ENFR
8news

Tech • IA • Crypto

TodayMy briefingVideosTop articles 24hArchivesFavoritesMy topics

Sundar Pichai on Agents Replacing Engineers, Google's Future, AI's Flip Phone Moment, and More

AIRowan CheungMay 21, 2026 at 03:59 PM20:10
Audio player
0:00 / 0:00

TL;DR

AI is rapidly reshaping content creation, software development, and daily life, with major platforms emphasizing human creativity, personalized assistants, and agent-driven automation as the next frontier.

KEY POINTS

AI-driven transformation of content creation

Advances in tools such as Veo, Omni, and multimodal systems are enabling users to generate and edit video, images, and text within unified workflows. This shift lowers technical barriers, allowing individuals with minimal resources to produce high-quality content. The trend mirrors earlier disruptions like CGI, but with a stronger emphasis on augmenting human creativity rather than replacing it.

Creator-first strategy in evolving platforms

Despite the rise of AI-generated media, platforms like YouTube are prioritizing human creators and their relationship with audiences. User engagement remains closely tied to recognizable creators, with platforms aiming to preserve this connection even as AI tools become embedded. AI is expected to enhance discovery and personalization without displacing creator-driven ecosystems.

Concerns over AI impact on creator economics

New features such as conversational search within video platforms raise concerns about reduced watch time and revenue. However, early expectations suggest that improved user experience could expand overall engagement and increase traffic. Companies are proceeding cautiously, testing features incrementally while monitoring their impact on creator income.

Personalized AI assistants as a key differentiator

Systems like Gemini are being positioned as deeply integrated, context-aware assistants that operate across devices and applications. By leveraging user data—with controls for privacy—these tools aim to deliver proactive support, from task management to content creation. The focus is on seamless daily integration rather than technical benchmarks.

Rise of AI agents and automation

A major shift is underway from simple AI assistance to agent-based systems capable of executing complex, long-running tasks. These agents can manage workflows such as software migrations or multi-step problem solving, often operating in parallel. Future productivity will be measured less by code written and more by tasks completed through orchestrated agent systems.

AI’s growing role in software development

Around 75% of new code in some environments is already generated with AI assistance, primarily through autocomplete and suggestion tools. The next phase, described as “agentic coding,” involves delegating entire projects to AI agents. Engineers increasingly act as supervisors, coordinating multiple agents rather than writing code line by line.

Human role remains central in meaningful activities

While AI is expected to automate routine tasks, human involvement will persist in areas tied to meaning, creativity, and personal connection. Activities such as choosing gifts, planning experiences, or engaging with art are likely to remain human-led. The broader goal is to reduce time spent on repetitive processes while preserving fulfilling experiences.

Multimodal future of content consumption

Emerging systems are enabling content to move seamlessly across formats, allowing users to convert text to video, video to audio, or other combinations. This flexibility is expected to redefine how content is created and consumed, making format less important than the underlying information or narrative.

Rapid pace of technological change

The current stage of AI development is described as an inflection point, with expectations that today’s tools will appear “primitive” within three years. The comparison to early mobile phones underscores the speed of progress and the likelihood of significant capability leaps in the near future.

Opportunities and uncertainty for younger generations

Younger users face both anxiety and opportunity amid rapid change. Adapting to AI tools, developing fluency, and maintaining a mindset of continuous learning are seen as critical strategies. Longer-term gains are expected in areas like healthcare and scientific discovery, potentially improving quality of life.

CONCLUSION

AI is accelerating across industries, but its long-term impact will hinge on balancing automation with human creativity, ensuring that technology enhances rather than replaces meaningful human contributions.

