🔰A Conversation with the CEO of Google

Sundar Pichai
1 follower
Mar 24, 2025 #Sundarpichai #Business
Sundar, I hope you're well. Can you see me? I can, Your Excellency. Good to see you
again. Thanks for having me. Thank you so much for joining us. We would have loved
to have you in person, but technology enables us to have this conversation
wherever in the world you are. I'm in France right now and so, you know, I'm
glad to be able to do this. I just heard about your AI investment in a big data
center here in France. Congratulations, and it's great to see. Thank you very
much. I'd like to maybe steer this conversation in a few directions, if
you'll allow me, and we have 30 minutes. So just to set the context, we're going
to start first by talking about the great technologies that Google has been
at the forefront of, and then we're going to talk about the company, the culture,
and the lessons for governments. And if you don't mind, we're going to end with
some personal questions about you. Great, looking forward to it. So, you know, for a
long period of time, every single moment someone talks about the technology
breakthrough, they would say we had a AlphaGo moment with this breakthrough or
with this technology. Today, what's top of mind for everyone is DeepSeek and the
breakthrough that it created. And I'm sure that you get this a lot, but I'd
love to get your thoughts on what the breakthrough was, what the lessons
learned were, and what we need to pay attention to as governments with regards
to this announcement that the company in China has made. Okay, I think the DeepSeek
team has done very, very good work, but for me, the lessons learned, you know,
stepping back, it shows how global this AI development is, right? It's happening
all around the world. It's a moment of excitement and inflection everywhere. You
can see there is frontier model development happening around the world. I
think what caught people's attention with DeepSeek was that you could have an
efficient model, right, and an open source and something everyone can
immediately access. And I think that creates a lot of excitement. But, you
know, it's something obviously we are focused on as well. We've always realized
the models which you serve people around the world, you know, it has to be very
efficient. And, you know, and the biggest takeaway for me, which we knew that
because we were working on it, we've done that with Gemini flash models, they are
very, very efficient, is that intelligence is going to be like widely
available, right? This technology is going to be at everyone's fingertips, and
it's going to impact the world profoundly. And I think the DeepSeek
innovation, you know, reinforces that point. You had a similar moment, I
think, a few months ago, when you guys made an incredible breakthrough in
quantum technologies and really showed the world that quantum is actually
closer today than it ever was at any point of time. I'd love for you to maybe
shed some light on that and tell us what governments should pay attention to
there. Should we talk about regulating quantum at this point? Is it something
that's going to come in the next year or five or 10? What excites you about this?
And what should we be worried about?
Look, we've been, you know, the progress in quantum is palpably exciting. We've
been focused on it for a long time. I think we have the most advanced state
of the art effort anywhere in the world. The latest breakthrough we had, we were
able to perform a computation in five minutes that would have taken one of the
world's fastest supercomputers over 10 billion years. It's longer than our
universe has been around. But more importantly, it shows that we can tackle
error correction in a way that we can build large scale, fault tolerant quantum
computers. I would say to me, the quantum moment reminds me of where AI was in
the 2010s, you know, when we were working on Google Brain and, you know, and the
early progress. So I would say, you know, that's what it feels to me in a five to
10 year time frame, I think you will start practically useful quantum
computers. In terms of what governments can do. Look, I think it's too early to
regulate technology. But I think it's important to, you know, come to a good
understanding of it, and then start preparing for it. The main first areas we
would need to tackle our encryption. But beyond that, I you know, I would think
about no different from AI. You know, how do you how do you set up a plan and set
up a quantum cluster to do some pioneering research work in the in the
next three to five year time frame? That's, that's how I would approach it.
So this question has been going on my head for a while. It'll take a normal
computer, multiple billion years to identify if the answer is correct. How do
you know the answer is correct? And what if it just throw out a number there that
doesn't make sense?
You know, look, I mean, that we do. We, the techniques we use have very, there's
a way to verify that answer is correct, which is different from how long it
takes to compute an answer. And so those are the techniques we use to triangulate
there.
