FULL INTERVIEW: Scott Dietzen - Former CEO of Pure Storage and Augment Code

Scott, I want to talk about leadership and board level perspective. You've been both a CEO and a board member leading and advising some of the most innovative companies in technology. Few people have seen both sides of the table as deeply as you have. What do you wish more CEOs understood about how to get real value from their board? And what do you wish more board members understood about the CEOs they work with?

I think first of all, when you step into the CEO role, it's the one job where you have to build both above yourself and below yourself. And I think you want to think carefully about trying to build a board that fills in your gaps or helps your leadership team. I love to have like a sales guru or a marketing guru or a finance guru on the board. Certainly as you scale, you need finance represented as well. And it's almost like your leadership team can have a board member coaching them in their particular function since no CEO knows all these different areas different deeply themselves. And then foster really open communication between your board and that leadership team to get full value out of your board. Too many CEOs try to be the conduit between the board and the leadership team. And I find you can get a lot more value with open communications. See the board meeting as an opportunity. Too many CEOs approach it like it's a proctological examination that they have to be scared about. Rather right. I'm egoless when I go into a board meeting because I'm really looking to make the company better, not try to defend my particular position. Never use a board to problem solve. This is one of the big mistakes CEOs make is they'll pose a problem and then try to get advice real time. To get the most out of a board. You want to do your own deep diagnostics on the situation and the issue and then have a plan go in with, here's our proposal for what we're going to do about this issue and let the board help you course correct rather than try to problem solve in real time because you can't get great value out of a board. And then lastly, remember, you're in charge if you disagree with your board and you follow the board's advice. If it turns out that you were right and the board was wrong, it's still your fault. So ultimately, you know, you've got to follow your internal conviction if you're sure that's the right path. I mean, take all the board's guidance into account, but you've got to be the one making the call.

I want to talk about being a CEO and scaling companies, you led Pure Storage through rapid hypergrowth and an ipo, an incredible case study in scaling culture, brand and technology simultaneously. When Pure Storage was growing fast, what was the hardest cultural challenge you didn't anticipate and how did you solve it?

Yeah, I think it's. It's really hard to establish and then maintain a great culture when you're, you know, we grew from, you know, 15 to 2,000 employees while I was there, and it definitely makes it difficult. One thing that's obvious is try to get everyone in the company focused on external lenses like how do customers view the company, how do partners view it, how does the competition view it? If everyone is kind of looking outside in rather than thinking about their own particular turfs, you know, and areas, you can avoid internal politics, which is a key part of building a great culture. Something else I used to do is insist on people not using I me when they're functions. Right. It's always we us. When you're speaking aspirationally about what the company needs to accomplish. The time to use I me is when you're taking responsibility for something that's broken, that has to be fixed, pass the credit, take the blame, is just a really powerful thing for managers to do to build a great culture. I do think it gets harder to maintain that culture as you hire middle management. You know, middle management tends to look at the company differently. If you're running a function, you know, your goal is to grow the whole pie. But sometimes the middle managers, especially if they're coming out of bigger companies, will tend to want to grow their slice of the pie. And you've really got to try to work hard to keep them focused on, you know, the overall goals of the company and not their individual turf. I love how AI is helping people increase span. You know, you look at Jensen at Nvidia with 70 direct reports. I'm not capable of being able to do something like that, but I think that's a great way. Something else is be really careful about early customers. Unhappy customers, I would say, are 10x to the negative what happy customers are to the positive. So you should have the conviction to say no or not yet to customers that you don't think you're ready to make happy. That can create friction with your salesforce. Although I found that customers will find ways to put you into other workloads if you're honest enough to tell them, you know, we don't think we're yet, they'll, they'll put you to Work in other places where you can be successful and you earn their trust in perpetuity when you have that conviction to say no. One other key point about maintaining culture is put customer feedback, partner feedback, in the loop, anywhere you can. We measured net promoter score on all of our sales and product support interactions and I found everybody behaved better. The score was being kept. It's kind of like the difference between an Uber and a taxi ride in your experience.

What's an underrated skill that separates good startup CEOs from great ones during periods of hypergrowth?

