Autonomous Software Development – Tyler Jewell, Dell Technologies Capital

This is texturing TV. Hey everyone, welcome back to another tech strong TV segment here today. I'm busy bus.

Y our next guest he's been on before, it's our friend Tyler Jewell Tyler's with Dell Technologies Capital. Hey, Tyler, welcome to Tech good morning Ellen drive to be back. Nice to have you man.

Hey, I you know I said you were from Dell Technologies. I didn't go through the whole title and everything. Why did you give him a little bit of your background?

I'm one of the investment managers at Dell tech Capital. I've been here for three years and previously I spent 25 years as a product operator in and around developer tools and devops and I've been an angel in DC investor for since 2005. Very cool.

So Tyler, you know, I don't expect our audience to remember necessarily, but the last time you are on Jesus how to be I'm gonna get six months ago five months ago something like that and you know you were talking about have we reached Petit devops in other words have we got to the point where we've rung out? as much as we're gonna get devops in terms of productivity gains and forward momentum and it's now maybe a case of diminishing returns if you will and and what that means for the industry right how fast you know, and why what would some of the reasons behind it and you were doing some research and at the time it was a very preliminary kind of report findings, but since then the report is now out and available the final version, correct? Yeah, that's right.

You can see the report on my sub stack Tyler Jewels sub stack we talked about it in three categories complexity Automation and autonomous development out there. Yep, and we're going to spend a lot of time talking about autonomous development today. But before we do, I just wanted to kind of backtrack a little bit and look when we spoke when we spoke six months ago.

You know, it was the the VC and well not only the VC the whole economy was still in that rip roaring. Dare, I use the word irrational exuberance of you know, people throwing around probably like crazy amounts of money, you know companies raising four or five 600 million dollar rounds. A billion dollar rounds and is like wow is this coming from since then, you know, we've had some corrections and there's a lot of folks in the valley who are panicking.

right a lot of company I met with a young CEO first time entrepreneur CEO A couple weeks ago. He said yeah, I just laid off my sales and marketing team and I'm like, yeah, I don't know if that's really a great thing to do. He's well, you know, we we need to cut the bird.

okay, you know, I mean the nice thing about where I said Tyler is I don't you know, it's not my it's not my Business to run it's their business, but I don't know if I'd fire my entire sales or marketing team. Anyway, so with with the with the change in economic conditions and the Outlook out there. What do you seen Tyler?

Right? How is this affecting kind of your view of you know, have we achieved Peak devops Etc? Yeah, you know so the the basic premise of what we concluded is that because of wage growth the due to inflation and you know consumer demand increasing at a skyrocketing rate for Digital Services that it's it's getting to be more and more difficult for the developer industry as we know it to meet the needs of consumer demand for all the digital services that need to be produced.

And so at the intersection of that you look at well geez, you know, that's great. That's great for devops vendors because devops vendors are fundamentally providing productivity enhancements and automation to make these developers more productive. The issue has been is that over the past 20 years, which we call the agile wave, you know productivity hasn't necessarily improved even though we've provided a lot more.

And tools the complexity surface area has actually increased because there's a lot more let's call it a process discipline that the average developer needs to learn about in order to produce software. And this is you know, Version Control continuous integration artifact and image repositories cybersecurity the complexity of you know, diversity of Technologies and protocols across the front end and the back end and so on and so on and so, you know with that, you know, as a result that increased complexity from just the tool chain that exists developers haven't become more more productive and we actually see some evidence of high attrition rates and that the developer industry is not growing at a you know, really a fast Paces. There's burnout Others High attrition the average professional developer careers, roughly 17 years.

There's some you know data that supports this which is a pretty short career life cycle associated with that and you know, and we just speculate hey what's going to happen now? When we look at the industry at large one, you know, we track all the different types of devops companies that are out there and you know, and we used to have only 15 categories and now we've got close to 35 different categories. So that's a lot of different types of tools that developers need to learn and more importantly the database we track it's 1400 companies in size now and two years ago.

It was only 700 companies in size. So it's doubled in the total number of companies and most of these companies, you know, we only track commercial companies. So they've all either one some early customers or more likely gotten VC bootstrap funding and wildly when you look at the amount of DC that's gone into those 1400 companies.