So let me maybe pivot a little bit about another venture that you guys are into,
you are into AI heavily. Today, most people actually don't recognize this,
but around 8.5 billion searches are done a day. And correct me if I'm wrong
about this, but that's more than the population of Earth. And you have around
49% of the internet traffic every single month coming to Google. So there's a lot
of, you know, advantages that you have in the AI domain. And you know, Google
will continue to be a leader because the world comes to Google to answer their
questions. There have been talks about, you know, people moving to chat GPT or
moving to other, you know, whether it's grok or other platforms for search. Do
you see that affecting Google in any way? Or is it so far that growth is
exceeding any sort of decline from these platforms?
Look, you know, broadly, stepping back, I, you know, I think we are in the
business of providing information to people. People's information needs are
exploding, if you will, right. So I think, you know, so it feels very far from a
zero sum construct to me. Having said that, you know, we are continuing to see
strong growth in search. We have recently evolved search pretty profoundly with AI
overviews. And we have found that people are engaging with it more. It's leading
to a growth in search usage across all demographics and so on. So it feels like a
positive moment. And but more importantly, I think taking the long term view, I
think AI is going to help expand the opportunity space, not just for us, but for
others working there too, right? With the kind of agentic capabilities that it will
have, you can not only ask questions, you can get things done. And so I think it can
help you be much more productive across a range of domains. And that's the
opportunity space ahead. And, you know, given how big the landscape is, I
definitely expect us to be a leading player. But it's natural to me that there'll be
others doing well in the space as well. That's what it is just like the internet
going to be a massive space. We were in the only company to do well in the internet
that have been other companies, right. And you know, the AI would be no different.
So zooming out, when you actually look at Google, what we realize, and this is
something that, you know, I looked at quite closely, it looks like you have an AI
layer that cuts across all of your different products, whether it's the Google
Pixel right now with the incredible camera capabilities of, you know, allowing
you to do things that other phones don't allow you, or maybe people are catching
up on that, whether it's you're deploying self driving cars and adding AI as the
core component of it. So it's clear to me, and I think it's been clear to the world
that you're an AI first company. Let's talk about self driving cars for a
second. You know, I know that Waymo probably is the most advanced company
self driving cars probably in the world today. Why aren't we seeing more
applications of Waymo across different countries? What can we do to ensure that
we usher this future in closer rather than push it further out in the future?
Look, I'm, I've been super excited about Waymo for a while. And I think we are
really hitting our stride in the past 12 months or so. We've obviously been
very, very safety focused. So the way we approach this, we want to build a very
safe and reliable Waymo driver. And once we do that, we can really scale up from
there. If we have definitely shown the progress there, you know, thanks to the
work with AI we have done over the years, we now have over 33 million rider miles
driven. And Waymo has had 78% fewer injury causing crashes compared to human
drivers, right. And so now the progress is clearly there. Last year, Waymo, I
think we served over 4 million passenger trips. And we are now averaging over
150,000 rides, paid rides per week in the US across San Francisco, Phoenix and
LA. So I think the first time any AV company has reached this kind of scale.
And we are now expanding, you know, we have a goal of doing this in 10 new cities
in 2025. We'll be doing our first international testing in Japan. And so we
do want to work with, you know, other innovative countries like you all, you
know, and that's the roadmap now we are putting in place. But you know, we are at
an exciting juncture. And I think I think the technology is working. And as you
rightly pointed out, we have to we have to scale it up by partnering with
regulators in countries around the world.
So we're ready to partner. Let's put the UAE on that list and on that roadmap.
I look forward to I want to ask you a question as well about this. So once we
achieve certain technological advancements, like self driving cars, for
example, or on artificial intelligence or others, it creates new types of jobs.
Although yes, it does eliminate, you know, some jobs in the past, I think that
these things cause disruptions. I've been thinking in the past couple of
weeks, what kind of jobs are going to be created within the experience of
someone getting into a self driving car to, you know, arriving to their
destination? You know, would we actually and I saw this example in Japan, in
Japan, you can actually rent a friend. So you can actually rent someone's time
for them to come and talk to you. Will we actually start creating these jobs
that will make the commute more appealing for us? Do you think that new
platforms are going to emerge from these kinds of services? Do you have any
ideas of what's going to happen?
