Opening Remarks — Day 1
Gene Kim opens the Enterprise AI Summit at Bletchley Park, arguing that AI is rewiring technology organizations much as DevOps did, but far bigger and faster. He shares what leaders at the frontier are learning, previews the experience reports and expert talks ahead, and invites attendees to share what they learn. Ben Grinnell of Newton explains how the event came together and covers logistics and feedback.
Chapters
Full transcript
The complete talk, organized by section.
Host Intro (Gene Kim)
Good morning, ladies and gentlemen. This is a special bulletin.
Earlier today, a coded message was intercepted. Its origin, unknown. Its contents, encrypted.
The message was brought here to Bletchley Park, where our codebreakers set to work at once, and I am now able to report the code has been broken.
The message reads as follows: Please welcome to the stage researcher, author, and vibe coder, Gene Kim.
Gene Kim
Good morning. Good morning, everyone.
I cannot tell you how excited I am to be here at Bletchley Park, the home of AI. I'm so delighted that you're all here, and we have such an amazing two days in store for you.
I think I can venture a guess, and I believe that we are here because we believe that something momentous is happening right now. I think we believe that AI, specifically agentic coding in the technology leadership community, is upending how technology work is performed.
And for some of us, we have felt this before in the DevOps movement. And like so many of you have observed, it's going to rewire technology organizations, but unlike DevOps, it's going to happen 100 times larger and, more importantly, 100 times faster. We believe that AI and agentic coding can massively elevate our ability to achieve organizational goals.
And I think we are very aware that there are many looming questions that we don't have great answers for, just to name a few. Like, what happens when teams of two can do what typically would have required eight? What do you do when the code review processes, the processes we set up—it's just not known how we can get assurance that we can actually put code in production, put them into consequential settings, and have some assurance that it'll actually work as designed? And there are many more.
Our friend Kent Beck, who many of us know in the technology community—he's associated with Agile, unit testing, extreme programming—he said every process, every technical practice, every ritual we've created were all made due to economic trade-offs. So every decision that we made in the last 50 years is now upended because none of us really envisioned an age where coding is becoming instantaneous and free.
And yet, we know that the reward is out there for us to use these technologies, even though we have many questions that we don't have great answers for yet.
But one of the most startling things I've heard came from a gentleman. He is a group CTO for a Canadian private equity firm, and he said we are now probably facing the greatest leadership challenge maybe ever. And I think that was a huge claim to make. So I asked him to walk me through why he believed that to be the case, and in 15 minutes, he absolutely convinced me that this is going to be an incredible leadership challenge.
And so here was the logic that he presented to me. He said for at least 60 years, a technology leader could just inherit a process that worked well enough, and if it worked, they could take credit for it. If it didn't work, they could make some minor changes and take credit for that. But we are now entering a world where every process we have is now profoundly insufficient.
And you're going to, over the next few days, hear about once you increase the development process by 10x or 100x, you break everything around it. And so what we're going to see is a rewiring of organizations that doesn't just stop at the technology boundary, but potentially all the way up to the CEO. And so I think his claim that this is going to set up the conditions for the greatest leadership challenge, I think that's right on.
And so for those of you who read The Phoenix Project, The DevOps Handbook, you will know that DevOps reshaped technology organizations. Some of you might not be old enough to remember an age, say, in the mid-2000s, where we deployed software once a year. And it was not so good, right? It created chaos and disruption across the organization.
And so many of us remember in 2009, two gentlemen named John Allspaw, Paul Hammond, they were the VP of development and operations at Flickr, respectively, and they said they were doing 10 deployments a day. And so by all firsthand accounts, the people who were in that room were throwing up in the aisles because it seemed so reckless, so irresponsible, even immoral. Like, what sort of maniac would do 10 deployments a day when you didn't need to?
And yet, this is now something that is very commonplace. And so you can imagine how much that changed work for development, QA, operations, information security, product. Everything changed about how we worked in technology organizations as we go from one deployment every year to 10 deployments a day.
And I would just like to state that what we're going to see over the next few days is the same thing is happening, but again, probably 100 times larger and 100 times faster.
So I just want to share with you a brief story about how I got introduced to this. This is a gentleman named Steve Yegge. He famously spent 10 years at Amazon and Google. He was a part of the first two-pizza team working for Jeff Bezos, then CEO and founder of Amazon, to help reduce customer contact.
And so I met him in June 2024, so over two years ago. And so we had many things in common. But one of the things that we both believed was that AI is changing coding and development from underneath us. And so we ended up writing a book together called the Vibe Coding book. And that has been an adventure like nothing I've ever been on.
And one of them was having the foreword written by Dario Amodei, CEO and co-founder of Anthropic, whose F-1 filing just hit last week, which many people have been seriously studying. But I love the way that Dario Amodei framed the notion of agentic coding.
He said vibe coding is an inspired term but also misleading. He said it's inspired because it evokes a whole different way of coding, right? That you're not hunched over a laptop using your meat fingers to actually type in code. Instead, code comes out through a series of conversations with AI, and AI is writing the code for you.
