing something physical and, you know, making $75,000.00 a year, you gotta be careful, you know?

Interviewee:                Look, I think we’re putting the spotlight on the wrong thing, right. I think the bigger spotlight, here, is that there is a $50 billion revenue industry around the dev tools, right? So those 1,100 companies that we track, they generate $50 billion of revenue a year. That’s going up 23 percent, year over year. That’s $10 billion increasing, a year. And I look at most of that revenue, most of those sales, as companies who are trying to provide forms of automation to compensate for the fact that the global size of the professional engineering pool isn’t large enough, right? Now, you know, and this market has gotten _____ fragmented. You know, 5 years ago, there was only half, there was only 550 companies.

So, you know, that market’s a lot more fragmented, there’s a lot more competitors, that’s adding to the confusion and the issues that are plaguing productivity and the engineers. Now, on the flipside of it, you know, I would also argue that we’re not doing nearly enough on advocacy, in that we should have three to four times as many professionals working in the software space. And there’s a lot of reasons why people are unable to become professionals in the software space, as well. It has to do with STEM education, there’s misogynistic and discriminatory behaviors that keep people out of the space. It’s also the complexity of the technology; there’s still a high degree of math and science that tends to be required.

And those things are also acting as barriers, here. And so, the flipside of this is, that’s why we have this $50 billion industry, this cottage industry, around trying to make people more productive, because we’re compensating for the fact that there’s just not enough people, and we’re not doing enough to get more people into the industry.

Interviewer:                I don’t disagree wholeheartedly, there. Here’s my take on it, though. There’s two ways of approaching the issue of we need everything – every company’s a software company, right, and we need more software. Either we’re gonna have more people doing software, which is, you know, a big shift, lift and shift, or we gotta, you know, to paraphrase Matt Damon _____ _____, we gotta automate the you-know-what out of it, right? Probably, we need to do both, right? But I would say, also, when you look at – let’s take opensource software, right? You and I have both been around a long time. I remember, you know, talking to doomers at army bases who told me that they were never gonna – you know, they don’t allow opensource software onto the network.

I’ve been in enterprises where I was told that. Today, 90-plus percent of enterprises are probably running opensource, you know, somewhere in the network. But more than just, like, running Linux or something like that, so many of the components of today’s software are, you know, opensource-based. And you’re right, there’s not a professional opensource – well, maybe there is – is there a professional opensource engineer community today? Is it enough to support it? I mean, I don’t know.

Interviewee:                You know, look, are there individuals who can just be professional opensource contributors and make a living off that? In some pockets, there are, particularly the people who focus on identifying bugs and security vulnerabilities, you know, [crosstalk] a different thing.

Interviewer:                Yeah, that’s a subset. No, but I’m thinking, like, Tyler, look at CNCF, for instance, [crosstalk] Linus Foundation.

Interviewee:                Yeah, but almost of that is professionally paid people, and anybody who, you know, works on a major project, even if you do it for free, at some point in time, there’s probably gonna be a commercial entity supporting that. And that commercial entity is gonna wanna hire the top contributors to that project, right? So there is, effectively, a professional path, you know. Now, whether that individual does not wanna be associated with the commercial entity, that’s more of a, you know, societal or a personal issue that they deal with, there. But, you know, but back to the complexity thing, you know, and a little story, here, for you, you know, it’s great that we – you know, 20 years ago, only 15 percent of software was opensource, right, 85 percent of what was built was coded, 15 percent was composed.

Now, the statistics show that 85 percent is composed and 15 percent is just, you know, the glue layer that pulls it all together, so it’s a wonderful thing. But, and you talked about that trend of, like, “I’ll never have opensource in my environment,” but the natural conclusion of all this complexity is that we – you know, we are of the mindset that there’s gonna be a bit of a crumbling of the dev ops market as a whole. Because the software process and the architecture has generally gotten to be too complex for a human to reason about the impact of a change. There’s too many moving parts, right? It’s like asking somebody, “What will happen if I walk up to this ant hole and I stick this stick into it? What will occur?” you know, you just can’t predict it, right?

And so, the only way to reason about that complexity of change is that you have to give it to a machine. You know, machines have the ability to reason about those sorts of changes. So the natural evolution of – and the last thing I’ll say with that, too, on the complexity front is that the prevailing paradigm for how software teams should be organized, for the past 20 years, has been agile. And the fundamental premise of agile is that quick changes and quick feedback make for a happier team and happier end users, right? Well, if you take that to its natural conclusion, you know, you’re making changes all the time, you’re getting feedback all the time, which in theory, you know, gets everybody to be happy. But what ends up happening is that the team gets burnout, _____ _____ gets overload, and then, you know, it turns into this frenetic energy that is not sustainable.

So, agile is a fundamental part of the program, you know, that we’re gonna lead up to. And instead, we need to focus on a way to have sustainability of the development team, and that we remove the complexity by focusing on machines who do the automation for us. And so, I think that what will happen is, much the way that, in the industrialization, you know, we turned an agricultural society into a machine society, the same thing in the software industry is that we’ll go through a software industrialization, where most of the engineers who are writing code, i.e. they’re farmers of code on that, will turn into orchestrators of machines on that. And so, we’ll have to step back, as an industry, and we’ll be programming bots who will actually write the code, make the changes, give us the feedback.

And that means, you know, machine learning throughout this whole industry. And then, you know, retooling the software engineering community as we understand it, to become orchestrators of the system, as opposed to, you know, farmers of the code.

Interviewer:                You know, it’s funny, I remember speaking to Luke Kanies. Luke is the cofounder of Puppet. He said there’ll be a day when software writes its own software, but he thought it was 25 years ago, and that was maybe 5 to 8 years ago. Tyler, we gotta run, but here’s what I’d like to do, ’cause this topic needs a wider playing field to play on, and we need more opinions. We do a video show called Dev Ops Unbound. It’s 45 minutes to an hour. We put together a panel of people, anything related to dev ops is fair game. If you don’t mind, I’m gonna have our people reach out to you. Let’s get you onto Dev Ops Unbound. I’d like to get some proponents, opponents, let’s really kick this around.

Interviewee:                Sure.

Interviewer:                ‘Cause I feel terrible we don’t have enough time to really dig into this the way I’d love to, but I think we got people in the waiting room. [Crosstalk], hey, for people who maybe didn’t catch the beginning of this, where was your repository of the dev tools [crosstalk]?

Interviewee:                It’s at my Substack, so tylerjewell.substack.com. You see the Developer-Led Landscape, there. And also, you know, we’re active investors in this space, you know, come check us out at Dell Tech Capital.

Interviewer:                Yup, and Tyler Jewell, for those who wanna know, T-Y-L-E-R, obviously, J-E-W-E-L-L.substack.com. Go check that out. Tyler, we’ll be in touch with you. Let’s continue this conversation.

Interviewee:                Yeah, see you later, Alan. Thanks for having me.

Interviewer:                Thank you. A pleasure. Tyler Jewell, Dell Technology Ventures, here on Techstrong TV. We’ll be right back.

[End of Audio]