Full transcript

we as humans are not wired for, this pace of It's not easy to live through a period of, like, massive change, I think. YouTube and the media landscape is changing. And, of people like me who grew up on YouTube, we don't want this amazing product to change What would you say to the skeptic about, the opportunities that these new AI tools bring? you know, you still have to put it in the hands of people so that they can express themselves better with these tools. And I think that's the holy grail. Look, I'm very confident. Three years from now, you would look back at what we had today and you will realize. they were, like, pretty primitive. It's like looking at a flip phone. You look at a flip phone, you're not going to be impressed, at one point people were so excited about those phones. something like my mom would ask, she does not care about benchmarks. Benchmarks are great for like, us nerds, right? But for the regular person, they don't really care about benchmarks. they have maybe ChatGPT on their phone. why would she switch to Gemini in one sentence what would you say is the reason to switch? All right. Sundar Thanks so much for doing this. Thanks for your time. It's a real pleasure to be here. Thanks for joining us again. Thank you. Yeah. So I got to ask. You guys announce everything at I/O, and it's like 50 different announcements all in one. Have you ever thought about just spreading it out? Because you could dominate the news cycle over a month. It's a great question. In some ways, you're right. It is a lot. It allows us to tell an end to end story across everything we are doing, what's our direction, how we are taking the company forward. So I think I appreciate the chance to tell it across our products, across our full stack approach. But yeah, sometimes it does tend to get crazy and some things may get lost. So it's a balance. Yeah. But I'm glad we do it this way. Yeah. I'm still digging into everything. I think Omni was definitely my favorite announcement for the day. It's really, like for three years, image and video has been kind of treated as separate from reasoning, but now you have Omni, which is kind of like Nano Banana for video. It's a good way to think about it. What do you think the future of content creation looks like with this in everyone's toolkit now? I think the important thing is I don't think the future of content is, you know, you still have to put it in the hands of people so that they can express themselves better with these tools. And I think that's the holy grail. That's what the audience wants eventually. Right. And I think I think back to CGI and when CGI came, nobody just wants it all CGI all the time in a way. You know, you don't feel the artistry or the creativity come through. And I think we are in that similar moment, what you can do. We are empowering someone from a high school can like over time, make incredible movies, movies, films. So the world is changing that way. I think will allow more people to express themselves, just like YouTube has allowed more people to be creators and producers and so on. So I think we are taking big steps in that continuum. But I think that's the that's the trick. You know, how do you do it? Which is why I like the fact that we are bringing a lot of this in YouTube for creators. We are putting it in tools like Flow, giving people a chance to edit around it, tell a story. And I can't wait for like, new ways in which people, people tell their stories. That's that's what I'm looking for. Yeah, I think it's just more tools to the toolkit, allows creators to be more expressive with what they want to generate. I actually talked with Mark Zuckerberg a few months ago, and, he told me that when we achieve superintelligence, that meta probably won't just recommend content on Instagram. It’ll actually start generating content on the fly as well. Basically like hyper personalized AI feeds. I'm curious to hear, like, your take on this. With AI tools like Veo and Omni. What's kind of YouTube's strategy on that? I think particularly in YouTube, I think we are going to be very creator first. creator first. And I think people come to YouTube, they have these connections with different creators. I find myself looking for you know, creators. If they've uploaded a new video, I want to know about it. But I think there is that connection there between users and creators. That's what I think makes YouTube, you know, special. So I think for us, that's going to be an important way we think about the evolution. Of course, creators can use these tools and there will be time users to like you see. But ask YouTube, you can say, well, we are using AI to kind of like give you, you know, a better answer, allowing you to explore in the direction you want. So we'll do those things too. But I think it's important in the platform at the end of the day. The human to human connection is always there. That's how I think about it. I agree. I want to touch on ask YouTube because this is a new feature. It was very short in the keynote, but was very interesting. Where you can really dig into YouTube videos further and compare them even, How are you guys protecting creators where, you know they might be fearful in that, you know, Ask YouTube or even the, the, ask chat with, with YouTube videos. It could potentially hurt watch time for creators. So what are your thoughts on that and how are you combating that. I think whenever you go through these new transitions, you know, you are exposing the users to more videos. That's true across a lot of what they do. Right? And I think overall in our experience, what we find is it will drive if you make the product better for users. People tend to use it more and and it grows and the growth has to come from somewhere. But look, we are very, you know, YouTube creators are an important part of the ecosystem. So we'll listen to their feedback. You know, I think it's a combination of making sure we are helping users with what they are trying to do, that a lot of information queries coming into YouTube. And actually I think things like Ask YouTube will even pull in more information queries. So it's an opportunity to the platform will grow. And that's how I think we can make it all work for creators too. But you know, we'll have to prove it out. But how things work out. Roll it out slowly. That's