But he also points out that vibe coding is misleading because it sounds jokey and unserious, maybe even frivolous. But he said at Anthropic, where they pride themselves on having some of the best developers on the planet, make no mistake, it is the only game in town. And that was back in the ancient days of 2025.
So it is just a time of incredible change. And I'm going to evoke Kent Beck again. He said for most serious people, the safest answer to most questions is nobody knows, right? So all these questions that we have, the safest answer is nobody knows.
So one of the people that we cite in the Vibe Coding book, his name is Dr. Erik Meijer. He's considered by many to be the greatest programming language designer of all time. So if you've used Visual Basic or C#, LINQ, Hack, and so forth, he was either at the center of that movement or he was close to the center.
And there was a phrase that he said in a conference a year and a half ago that I just found so startling and even shocking. He said we are probably going to be the last generation of developers who will write code by hand. And so this is a very kind of funny and startling thing to hear for someone who spent his entire life dedicated to making programming languages for humans, right? Why would he say that humans won't be writing code anymore?
And he said time is our most precious resource. It can neither be stored nor generated, right? And so if we can find a way to save time, it's this moral imperative that we must take.
So that was certainly an aha moment, but the second aha moment came from this gentleman here, Dr. Tapabrata Pal. He was famous for his work at a US credit card issuer called Capital One. He started the DevOps movement there back in 2012. And so he was the last enterprise architect to still have an IDE coding environment on his laptop. And he was part of the movement to bring engineers back into the organization. They had outsourced 80, 90% of the developers, brought them all back in, and now they have something around the order of 17,000, 18,000 engineers.
So during his time as VP of architecture at Fidelity, one of his responsibilities is to own the software supply chain to make sure that every dependency is secure. So he would be the person you go to, to ask which applications—the 30,000 applications at Fidelity—which ones have Log4j.
So he's owned this application for two, three years, and he's had this vision of what it should look like. But whenever he asked his development team, “Can you do this? It should look like this.” These are all backend Java engineers. The answer was always, “Sure, but it will take nine months, and we'll have to hire two frontend engineers.” He was so frustrated that he spent five days vibe coding it, and it went into production a year ago.
And I think this was the first glimpse that we can use these crazy new technologies and put them into a mission-critical enterprise context.
And so over the series of the last two and a half years, we've been putting together experience reports of technology leaders who are at the frontier bringing this amazing new technology into some of the most consequential missions. And I think they'll be the first to admit that we still don't have great answers to many of these questions, but the reward seems so high that these are the leaders who are taking those organizations there despite that. And hopefully we'll answer some of these questions.
So I'm going to share some of the highlights for me from this April 2026 conference that we held in San Jose, where we had organizations from the tech giants, Facebook, Amazon, Netflix, Google, Microsoft, but also OpenAI, Netflix, as well as a startup, as well as a 185-year-old tractor company.
And so my friend, Amy Willard, said we all got rebooted at the same time, right? Basically, as we are trying to figure out what are the new ways of working, it doesn't matter whether you are at a frontier AI lab, at a tech giant, or at John Deere. Again, she said, “The best answer to most questions is nobody knows.”
And so one of the things that Amy presented was how at the very top levels of leadership at John Deere, they're encouraging every frontline worker—not just developers, but everyone working on the production plant floor and the dealerships now—to use AI as part of their daily work. And they have over 90% engagement across all employees.
And the reason why I thought that was so amazing: two years ago it was viewed with suspicion, right? Why would I want to use AI when they're just going to automate my job away?
And so my friend Steve Yegge, who co-authored the Vibe Coding book, made an observation that said, “Wow, John Deere is ahead of Google.” He made the observation that most people at Google weren't able to use the best coding assistant tools, that adoption was actually quite low. And of course, that makes sense when there's something like 40,000 developers at Google.
And so I guess there's two ways you can interpret this. One is, look how good John Deere is. The other interpretation is, look how bad Google is, which is, I think, how most people chose to interpret his tweet.
So one of the people who responded is Sir Demis Hassabis, Nobel laureate, CEO of Google DeepMind, and told my friend, “Tell your buddies at Google to do some actual work instead of spreading absolute nonsense.” And I thought this was just an amazing interaction to show how completely charged so many of these topics are.
So over the next two days, we have two very special types of talks, continuing the tradition of 14 years holding technology leadership conferences.
The first type of talk you're going to hear are experience reports. Experience reports are basically where leaders will tell you, here's the organization we compete in. Here's where I fit in. Here's the business problem we set out to solve. Here's what we did and where we started, including tools and techniques so that they can be replicated, the outcomes, and challenges that still remain.
And so I love that because it very much matches the scientific method where you state a hypothesis, right, and then they share what actually happened. And so this is a way that we create a community of leaders that help other leaders. And you're going to hear many of them over the next few days across so many interesting industry verticals.