I think the critical thing is just recruiting. It's all about the leadership team you build as a, as a CEO that's running all the individual functions underneath. Now, I tend to favor up and comers, people that have a chip on their shoulder, that have had success but maybe haven't been proven out in the big job yet. And so they've got that supreme hunger, that extra gear in terms of how hard they're going to work to be successful. I worry about complacency. You know, people that come in that are a little too comfortable tend not to put the extra energy and drive in that you need because it's so hard to grow a company. You know, when my, my kids are in growth spurts, they, they need to sleep a lot and eat a lot because growth is just an amazing amount of work. You've got to try to stay ahead of it too. You know, there are going to be leaders that are great at a certain scale that will start to hit. And so you need to be able to hire bosses for those early hires. You really want to involve them in that process, get their buy in so that, you know, you don't alienate them. They get a vote for sure, but not a veto as you're trying to scale the team. And I would say be really careful not to get behind. You know, it's so easy to defer making changes when you're in the CEO role. You know, I like to push myself as soon as my bit flips that somebody, you know, either isn't filling the function well or needs to have more and oh, a boss because they're going to hit their scale. I want to do that sooner rather than later because you can get behind, you know, somebody departs or somebody else isn't successful. And then you're trying to do multiple executive searches at once. And I find I can really only do one good one at a time. You know, remember your, your role is, is not to defend your team, it's to make Your your team better. Not everyone can make an elite team. And so you've really got to be aggressive in making sure that you've got the best of the best, bringing everything together.

I want to talk about the AI industry and its future through your board work at Redis and augment code. You're right at the intersection of AI data infrastructure and developer productivity. How do you see the line blurring between data infrastructure companies like Redis and AI native companies? Are we heading towards full convergence?

I do see all of this coming together. AI is already changing software engineering. I think it's in the process of changing customer service. I think it's going to change all these internal business processes as well. And so you really want to be out in front of this whether you're AI native today or not. And even the AI native companies still do a lot of other tasks, more traditional ways and they really need to be thinking about how they can bring AI to bear across all of these different disciplines. I don't mean put it in charge, I mean leverage AI in order to get greater productivity out of the existing team. I think successful infrastructure companies in particular are going to embrace a new software architecture that incorporates AI into the software, which makes the software more flexible, more secure, more resilient. Hopefully you've got to be careful about that, but it's going to make software a lot better. I envision a world where we get to software abundance, where businesses can have all the software that they desire and that software will be easy to use and reliable. Most every piece of software disables of points in some way or another today. And I hope we can pay down all of that debt collectively in the years to come.

You describe yourself as a geek, outdoorsman and a dad. Grounding traits in a high intensity career. What's something you've learned from parenting that surprisingly made you a better executive or board member?

I think it's really delegation for me. Try not to control outcomes, but really try to empower the team. Both my kids came out fiercely independent and you know, wanting to get dad to, to fulfill whatever their whims are. And I was surprised by that. You know, I kind of expected dad to be in charge. The, the actual parenting experience for me for was anything but that. They saw, they saw dad as more their servant and so obviously you have to have guardrails up to try to keep them in appropriate lanes. And I think you lead most of all by example. I can't tell my kids what to do, but hopefully I can show them what and over time they'll follow that lead. I think that's so critical for CEOs, right? You want the team to work hard, you've got to work hard, be out at the front line, interact with the customers, talk to the engineers, understand what's going on across the team. When everybody sees you caring and deeply involved, just like kids, they'll do better work.

When you're evaluating a startup as a potential board opportunity or investment, what's the single signal that convinces you the team is truly special?

Distill it down to one signal. I think you've got to see, you know, smarts, you've got to see real hunger. And then I look for deep domain expertise as well as deep knowledge of the customer, of the customer pain. I think you need all four of those ingredients. You've got to be able to connect the dots between the technology you want to build and then how it satisfies that customer pain. You know, people talk about vitamins versus painkillers. Ideally you're selling a painkiller because it's a lot easier to sell a painkiller than it is a vitamin to corporations. So if you can deeply understand that customer pain and then how the tech can help alleviate it with the hard work and smarts, you're going to be successful. So I want to see all of those four ingredients roll together.

Do you think the industry is over indexing on model performance and under investing in data quality and integration layers?