right. That's why we are testing it. And you know, we we care about watch time and like you know and yeah. So we are very thoughtful as we work through these things. Actually on a personal I'm a university dropout and I've learned everything off YouTube. Like I wouldn't be sitting here right now with the Google CEO if I didn't have YouTube to actually learn. it's great to hear. And that's why we do the products we do. Yeah. I mean, but without a doubt, like YouTube and the media landscape kind of as a whole is changing. And, you know, a lot of people like me who grew up on YouTube, it's a little bit fearful, right? It's like, you know, we don't want this amazing product to change because it's been so amazing already. What would you say to the skeptic about, you know, the opportunities that these new AI tools bring? Look, I think the technology shift is pretty profound. We are trying to make sure YouTube creators are the forefront. And and they are among the first movers of learning how to use the technology and interpret it in a way that works for their audiences. And I think they're pretty committed to being in this journey together, in a way. I think the strength of YouTube as a platform in the past and the future will be based on the strength of the creator ecosystem, right? And and both in terms of, you know, to us, we view creators as businesses, next generation media businesses on this platform. Right. And like we feel a lot of responsibility because we view it that way. And so every evolution we take, we have that deeply in our mind. Neal and team are constantly thinking through those things as we make progress. Have you thought about sort of educational resources for creators? Because even even on the creator side, I've seen a lot of my peers even who are, you know, fearful of even using the tools because they think their audience is not going to watch them if they're using these tools. But I think, you know, as we're talking about, these new tools basically create endless opportunities for creating new types of content, right? You couldn't create with limited resources before. Have you guys thought about, you know, any sort of educational resource or, you know, even YouTube videos on how to use tools like Omni to better educate creators. It's a great idea. If you ever want to work on the YouTube team we’ll sign you up. Yeah, but, like, look, I do think, you know, take a feature like auto dubbing, right? That's dramatically expanding the reach of, like, YouTube content to more and more people. And, you know, so I do think getting more creators start out these tools. And that's how we get the feedback of what's working for users with this not do. And I think that's important. So we have to do a better job at it. So I definitely will be talking about this to Neal and the YouTube team. Okay. Sweet. Yeah. I mean, I guess it's even more important as these capabilities increase and not feel like we're at that point with Nano Banana, Omni, which is nano banana for video that we are at that point, like there are creators getting a significant edge. And I feel like we gotta level the playing field with education. I mean, it's part of what we do with our company as well. But, anyways, I'd love to kind of get into like a broader question for something like something like my mom would ask, she does not care about benchmarks. Benchmarks are great for like, us nerds, right? But for the regular person, they don't really care about benchmarks. You know, they have maybe ChatGPT on their phone. But why would she switch to Gemini like, in one sentence what would you say is the reason to switch? Look, I think what makes Gemini is how seamless it is in the context of probably a lot of things she's doing in her day to day life. And I think we are making it very personal with things like personalized intelligence. I think over time, it will have a better understanding of what she's doing in her life in a way that it can help her better. I think with things like Spark, we are bringing the power of agents, which we as developers or people on the cutting edge understand. But she doesn't need to understand all that. But being able to give Gemini action items, tasks and what should make her life better. That's all we want her to feel right. And so I think that's the promise we are working on. We want this to be this personal, proactive, powerful assistant to her. And I think if you look at all the things we are doing, I think we are making a lot of progress. And I also love the creative aspects of Gemini, you know, with nano banana, with Veo, with Omni, I think I think people can express themselves through these tools as well. Or if you're a coder, you can program. But ultimately these are all these people are expressing themselves. So I like that way we are differentiated as well. I agree. Let's talk about the personalization a little bit more because I know, on at the keynote, you guys kind of demoed the glasses. They’re not available, till I believe the summer. But also the cross compatibility with the watches. And you've got the phones, you got the laptop. So I'm curious on the personalization front, is the memory, like, kind of hold across all the platforms? Is that like a significant edge for Google or how are you thinking about that? I think look, with personal intelligence, we want to give users controls in terms of what they want to link but if it is, it's across your account, across all of these apps. And and remember the agents that also built 24/7 in the cloud in Google Cloud running as a VM (Virtual Machine). So it's think of it as personalized agents which are working for you behind the scenes and as context across what you do. We have to be very responsible and make sure that users feel in control. They can stop things, etc. or they can do incognito versions of these tasks. But I think that's what's delightful today, how Gmail works. You know, you don't think about whether you have access to your Gmail on your laptop and even even for someone like your mother, it just works in your life. Google docs works the same way. And I think that's the magic of, how we can build these things. You're really solving problems for users over time. Ultimately, the power of these tools come from freeing up some of the tasks you do so that you're spending your time more meaningfully on whatever you want to