**Matt:** Matt Thompson (Google Cloud Developer Ecosystem Lead) **Ricky:** Ricky Robinett (Google Cloud Global Developer Marketing Lead) **Host:** I noticed you are into deep sea photography. **Matt:** I am. **Host:** I'm definitely not as cool as you. **Matt:** Same. (Laughs) **Host:** That's cool. I'm actually wondering, but I could imagine... what does it feel like when you see, like Gemini 3 Pro, nails a complex coding task? Would it remind you of the deep sea experience? So to answer your question, really cool things I see in code do reflect nature. Matter of fact, we were just talking about your Street Fighter AI experiment. The learning curve that you showed is exactly what happens in nature. And we see that when diving. You'll often find a sea creature outside of its home area—either because the current's taking it or something's chased it. It's either going to survive or die. And it's all based on how fast it learns. So for me, what I see in what you built is what I see in nature. It's adapt and survive and grow. So, very similar. Although I think I like diving more... but [AI] is very cool. **Host:** So for Ricky... if Matt is the guy diving into the deep and bringing back the pictures, then you are the one who is responsible for explaining to the crowd what the values of the pictures are. So, what is the job of an evangelist like? What's the difficult part? **Ricky:** Yeah, I was going to say, Matt's risking his life in the water and I'm just sitting on the beach waiting for him to come back like, "Looks good!" (Laughs) I love Matt's point on human beings being created for something. And I think that ==throughout history, human beings are created to create, to tell stories. And I actually think your underwater photography is such a great parallel. Because when I see the photos... I see the photo, but I also want the story. I want to understand: Where were you? Is it a rare photo? Was it hard to get? Oh my gosh, you had to fly 30 hours and take a boat...== And so I think for the same thing with developers, it's being able to show them those stories and make them the hero of those stories. Honestly, talking about our products or our tools isn't the thing that gets us most excited. What gets us excited is: "Oh my gosh, I met a developer, they built a Street Fighter reinforcement learning [model]!" And like, I'm going to tell that story all over the place. And I think just collecting as many of those stories and celebrating as many of those developers is really what my job is about. I feel so lucky to do that. **Host:** Thank you. So you think human's job is storyteller? **Ricky:** I think that is the unifying thing that makes us human. I could talk all day about this... but my thing is magic. What I love about magic is really you're trying to show someone a thing that feels impossible. And you're trying to get them to embed a story in their head that could never have happened. The exciting thing about all this technology is it *feels* impossible, but it isn't. And those are the best stories. When someone sees it and they're like, "Well there's no way I could do that," and you say, "Let me show you how." **Host:** Back to the practical part. The advancement of Artificial Intelligence has raised a lot of discussion about "one-person team," "super individuals." Have you guys ever encountered some of... **Matt:** (Points to Host) He's one. **Ricky:** (Laughs) Yeah. **Host:** Tetris... I saw your Tetris game. **Ricky:** Yeah, yeah. I will shout out someone though, I think this person is a great example of what I think the future looks like. There's a company called *Supermemory*. The CEO—the kind of person who did all of it—his name is Dhravya. He is incredible. I got to work with him and was just so inspired by him and just... able to do so much as an individual. Now, I think there's something there, but the thing about Dhravya that I think is so true of what we will see is: ==[he is] such a great community builder too. And he inspires others. He wants to encourage other people to see what they can do. He wants to bring people along on the journey. And so, in some ways I think people can go farther by themselves than they ever could, but I think the most fun is to do it together. And I think Dhravya is such a great example of someone who... yeah he was the solo founder, but is bringing everyone along on that journey and lifting people up on their own journeys.== **Matt:** From a Google perspective, we have a very interesting view from our own internal engineering teams. One of the conversations I get to have with customers all the time is... they come visit our campus and one of the questions they always ask is: "How does Google build product? What can we learn from that?" And with AI, it's the idea of taking a 10x developer and having *everybody* be a 10x developer. So every developer inside Google is now using AI to go not just faster, but more efficient and with less bugs. So about 30% of all the code that Google produces today is generated using AI in our products and everything else. So, "Super Developers" exist, I agree. We kind of look at every developer inside Google becoming that 10x developer. And then there'll still be those few *100x* developers that, as you described, are just amazing in what they do. **Host:** That raises some anxiety, especially for Computer Science students. What should they focus on to stay relevant in this era? **Matt:** I have two kids. I have a daughter who just graduated in Computer Science from university. She now wants to go to law school. In part because of this challenge, and in part because she's very interested in the ethics of AI. I have a son who's halfway through university studying Computer Science / AI. And he thinks about this question all the time. So, he's decided that if he's in the very top of his class, there's no fear. If