Um, it's in the tens and tens of billions of dollars and the amount of VC that's gone into that is is the let me say it differently the amount of exits in liquidity from all those companies either through IPOs or nice Acquisitions is a small fraction of the VC money that's gone in there. And you know, it's wild is you talked about the market Contracting if you look at all the software companies that like to buy other software companies in the devops space and you look at the markets capacity to have new IPOs, um or the capacity for companies to get to a point where they can be profitable. There's only about a hundred and fifty or so companies here that cross that threshold right?

And and so if you look at restricted Venture Capital that's going to become, you know, less that's such a capital available and Number of companies that can get Breakaway speed unfortunately, you know, there's probably 500 to 600 companies in that database that won't either be able to attract VC money won't get to an IPO scale or will be attractive enough to get Acquired. And so what happens to those companies if they can't if they can't get to profitability and they can't raise money they probably go away and so, you know, we're speculating that you know, like one third of the entire database could be at risk for bankruptcy over the next three to five years. You know, my experience has been.

They don't necessarily I mean some of them do go away. Some of them go away in their IP got salt right? I look I had a security company that we were very close to styling the one of the biggest security companies of the time and instead.

They bought the IP of this company that went belly up because it was cheap obviously right and wasn't great IP in my mind, but whatever and we lost we lost the potential exit because of it. You know, but but Tyler this isn't in my mind, I don't think this is new. I think this is kind of the cycles that we've seen in technology for the last 20 years.

Absolutely. Yeah, I mean we saw this in 2001. We saw it in 2008, you know, the cycle is normal and it's healthy.

You know, what's what's Wild is if you as a VC, we take a look at how much money did limited Partners Fund to VCS like in the previous year and you know in 2021 was to the degree or something like 320 billion dollars went into new VC funds. Um, and even if you take a 70% contraction on that, which is what we sorry, but which is what we saw 2001 and you know in 2008, you know down to roughly a hundred hundred billion dollars of new VC funds that were made available. It still only takes us back to like 2014.

You know, so so a 70% contraction in the VC Market is still only you know, you know, it's a post financial crisis level of funding. It's still a lot of money. That's a lot of money.

Yeah and look the fact is I mean from what I've been told and read this Stella a lot of VC some of these bigger VC, you know. Funds that we are you know organizations that we see out there the Tigers the inside so they're still sitting on a lot of money that has not been deployed. Well, theoretically fit theoretically, you know, so here's a while statistics.

So this year even with the market contraction in the first six months of this year VC funds raised almost 320 billion, which is equal to what they really last year this year this year. So this year they're gonna in the brazing more but having said that just because they raise the money doesn't mean that the dry powder is allowed to be deployed. Um in that, you know, so most DCS get their money from limited partners, right and you know, so whether that's institutional or family offices or wherever that may come from those that money flows in and those institutionals are kind of constantly recirculating their money into the VC space.

So even though you know, a major institutional goes to a V6. Yeah. I'm going to commit You know a hundred million dollars to your to this fund that you're raising the VC still has to do a capital call when they want to make an investment and you know, and it's pretty easy for The Limited partner to sit there and go.

Well, you know, I'm watching, you know, yes, I committed a hundred million dollars to you. But you know, I you know, I don't think it'd be a good idea for you to do the capital call meaning I have to actually give you the cash because you don't give the cash up front and you sit there's like hey, you know, I just kind of want to wait three to six months to see how everything else pans itself out here. And now look the VC can make the capital call and the lp is forced to give the money.

But when that VC needs to raise his next fun and he's looking for that that person to commit another 100 million dollars or maybe he goes shopping somewhere else. So even though that the dry powder is out there doesn't mean that the capital calls are going to be allowed to go through so it's a it's a it's a weird dynamic that plays itself out. Yep, something will be watching Tyler.

I want to Pivot a little on you. How's that for a word with the VC? I want to Pivot a little on you and talk, you know, one of the things in the report is this third wave if you will, I think you call it of autonomous.

software development Okay, that's if you don't mind. Yeah, the the basic premise is that the soft, you know software is an industrialization cycle, right? You know, it's a creation of software is the type of supply chain that's out there and the devops industry organizes itself around a development Paradigm and you know in the 80s and 90s and early 2000s.