You know, maybe two observations, right? I think one is, I think there was a
recent study by MIT economist David Autor, which said, you know, if you go
back to 1940, since then 60% of the new jobs that that have been created kind
of didn't exist before that, right, you know, and so there's this constant
evolution we go through. Just imagine even five years after the internet
started in the year 2000, envisioning that there could be someone called a
YouTube creator. Today, there are millions of creators, and they're real
small businesses, right? And you know, they generate value, they employ other
people and so on. So we have to keep that potential in mind. To me, there is
no doubt in an AI driven world, just like the internet did, there'll be many,
many new opportunities created. And I think over time, there'll be value and a
premium placed on human experiences and human interactions. Right? I think if
anything, the prevalence of AI would place even more a premium on human
experiences. You know, we've seen that with chess, right? You know, chess has
been AI is the best chess player in the world. And that's been true for a long,
long time. You know, chess is more popular than ever before, right? And more
people are playing chess now, right? And so I think it's worth keeping that in
mind. And you know, so while the technology is going to be very
prevalent, it's going to improve productivity, it will cause disruption,
but it'll also create many, many new types of opportunities. When I look at
our latest video creation models, video generation models, VO two, you know, just
like the internet, it's going to enable many, many more people to create content
than ever before, right? And so I think we are expanding the type of
opportunities available to people. Amazing. Let's pivot a little bit to the
company. And you know, everyone in the world today, if you ask them what
innovation, if you ask them about companies that are able to enact change
in a way that allows them to maneuver around any change that happens in the
world, Google comes top of mind. And most people think that Google is able to
innovate because they offer beanbags and free coffee to their employees. I'm sure
it's a lot more than that. The first thing I want to ask you about is what
makes the culture of Google so special? Is there a formula? Is there a thing
that you do to inspire people? How does it work?
Okay, I think it's a combination of things. You know, I think our founders,
you know, had, you know, you know, really set a very ambitious framework for the
company, right? We used to say, we need to have a healthy disregard for the
impossible. And, you know, we always felt that if you encourage people to work on
very ambitious things, even if you end up failing, well, there are many
advantages to doing that. One, you have very little competition, the more
ambitious the thing you're working on, Waymo is a great example, right? You are
quantum is or AI is, you know, there are fewer people working on it, and you
attract the best people. And even if you end up not achieving that ambitious
goal, whatever you end up doing will be very, very valuable. So I think that's
the kind of framework we have always had. I would say, you know, you have to
have an optimistic culture, you know, a culture of curiosity, you encourage
optimism, you know, you, you really push people to take risks. And you don't
punish failures, right? You encourage risk taking in your culture. That's a
combination of all that. And but you know, we've always had to experiment. You
know, for example, just about two years ago, I set up a labs group, so that
people can have small teams to go ship stuff. And out of which we shipped
something like notebook LM last year. And you know, which which ended up being
very, very popular, right? So you have to constantly tinker and make sure you're
innovating from the ground up. So you approach things with the approach of a
scientist, I went and saw how many projects Google killed, or let's say
stopped over the last couple of years, it's around 296 projects in total, maybe
the number is a bit higher or lower. And what struck me is the fact that you're
still trying to innovate, you're still trying to iterate. And yes, you might
have breakthroughs like Android, for example, or like Chrome, or like Waymo.
Most companies, when they go through so many iterations of killing projects, they
tend to stop innovating altogether. But you guys maintain this. How do you
identify when you continue a project and when you actually stop? What triggers you
to say, okay, you know what, we've got to go, we're gonna park it for now. And we're
gonna focus on something else. Look, I think there are two types of it, right?
One is at the core technology layer, you're making long term bets, right? You
know, I mean, 2015, when I became CEO, I said the company is going to be AI
first, and we decided to take a full stack approach to AI, right, all the way,
building research, building infrastructure, we started building our
tensor processing units in 2017, we are in our sixth generation. When you make
these long term technology bets, AI, Waymo, quantum, those, you know, it will
eventually happen, right? So you see it through. So I think that foundational
layer, you take a long term view, and you don't waver at all. At the application
layer, you are right, when you when you take efforts to build products, you're
going to have some successes. And some things won't work out. But you know,
that's, that's true, by definition. I think, you know, there's judgment there
as to when, you know, when you, you know, when you keep pushing through, and at
what point do you see it's not getting traction, your approach is wrong, and
you have to course correct. And you know, there's definitely judgment there. We
don't always get it right. But I think it's important to important to go
through that reading process to, you know, that's what keeps the company fresh.