do. Right. That's the promise of this technology. We all have to work hard to deliver on that promise. That's how I see it. Right, agents in your phone now too right? Agents everywhere. Yeah. Let's zoom three years from now. Let's say we're sitting again today. In three years. What's one prediction you'd make that you think is not obvious right now, but will be obvious then? Look, I'm very confident. Three years from now, you would look back at what we had today and you will realize. Like, they were, like, pretty primitive. It's like looking at a flip phone. You look at a flip phone, you're not going to be impressed, right? Like that. Like at one point people were so excited about those phones. I think we are in that rare moment in technology. The curve is so steep. People will look back three years from now or three years ago and say, oh, you mean you couldn't do those things? Right. I think it will surprise people. I think there's a lot of forward progress ahead. I think we are all laying the foundation for agents, but agents will be very common. I think people will be relying on them to get things done in a more seamless way. So that's one thing I see in a very tangible way happening. I do think any content can go from any modality to any modality. We have shown hints of it with Nano Banana, Veo, and Omni. But truly it's going to work in all these ways, so doesn't matter how the content was originally created, people can consume it in many different forms, modalities. I think that's that's unique, right? Right. I also read that now 75% of new code at Google is generated by AI. Do you have a prediction on when that could be potentially 100%? Look, I would say the 75% is like where people are coding and they get auto generated, you know, auto completions and they're accepting it. I think the frontier shifting from that, that metric matters less to what we are talking about in Antigravity today with like agents and sub agents. And you're managing a team of agents. So the new metric is going to be what task did you get done by just deploying agents. Engineers, each engineer will have like these many, many agents at their disposal. So I think it's a whole new metric. And I think that's the shift already underway. So internally using Antigravity people are able to do what I call agentic coding, long running tasks. Can you take a code migration project? You have to take something from C++ to Rust or Python to Go, and and it's a complex code base, a small team of engineers, can you deploy a bunch of agents, use the current code and the program or the binary as a spec and like get the output you want. Those are, you know, where the frontier is I think. It's agent management, agent orchestration. And these sorts of things are the way engineers might shift their responsibilities in their job. Is there anything specific that you think we do now? It could be across, you know, engineering. It could be across any anything really. Is there any like tasks that you think that agents just won't be able to do? You know, the human in the loop really will need to be there. You know, I think everything, you know, take shopping, for example. Shopping is a really enjoyable task for a lot of people, for a lot of things. Right. And and then there are times shopping is not, you know, you want to do your weekly groceries when you may be super busy. You have two kids, you just want it done right. That's different than buying a gift for a loved one, right? Like, you know, so I think so. Part of it is the system should work in a way so that you are engaging in that higher, meaningful level. If you want to plan a trip, you want to spend your time on understanding the activities you can do and which one is better, and so on. Not all the underlying mechanics of like, oh, I need this metro ticket to go from place A to place B, and where should I go, you know, that's the difference, I think. How do you shift? How do you take the mundane tasks and make it easier for you? But how do you take the aspects of this which are joyful and meaningful, involve spending time with other people? I think everywhere it's that how do you free up doctors to actually spend more time with their patients, right. And and less on like billing or medical records or transcription of your notes or getting second opinions on things you want. Right. And like, you know, that's that's how I see this evolving. All right. I know we're almost out of time here. So last one. There's a sort of state of confusion among young people and even Silicon Valley. I see this kind of like AI frenzy. Right. What advice would you give to, let's say, an 18 year old for how to best set themselves up for the future? Look I, I think we are, you know, we as humans are not wired for, this pace of change. It's not easy to live through a period of, like, massive change, I think. I do think if you're young, you have a lot of opportunity ahead of you, right? I think, you know, with, with the progress we are seeing in scientific discovery, health care, etc., most of the young generation hopefully will lead longer lives with better health span than the previous generation. Right. So I would just say at this time to kind of slow down and absorb and learn how to use these tools and get comfortable and, and realize that these tools are going to help you do things better than ever before. But keeping that mindset, no different from how you said earlier, you used YouTube as a tool to learn everything. That's how you approached it. And like you learned a lot of things. And by the way, I've done the same thing and, you know, and, and so I would just view it as a continual learning moment and you like, you want to become comfortable and native to this technology, right? Right. That's how I would approach it. If I were young. Would you still advise going to college? I would, you know, I think a lot of those things are about growth in your life. Meaning for some people that they don't need to do that. You know, so I think these are uniquely personal decisions. But if I were 18, would I go to college? I would. I learned a lot in college. I, I met my current wife there in college, and so I wouldn't trade all that for anything else. Right. All right. Well, that's all the time we have. But thanks so much for doing this. All right. Thanks Rowan, all the best. Likewise, take care.

More from AI