That was really a waterfall, you know, it was a mindset of you design software in three to six month increments and there was a lot of reasons why waterfall was the predominant Paradigm and what ends up happening is the industry and the tool chain kind of get developed up and around and supporting that Paradigm and for the past 20 years. It's really been around the agile Paragon this basic premise of you know, the faster you ship something the quicker you get feedback the faster you can react to that and you create kind of a, you know, a virtuous feedback loop. And and you know, we've created hundreds of billions of dollars of market value in the tool sets, you know these 1400 companies that have been built up in and around this agile Paradigm.

The issue with that Paradigm is that you know, the basic principal around Pete devops is hey, you know agile is not sustainable and that it promotes faster changes faster feedback faster changes faster feedback and you get on a vicious a vicious flywheel and humans are unable to sustain themselves and producing faster and then taking, you know broader sets of inputs. So what happens like if you play this out to Infinity, you know, you end up with people either burning out or they're not able to participate in that sort of supply chain. And so so that would pretend the potential in of the agile way now and we basically said, hey, we think it's about 20 to 30 years at the agile love will play it out.

We'll see a lot of consolidation, you know people We try to do these Mega agile platforms, but at some point in time you realize that humans are not going to be capable of moving faster. So we need to hand that over to robots to do that work and there's a large portion of the software process which are tasks that can be completely automated and managed by robots on top of that robots are able to reason about changes and very complex systems at fast speeds much better than humans can do so and the impact of those changes and there are also able to operate on mass quantities of data to be able to interact on the feedback. So I wonder what I want to stop you for a second.

Yeah. Let's make clear to our audience when you say robots. Well not talking about Robbie.

Or this from Lost in Space or that or an industrial robot building cars or something? Yeah. These are software robots right over automots right that are you know, think of it as a a software pair programmer, you know, that's sitting alongside the developer that's there.

Okay, you know and we call this, you know, this would be autonomous software development which would be the construction and maintenance of software with less human involvement or without human intervention on that front and And and you know, and we we pretend that there would be a completely different architecture and that you know, what would really drive this is you would have large bases of data and developers would actually transition away from being a construction specialist and into people who describe the intent of a software system. My job is to describe the intent of how the system should work and in the robots work off that intent and massive amounts of data that's coming in to construct deploy and monitor that software and then make changes as either the feedback comes in or the intent changes. Okay.

Got it. Yeah, absolutely. You know, it's funny.

It's not just software development security too. I actually had a conversation I speak to a lot of security people obviously devsecops that kind of thing and to them, you know the ultimate win here. To make devsecops better is to take the humans out of it really just to automate the heck out of it so that it could move at this as fast as we can develop and deploy that's how fast, you know, the security works with it.

And in order to do that. We need to take the human element of Out of the out of the price. I mean humans are prone to errors humans are you know even innocent innocent mistakes, you know repetitive, you know, repetitive tasks produce, you know, simple mistakes if it's a repetitive task.

A humans just never gonna do that as well as a robot does aroma she's gonna do anything over and over. Other I think. fact to hear and that is you know, one of the things that I see in devops Tyler is that we've actually put more on the developers plate it seems Right.

Oh, yeah, the developers responsible deployment. They were developers responsible. Make sure it does things done the developers responsible for security, you know, so we were taking a hard job making it harder this comes back to the complexity argument, you know as a developer really more productive and you know it it takes many years to Um, established proficiency and certain areas of the development process, right?

Whether it's unit testing or integration testing or maybe security analysis, you know, even deployment or infrastructure is configuration. So you pile all those on to full stack developers who need to be generalists on all those different things. It's very difficult to acquire that proficiency and even more difficult to maintain it because it's not the Technologies constantly changing on you and your tool chain is so huge that you have to be trained on the tools itself.

Right? So, you know, the level of retraining and retooling is massive first today versus where it was 20 years ago. A great I don't agree Tyler Webb.

Well, we're overtime as usual here. But can you just give us the sub stack address again? Yeah, I go to Tyler Jewel sub stack or look for the developer LED landscape and that will get you right there.

Yep, and it's Tyler t y l e r obviously and Julie's j e w e l l. That's right. Tyler man, hey, thanks for stopping by and bringing us up to speed here.

Keep us abreast of what else you thinking. You know, you always have an open invitation. Thank you Alan.

It's good to be here. All right, Tyler Jewell Dell technology Capital. We're going to take a break here on Tech strong.

We'll be right back.