And you're constantly pushing to do new things.
It's, we were having a conversation yesterday with a global CEO as well. And
it's the gambler's dilemma. When do you cut your losses? And when do you keep
investing? And you know, when when do you hit it big? I want to maybe ask you a
final question about the culture in Google. Do you tell people to use AI? Do
you have any training programs for them to use generative AI? Is it something
that you push for? Or is it something that you're waiting just to see how it's
going to affect the employees and the culture in Google?
Look, we add this is a internally at Google, you know, this concept of dog
fooding, you, you eat your own dog food is what we call it, you know, we really
want people to adopt our technologies and be at the cutting edge. There are two
examples, I would say one is, we are really trying to help engineers code
using AI, right. And already, we have made deep progress on having AI tools
give people suggestions that can complete as they are writing lines of
code. In fact, 25% of all code being written now has those suggestions, you
know, from AI, which are accepted by engineers. But now we have smaller
projects, which is pushing it to the next level, like, you know, how can they
give the AI agents tasks and that can go and actually do much deeper work for
them. The second is we are taking many, many functions and business processes
across Google, and trying to think how we can make them more productive using AI
in a deep way. How can we do customer service in an end to end way that is
better, right? And so, but to answer your question, yes, we are deeply
encouraging people to use AI, right, and we measure it, we track it. And when we
find solutions that work, we bring it to our cloud partners as solutions for
them to try in their companies, right. And that's how we create that iterative
cycle.
There was a initiative that was launched in the UAE. So the Crown Prince of Dubai,
once he saw the impact of generative AI, said that we want to do a citywide and
now maybe a countrywide program to teach a million people how to prompt properly
on these tools and become much more effective. And there is a genuine belief
here that this tool can be wielded to change the world. And we need to get
people ready to be able to prompt better, to be able to extract value better, and
to do more. So I'm glad that, you know, whether it's in coding or in other
domains, we're able to do more with people. I'd like to pivot now maybe to
more questions about your journey, personal questions, if you don't mind.
And I was actually inspired to hear your story, how a boy that was born in India
and, you know, had the ambition to change the world ended up running one of the
most valuable companies in the world and maybe the most recognized company that
everyone knows around the world. Why do you think you made it? What was the
secret that got Sundar to start in a village in India and end leading one of
the world's largest companies?
You know, I'd have to reflect on it. Look, I think for me, you know, I was
fortunate in some way, you know, I had to wait a long time to get access to
technology. So growing up, every moment was very, very vivid for me, by waiting
for a telephone for five years, etc. And so, you know, I saw firsthand how it
changed my life, and the lives of people around me. So I think I had this
passion to kind of get access to technology, and to make sure I was
working on things that would bring technology to as many people as
possible. So maybe in some ways, I had a personal mission, right. And, you
know, I was fortunate to find a company like Google, when I read Google's
mission statement, in which it said, you know, wanted to make information
universally accessible. You know, that's what, you know, really excited me to
come to Google. So I think part of it is being, you know, having, you need to
get excited by a mission, that's what pushes you to do your best work. Now,
beyond that, I would say, you know, challenging yourself to work with people
who are better than, right. And I think, you know, so you have to, you know, you
learn and grow when you're pushing yourself to be with people who you think
are better than you in many, many ways. And so, you know, I've strived to do that
over the years. And you know, it's what helps you learn and grow. So maybe it
may be a combination of that. But you know, I would encourage everyone to each
person, the path is different. But if you can find something, you know, that
that resonates with you in a deeper way to your heart, I think that brings out
the best work you can do.
It gives us hope that maybe a child born somewhere in the Middle East, in a
rural village can one day in the future become the CEO of one of the largest
companies in the world. And I feel like there's a lot of lessons from your story
there. Every single person I met, everywhere I've been, is obsessed about
something. What are you most obsessed about?