I think while our models are rapidly advancing, it makes sense to keep the emphasis where it is on this general purpose AGI capable model. But I do expect these models are going to plateau unless we have some algorithmic breakthroughs in the months and years ahead. And when they plateau, I think the market's going to change pretty dramatically. On the enterprises. What I expect to win will be model specialization. You know, if you either start with an open source model and you add custom training on top of it, or you use distillation to take one of these larger models, I think we're going to have faster, more deeply knowledgeable models in particular areas of expertise. Right. In the case of augment having models that understand all of medieval history or know the works of Shakespeare, it's not particularly relevant to coding. So just like humans specialize, I think we're going to see a lot of model specialization in these enterprise use cases for consumers. I think these general purpose models will continue to be supremely popular. But in the enterprise, I think the world is going to move toward much more specialization and that's going to dramatically Change the AI landscape in the years ahead.

Scott, you have helped build and lead companies across several major technology shifts from weblogic to Zimbra, Pure Storage and now augment. You've consistently been ahead of the curve in identifying the next wave. What consistent pattern have you seen in how breakthrough ideas actually form?

I think part of it is I just gravitate to things that are very interesting to me. So I do like big markets, right. I want my work to ultimately have an impact in the world. So try to look at bigger problems or bigger market opportunities. I like deep tech. Right. The thing about deep tech is it's hard and then ultimately if you get a solution and that solution ends up being defensible for long periods of time because it is so tough to get it right, ideally you have a tech disruption involved. I had the good fortune to have a front row seat to the Internet tech disruption. I did my PhD research on machine learning and so getting a chance to see that technology finally coming into the market and changing the world is hugely exciting. There was flash memory, how it changed the software recipe for taking advantage of storage. So that tech disruptor is really helpful for changing the landscape on the incumbents because the incumbents are entrenched with customers. So you want something that's changing and then you need some innovative recipe that is going to allow you to, you know, close remove customer paint. It's not tech for tech's sake, right? It's tech that can be put to work to change things either for consumers or for enterprises to make them better. And then, you know, it goes back to the quality of the team. Do you, does the team have the knowledge and the drive and the hunger to be successful? When those ingredients are in place, I think you can do really well against the market. You know, people talk about 1 out of 10 startups being successful. I've had much better luck than that based on, you know, weighing those factors.

What's the most contrarian view you hold about the future of AI or the startup ecosystem that most people disagree with?

Yeah, I would say the. It goes back to the point I made about models specializing in the future. Everybody now is chasing AGI now. We want to get the largest, most capable model. And I certainly think that's a very interesting pursuit in general. But for enterprises, I think what's going to win out instead are highly specialized custom models, models that are very good at particular tasks. Again, just like we train humans for, they'll be faster and they'll be much easier to train and to post train if we Focus on that. So as we start to see models plateau because everyone has access to the same compute, largely has access to the same data, are using the same algorithms, I think that's inevitable, that we're going to see plateaus start to happen. Then the shift, the shift is going to come. It's going to be open source models with a bunch of post training, distillation with a bunch of post training. I think for customers, you know, it's going to be about making sure you can use retrieval, augmented generation to make your knowledge available to these AI models so you can put it to work. So that's, you know, the history of your code base and all your code reviews, or that's your customer support system and sales systems and all the interactions that have taken place, you can bring that knowledge to bear to better serve your customers. I don't think you'll train on that data, but I think that data needs to be available to the models and then we've got to come up with reinforcement learning that works in that setting to make those models better at putting your data to work in order to better serve your customers and partners and make your internal business processes more efficient.

Scott, you have had a high impact career. You've been a CEO and board member of some of the greatest companies in technology like Pure Storage and Augment Code. You have employed thousands of people. You are a father, you are an outdoorsman. When people look back at the impact that you've made, what do you hope that they remember most?

I hope I am thought of as a high integrity individual that really cultivated the fantastic talent that I've gotten to work with. I mean, I'm ultimately profoundly grateful to have been blessed to work with just extraordinarily talented individuals. You know, the work, the software work that had to come together to make products like weblogic and Pure Storage and Augment, the successes that they were and are was supremely difficult. And one of the things that's true in both a CEO and a CTO is we CEOs and CTOs get a lot more credit than we deserve and it's the teams around us that get so much more. There's what I most want to be remembered by, is, you know, the guy that lowered the average IQ when he walked in the door, passed the credit to the teams that did the work and deserve it.