YT YouTube2Text

YouTube2Text → Transcripts → Big ideas begin here: Sergey Brin at Stanford

Big ideas begin here: Sergey Brin at Stanford — Transcript

by Stanford University School of Engineering · 10,436 words · 1,467 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:02(upbeat music)
1:06(upbeat music continues)
1:21(people clapping)
1:26- Welcome everybody.
1:28This is the closing event of our centennial year.
1:32It's pretty exciting.
1:33I'm Jennifer Widom.
1:35I'm the 10th Dean of Engineering,
1:37and you can do a little math there
1:38and figure out we deans like to stay around
1:40for a while here.
1:43It's been a great year.
1:44It's been a year of celebration, reflection.
1:47Looking to the future.
1:49For those of you
1:50who haven't heard about the history,
1:52we've had engineering at Stanford
1:54since the beginning of the university, 1891.
1:57We started with chemical engineering,
1:59electrical engineering, mechanical engineering,
2:02and mining and metallurgy.
2:04And it was in 1925, 100 years ago,
2:07when those four departments were brought together
2:09to form a school.
2:11And those four departments are still with us today.
2:13One of them has been renamed,
2:15it's now Material Science and Engineering,
2:17rather than metallurgy.
2:19And we have five more departments
2:21and many, many interdisciplinary programs.
2:24We've had events throughout the year
2:26to celebrate the centennial.
2:28They've been fantastic.
2:30We started with a panel with the five deans
2:34of the 10 who are still with us.
2:36That panel was moderated by Jerry Yang,
2:38who I think is here today, a great friend of the school.
2:41Our second event was on May 15th.
2:44We had a big party out on the quad.
2:47We expected 2000 people.
2:48We got 3000.
2:49And we didn't run out of food.
2:51We had a great showcase of projects and research.
2:56The next event was a fireside chat
2:58with Jensen Huang and John Hennessy.
3:00That was actually on this very stage,
3:02around the end of May.
3:05We partnered with Stanford football
3:07and had a school of engineering themed football game.
3:11Nobody told me when I took the job as Dean that one
3:14of my jobs would be
3:16to drive a motorized couch on national TV
3:20with Andrew Luck as my passenger,
3:22but I nailed it and that that was great.
3:27We had on Reunion weekend,
3:29a trivia contest on the history of the school.
3:32The alums enjoyed it.
3:33And now this is our closing event.
3:37If we look back on the hundred years
3:39of the School of Engineering,
3:40obviously the formation of Google was a shining moment.
3:44Sergey Brin, who you'll soon meet,
3:47met Larry Page when he visited to think about coming
3:50to our PhD program.
3:51That was in 1995.
3:53They worked together on a project called Digital Libraries
3:57that was funded by the National Science Foundation.
4:00So if you ever have any doubt about the impact
4:03of federal funding,
4:04Google came directly from an NSF project.
4:07We all know what happened next.
4:09They developed an algorithm called BackRub,
4:12which became PageRank.
4:14And by the way, that server right there is the first server
4:18that ran the PageRank algorithm.
4:20So a bit of history right in front of you.
4:22We're gonna hear more about that time shortly.
4:25But I do wanna say that's only one example
4:27of the entrepreneurship that has happened across the decades
4:31and the the year,
4:33the century of Sanford Engineering.
4:35Thousands of other companies have been founded
4:38by students, by faculty, by alumni.
4:41They've generated literally trillions
4:43of dollars in economic growth.
4:45And that foundation was laid by the third Dean
4:48of the School of Engineering,
4:49someone by the name of Fred Terman.
4:52Fred mentored William Hewlett and David Packard.
4:55He also mentored the Varian Brothers,
4:57and he also helped establish the Stanford Industrial Park,
5:01which is now known as the Stanford Research Park.
5:04Still going strong today.
5:07So many breakthroughs over the years
5:09in Stanford engineering in aeronautics,
5:11electricity transmission,
5:13microwave radar, semiconductor work
5:15that really sparked Silicon Valley, cybersecurity,
5:19that all of us rely on,
5:20internet transmission protocols
5:22that all of us rely on,
5:23the foundations of AI, bio electronics,
5:26lithium ion batteries.
5:27The list goes on and on, and there's surely more to come.
5:31Now, I do wanna acknowledge
5:33that when we look at the work that we do,
5:35a lot of the details are really done by students.
5:39And today's event is also a class.
5:42More than half of the room here are students
5:44in the Entrepreneurial Thought Leaders Program at Stanford.
5:47And I just wanna say to the students,
5:50you're the sequel to our histories.
5:52So thank you students for coming.
5:55(people clapping)
6:00And today's conversation is also really about students.
6:03John Levin, the president, who you'll meet shortly,
6:06and Sergey Brin, we're Stanford students
6:09just like all of you.
6:10Today, they're helping define the future
6:12of technology and the future of education.
6:15So this is really an ideal way to close our centennial.
6:19Now I wanna set the stage
6:21of the early 1990s when Sergey Brin arrived
6:24as a CS PhD student, computer science.
6:28Email was just becoming the way all
6:30of us were communicating.
6:32Entrepreneurship was just beginning to accelerate.
6:35The sixth Dean of the School of Engineering, Jim Gibbons,
6:38hatched the idea
6:40for the Stanford Technology Ventures program,
6:42which hosts the class that's here today.
6:45The Science and Engineering Quad