he's at the bottom of his class, it's going to be hard. I think that's probably true. I was having a conversation with somebody else today... "Will there be as many developers on the planet in five years as there are today?" I actually think there will be *more*, but I think they'll be different skills. What do you think? **Ricky:** I think that is true. I also think everyone's going to have to reflect on *why* they build. If the 'why' for you building is like, "I needed a job," and that's totally fine... it's a scary time, understandably. I think a lot of people have anxiety about that. ==If the 'why' you build is like, "Hey, I had this idea that I believe needs to exist in the world," like... the more people that can build, I think there are kind of limitless ideas that can exist in the world. And so that's not going to stop. If the 'why' you build is like, "I feel this innate desire to create things," that's not going to change.== And so, you know, what's going to happen with the *profession* of software development is a thing that everybody is trying to figure out. But what's going to happen with being a *developer*? I think there's going to be way more people who can bring their ideas to life, and that can only be a good thing in the long term. But... to the same point, the people who are in the top of their class looking for a job, it's going to be great. But even if you're in the bottom of your class, but you build a bunch of stuff and be creative... no one's going to care [about the grades], they're just going to see what you built. **Host:** Like you mentioned, a case where a guy is just so great at utilizing the power of AI and builds everything, and he is more devoted to building the community. Maybe connections... **Ricky:** Yeah. I think the role of humans... A.I. is to understand the impact, the ethics. I think that's really important. By freeing up *how* we build it and accelerating the pace, that gives us more space to ask *why*. Why we built it? Should we build it? I also think... I've worked at startups, I've worked at big companies. All the same problems with stuff we build, which is: ==You spend all this time building it, and then people may not want it. And so I think the role of the human is understanding the MVP (Minimum Viable Product), understanding product-market fit, understanding the feedback loop. And all of those things are, to me, some of the most fun things too. So yeah, I think the human part becomes more about the engagement, the feedback loop on it.== **Matt:** In our role, we get the opportunity to both speak to a lot of developers around the world and also understand some of the research that happens. And some of the most interesting research, you can go all the way down to children. So when one child plays with a toy, you'll see some creative ways that the toy is used. When you add a second child to play with the *same* toy, you'll often see tension, but even more creativity—more than 2x. And so if there's a multiplier effect as you add people to the network, you're actually adding creativity as people come together. ==People ask me all the time, "Why do developers like to engage with people?" Developing is often a solo experience. Headphones on, in the flow, I'm creating something. And yet they're very social people. When the headphones come off, they want people to be around them. That experience is because the ideas that get passed back and forth is more than X times the number of nodes.== So, I don't know that AI will ever replicate that. A matter of fact, I'm betting against it. ==I think AI is great at following down a path and making it easy. But picking an entirely *different* path and deciding "let's go try that"—that's a human endeavor.== That's what gets me excited. Because then I get to go... I know the number of ideas, if I sit with five other developers, the number of ideas will come out of a half-hour discussion beyond anything I can imagine right now. **Host:** I really like your metaphor. Maybe AI is more of an "Information Organizer" and humans are the ones who invent new ideas. Actually, one of my teachers in my graduate school... he said he loves to study his child. Like, he is specialized at Artificial Intelligence and he likes studying his child. **Matt:** It's a massive parallel processing unit developing right in front of your eyes. Yeah. **Host:** And maybe that's complementary to what we have in Artificial Intelligence here. **Matt:** Very much so. **Host:** To the models... we always say "Talk is cheap, show me the code." But recently we have seen a lot of new models and everyone is claiming them to be the state-of-the-art. So it's somehow becoming like tech specs are cheap. So what is Google doing to make developers actually believe that something like Gemini or other AI tools are really good for their daily workflow? **Matt:** It's an excellent question. And so you said "Shut up and show me the code"... the corollary to that is "Code is Truth." It's something that I've always believed. I've had technology leads come and tell me, "Matt, stop explaining. Code is truth. Show me your code." From an LLM perspective, there are obviously all the benchmarks that everybody looks at. And one week we're number one, next week maybe somebody else is number one. But there's actually a usability piece here. One of the big things that we think is different—and Ricky showed it with the CLI (Command Line Interface). The Gemini CLI is a very simple way to get access to a world of capability in a developer-friendly manner. And yet something that simple has opened up so many different new ideas... from Ricky using it for his email, to creating Tetris. It's incredible. That's what we're actually trying to open up. So I'm actually going to argue: We care deeply about the capability to generate code effectively. **Ricky:** I think that's a