Yeah, look, I mean, at this moment, you know, I'm spending a lot of my time on AI
and the rate of progress, how profound that technology is. It's the first time I
felt even if I travel for a couple of weeks outside, and I come back to the
company, I go back to the teams working on AI to catch up on the progress that has
just happened in the last two weeks, right? So we have rarely seen a
technology, which is progressing. So you know, obsessing about making sure we are
driving the technology forward, making sure we are building applications around
it. And that we are doing it responsibly, we want to be bold, right with this
moment in time, innovate. But the technology obviously needs to be deployed
responsibly. And so thinking through it, you know, is partly my obsession
nowadays. And so I think it does come back to AI.
And since AI is the reason why you're obsessed, any fears that you have for
what AI can do that we're not paying attention to? My view always is Google
can do a lot more, but I believe that you guys are restricting yourselves from
just deploying everything because you want to be responsible players on this
domain. But what do you fear most when it comes to AI? What's something that you
know, as government officials need to pay attention to?
Look, about everything, you know, I would say as governments, you want to focus on
the AI opportunity, right? And I think UAE has been, you know, I want to give
credit, you know, it always surprised me that, you know, you became the national
AI minister back in 2017, I think. And, you know, I think that is pretty
farsighted. And I do think for governments that's going to be this, there's
going to be no technology which would impact your economy as much as AI. So
stepping back, making sure you're investing in your country's infrastructure
like UAE has done, making sure you're skilling your population. You gave the
example of teaching people how to prompt in Dubai, right? And so skilling is going
to be important. Unlocking data, you know, in the public sector, right? And
using that data to come up with solutions using AI. These are all tangible ways in
which I think, you know, government should think about AI. On the risk side,
which you asked me, and I think all of us need to be responsible with this
technology, the opportunity space is huge, but obviously, it's a very powerful
technology, it can be a self improving technology. And, you know, so making sure,
you know, you're developing competency, to evaluate risk, assess risks, and take
the appropriate steps over time, you know, it's going to be something we all
need to invest in. And there have to be global standards around it, right. So for
example, you know, as the technology can create real deep fakes, you know, what
are what are regulations to prevent that, right? You know, what are standards, all
of us need to deploy in our technology so that you can detect that something was AI
generated, right. And so those are all going to be important provisions. I think
both end up being equally important through this moment.
What can we do about defects? Because, you know, I think you face this more than
anyone, you are a famous person that has a lot of video content online. Is there
anything we can do about that?
Look, I think we are definitely developing technology, you know, and we
are open sourcing it, we call it Synth ID. And so we are watermarking image
generation. And, and we are also open sourcing the tools to detect it, we need
to work with other companies to create a standard around it. But over time, you
know, I think broadly, we want to encourage AI innovation. But there'll be
areas where regulation makes sense. You know, like today, in the financial world,
you have laws against fraud, right? So you're going to need the equivalence of
that, you know, in AI world, and you know, something like deep fakes would need to
be tackled that way.
Let's assume that you and I try to ask this question to everyone that I meet,
let's assume that you woke up tomorrow, in my body, and you were the Minister of
AI, what decisions would you take that are currently not being thought of? Like,
what is one thing that you're going to focus on that you think will have the
biggest value for a company like the UAE, or any country in the world?
You know, beyond me, you know, I would reemphasize, you know, the four areas I
would focus on is supporting infrastructure, right? I think, you know,
I think most people are underestimating the infrastructure transition. I think
UAE has been at the forefront, because it's not just compute data centers, it's
energy transition, which obviously, you know, you've been at the forefront on.
Skilling, obviously, is another important area. I think the amount of work that
needs to be done to help your population get ready and transition, I think is
something that needs to be a lot of effort put into it. I would definitely, I
think that much tougher challenge has been encouraging governments to unlock
data, right? And, you know, how can governments better utilize and share data
to improve public services, like health care, education, transportation, disaster
response. So I think, I think being at the forefront of that, I think is a real
opportunity. So that is something I would think about. And then finally, how do you
have balanced regulation, which is pro innovation. But, you know, for high risk
applications, you know, how do you how do you think about, you know, responsible
regulation, all of that, you know, needs to be in the toolkit, you know, is what I
would emphasize on. But you know, above all, I would say, being excited at this
moment, embracing innovation, because it is something that's coming at us in a
fast way, is going to be very, very important.
Sundar, thank you very much for your time. That was great. Thank you for taking all
the questions. I wish you the best of luck in France, and hopefully we'll see
you here next year.