So the other type of talks are the expert talks. These are people who have subject matter expertise that will help get us from here to there. And again, we have amazing talks over the next few days around so many interesting bodies of knowledge that we think are relevant for all leaders.
So I am going to share with you this. I think we're in the most incredible time of distributed learning. In an era where the answer to most questions is nobody knows, the best thing you can do is experiment. And the only thing better than labs, to quote Amy Willard from John Deere, is to hang out with other leaders who are experimenting.
And so I think part of our aspirations is really to create a global, single lab notebook where we can record what people are doing, the experiments they're doing, and what the outcomes are, so that we can help answer some of the questions I painted before.
And my friend Andrew Ramirez from GitLab, he said, “That's not good enough. That's not fast enough.” And what would it take to get a community of 1,000 technology leaders within a year all sharing notes? And it's been an exhilarating ride to get those running.
And tomorrow, I'll share with you more about what we're doing with the Software Engineering Leadership Forum. But we've got eight virtual sessions, two in-person roundtables. We're elevating some of the top problems that we think the community is facing and coming up with answers that we believe are a good starting point.
So with that, I'd like to turn it over to Ben. It was Ben Grinnell's idea to bring this conference back to the UK. And I want to thank Ben for everything that you've done to help make this event happen.
Ben Grinnell
Thanks, Gene. And thank you, everyone, for coming. Can't believe we've actually pulled this off.
Gene Kim
Me too.
Ben Grinnell
So just to give you a bit of background, my journey with AI was started 30 years ago. I was a student at Loughborough University studying political control, and a guy called Kevin Warwick came in with seven dwarfs, which were eight-node neural networks. And then when you look at these—
Gene Kim
What year was this?
Ben Grinnell
1996. And the seven dwarfs were all moving around and being called Dopey and Bashful, and they'd just been trained in different environments.
If you look up Kevin Warwick, he's a professor of cybernetics. He was the world's first cyborg as well, because back in '97, he connected his nervous system to his wife's through a chip in his arm. I think they're still married.
Fascinating guy. And I looked at that and thought, “What if you could scale that to the size of the human brain? That'd be amazing.” And I've been waiting 30 years, and we finally got there, so that's been exciting.
The other thing I did in '96 was went to the IFAC Control Conference and sat in a self-driving car, and that's taken far too long as well. So I'm very grateful that you're all driving adoption in your industries and companies because that's the speed we need now.
I then spent 20 years building a consultancy from a startup to a 2,000-person business, and retired in 2024. And I went around the world and crossed a few things off my bucket list. And I was sitting in the Galápagos Islands and a bit annoyed about an email I read because Gene and Steve sent me a manuscript of the Vibe Coding book. So having played with the sea lions and things, in the evening, I was reading this book and thinking, “Why the hell am I retiring? This is far too exciting.”
I was also having a conversation with Steve at Newton and came back from there and started the job at Newton, which has been a fantastic 15 months now of a rollercoaster ride of really adopting AI across an enterprise and really helping a lot of our clients do that. So that's been fantastically exciting and has led to this event.
I started working with Gene in 2014. I went to the first inaugural DevOps conference, as it was back then, in San Francisco. And I was struck by the learning community that he was building. It felt very different from every other conference. There was no commercial element to it. It was really just let's get people together and learn.
And I remember looking at Target and Nordstrom, the IT leaders sitting in the forum discussing how they were going to improve their businesses, even though they're roughly competitive. And it was great to see that sort of just learning community, and that's so important. And the subject might have changed from DevOps to AI, but the environment, the learning in the community, and the ethos hasn't.
So that's why I was really pleased to say, “Gene, we can do this in the UK again and bring this conference back.” So it's been fantastic that through my position at Newton and through knowing Gene, I've been able to do that. So really excited.
And thank you to everyone who agreed to speak. Chatted to all of you, so if we can't do the photos, Gene and I have chatted to everyone. I'm so excited about how much I'm going to learn the next two days because I've heard snippets of those talks, and it's going to be fantastic.
So I hope you all enjoy it. Just got a couple of housekeeping bits to do. I think—conference agenda there. Access Wi-Fi there. I'll leave it up for a couple of seconds. Please make sure you're logged on so you can use your phone throughout the event.
One of the most important things in a learning community is feedback. Feedback is a gift, and sharing is caring. So every time you hear a talk, and it might feel a little bit like, “Oh my God, I'm learning so much,” please just log on and do the Uber rating for the speaker because that's how we learn which talks to do next time and which ones are really resonating with the audience.
So it's not much of an effort. I know it's an ask, but please make sure you give the feedback. It'll look like this. Very easy. It's even easier than Uber. There's only four—three stars. But it'll be very valuable for us next time we try and do something like this.
I think that's it from me.
Gene Kim
Thank you, Ben.
Ben Grinnell
Thank you, Gene.
One membership. The IT Revolution library.
IT Revolution's books, the video library, the courses, and the MCP. The papers stay free.