really good answer. To me, when I talk to developers—you touched on the anxiety—there's anxiety, there's excitement, there's confusion. And I think our approach is: Hey, we're not here to tell you what you *should* do. Every company is going to say they're the best. What we're really focused on is how do we help you learn? How do we empower you to use this stuff? And that could be our models and tools, it could not. But my goal is really, as every developer is sitting down and saying "How does this fit into my workflow?", that we can meet them where they are. Whether that's literally like we're doing in Beijing and meeting them where they are, or it's virtually over video, or it's through written content. But I think for us, if we can just help everyone navigate this, the tech will work itself out. I just don't want anyone to feel like they got left behind. **Matt:** And I will say, one of the things that I've really enjoyed about Ricky bringing this message is... it's a very inclusive message. And so we started with the ideas that, "Hey, I'm coding, I've got my headphones on, it's all me." But then you've got this piece that it's really about you and all the engagement you have with everybody around you. And I think... I have a son who is very much "this is all about me." But even he is starting to see that as he engages and shares ideas, there are new ways of approaching it. And I think that's the unique learning that humans have. **Host:** I feel the same. I'm that sort of "that's all about me" kind of coder. But what I have been experiencing is that when AI saves me a lot of time... maybe just one day I finish that project... then I put my headphones off and I gather my friends and say [I share my work]. But before, I didn't have the time. **Ricky:** You're showing all these people through content, through the things you build. [You] are the ones who are the most humble about it, right? I think that's really beautiful too. You could be like, "Yeah I totally get it, I'm awesome, I'm making a video, I'm thinking big." But you're like, "No, I wish I could do more, I should do better." I think that's so beautiful. **Host:** I think I hear a lot of human values from you guys. Ricky is more about communicate, communication, help each other. And Matt is about... AI can do all the acceleration, like the librarian, but inventing new things... that lies on the hand of ourselves. We need to feed the AI technology. So at the end of our conversation, do you guys have some golden advice for like 10 years later? **Ricky:** I had a boss who had a mantra of: "Enjoy yourself." That mantra came from a song and the first part of that song said: "Enjoy yourself, it's later than you think." For me, I do all these things because I love doing them. I am sitting here with Matt, I am sitting here with you, because there is nowhere I'd rather be right now than having conversations with you. And I think if everyone was able to find that *thing* in this moment, in 10 years it's hard to imagine not being happy that you made the decisions that put you in the places you want to be. I had someone ask me: "What do you say to someone that's not going to use AI? How do you explain to someone that if they don't use AI they're going to get left behind?" I was like, well I don't know if that's necessarily true. That assumes we're all going to the same place and in the same direction. So I think understanding where you want to go... if you want to lean into the AI stuff, we're having so much fun and we would love to help you. If you're like, "Hey, I want to be organic, human-created code 100%," I think there's still going to be a place for that. And if you're the best in the world of "human organic code," you are going to be one of the most valuable developers in the world. So I don't think there's a right answer. I think there's an answer for *you* and the more you can find that, the better you'll be. **Matt:** I want to go trademark "Organic Code" because you know how you go to the store and there's organic chicken? We've now got organic code. (Laughs) I think I can sum this up really in a short period of time. I hope. I'm so excited about the *now*. And I think back over my understanding of history... in a deep level probably goes back 100 years, and then a little understanding going back a thousand years. So when I think 100 years in the future, this will be part of the history. And history has shown over time there are parts of quick acceleration—the exploration of the world, when the Chinese traveled all of Asia in sailing boats, or Western Europe traveled across the Pacific. Massive acceleration of discovery and understanding. We're right at the beginning of that again. This is... there was physical exploration, and now we're going to be in metaphysical exploration, whether it's the sciences, whether it's math. I'm very excited for the type of science problems that will get solved in the next 5 to 25 years. I'm very excited about the type of opportunity services that people create, the type of communication we have. And then ultimately for me, I believe that the opportunity to explore the rest of the universe will be tied to this period of time as well. I don't know how, but I think AI leads to our ability to get outside of our own galaxy eventually. And that's hugely important. So the history books written about today will show 100 years from now that *this* was that time. And that's what excites me. **Ricky:** In 10 years you and I are going to be floating in the ISS together exploring space. **Matt:** I love that. That's awesome. **Host:** Yeah, I think that's a really exciting picture. And if we want to be ourselves, whatever we like, we can do that. That's the golden advice. So thank you guys for the conversation. I really appreciate this opportunity. **Matt & Ricky:** Thank you very much. Yeah. Thank you.