6:46was on drafting paper at best, maybe just in people's minds.
6:50And by the way, I arrived the same year that Sergey did.
6:53I joined as an assistant professor in 1993.
6:57There was also an undergraduate at Stanford
6:59beginning his senior year that same time.
7:02And that was John Levin.
7:04And one wonders was he thinking
7:05he'd eventually become the president?
7:07Who knows.
7:09John was an undergraduate in math and English here.
7:12He went to get his PhD at MIT
7:15and returned as a faculty member in 2000.
7:18He was chair of the Economics Department,
7:20the Dean of the Business School,
7:21and became Stanford's 13th president in August of 2024.
7:26I wanna say that John is really a wonderful champion
7:30of the School of Engineering.
7:31He's been so helpful throughout our centennial year.
7:34He deeply understands our culture
7:36and our spirit of entrepreneurship.
7:39So it is my pleasure now to bring John Levin,
7:43president of Stanford, and Sergey Brin,
7:45to join our conversation.
7:47(people clapping)
7:51- I'm John Levin.
7:51It's great to see all of you here
7:54and to have this chance to celebrate the final event
7:58of the engineering school's centennial year.
8:02We have an extraordinary guest
8:05and fitting guests for the final event
8:08of the centennial with Sergey Brin,
8:11who really needs very little introduction.
8:14But I thought I might just start
8:18before we get into questions with Sergey,
8:20to roll the clock back to,
8:23as Jennifer just did.
8:25She gave such a wonderful sweeping history of the school,
8:29but to roll the clock back to that particular moment
8:31in the 1990s when Sergey came
8:35to Stanford in 1993 as a PhD student.
8:38Jennifer was a new faculty member.
8:41I was a Stanford undergraduate at that time.
8:44I was a senior in the fall of 1993.
8:48And some years ago
8:52a reporter for the New York Times
8:53wrote an article about my Stanford graduating class,
8:57which was the class of 94.
8:59And the gist of the article was,
9:01it was,
9:04I hope that's better.
9:06The gist of the article was that it was the most fortunate
9:09grad college graduating class of all time,
9:12because we graduated at a moment that was on the cusp
9:15of the internet and technology taking off.
9:18And we were right in the middle of it in Silicon Valley.
9:20And the reporter called me
9:22when they were writing that article,
9:23it was about 10 years ago,
9:25and said, "I wanted to talk to you about this."
9:29At the time, I was a faculty member at Stanford.
9:32And she said, "You're the first person
9:36that I'm calling about this article."
9:38And I said, "Look, I'm, I'm really sorry to disappoint you.
9:41It's true, I was at Stanford in the early nineties.
9:44I was in the class of 1994,
9:47but I was clueless at the time.
9:51And in fact, I left Stanford in 1994,
9:54and I went to be a graduate student at Oxford in England,
9:58a country where they had forgotten the technology
10:00for the ice cube.
10:04And by the time I made it back to Stanford,
10:06which was a few years later in 2000, you know,
10:09things had already taken off.
10:10So I really can't help you."
10:12And she said, "Well, don't worry about that.
10:14I like to start these articles by talking
10:15to people on the periphery, and then I work my way in."
10:18(people laughing)
10:20So if I was on the periphery, there were a few people
10:24who were at the absolute core
10:27of the internet revolution.
10:30And and there's no one
10:31who was more at the center than Sergey Brin,
10:34who did see the potential
10:36and the future of where technology could go
10:39and did something extraordinary and world changing,
10:42of the kind that we, you know,
10:44we hope many people,
10:45has been done many times from the engineering school,
10:47and we hope will happen many times
10:48in the school's next century.
10:49So, Sergey, thank you for being here to be part
10:54of this celebration of the school.
10:56- Okay, you guys are flattering me way too much.
10:59I think there was a huge amount of luck there,
11:01but anyway, thank you for hosting me.
11:03It's a pleasure to be here.
11:05- So let's go back to that time
11:08and take us back to, you're a graduate student at Stanford.
11:14Tell us a little bit what it was like
11:16to be at the engineering school
11:19and how it helped to shape you
11:21and open up the opportunity to create Google.
11:26- In hindsight, maybe I didn't appreciate it at the time,
11:29but it was a very creative and free time,
11:35I would say.
11:36We started in,
11:38or I started my PhD program in Margaret Jacks Hal.
11:40So on the main quad actually.
11:43And this old kind of building with, you know,
11:47creaky little rooms and doors.
11:50I learned how to pick locks there
11:54thanks to the MIT guy to lock picking.
11:58And yeah, I'm kind of surprised honestly
12:02how much freedom I was given in hindsight,
12:07because I, you know, I could spend my time.
12:11Initially I was trying to reverse shredders,
12:15like, you know, you shred your documents
12:17and then scan it and put it back together.
12:18I never fully got that working, actually,
12:21but I don't know, nobody ever told me not to do that.
12:25I had a couple advisors over the years,
12:28Hector, I'm sorry, he's passed away,
12:31but what a sweet heart he is.
12:34Yeah, what a nice guy.
12:35And then Jeff Ullman.
12:38And they, I don't know,
12:39I guess they would periodically ask me what I was doing.
12:43Yeah, they didn't really put many limits.
12:48When we moved to the new building,
12:50the gates building of computer science at SAIL,
12:52it was a little bit naughty back then.
12:55So I knew my lock picking days were over
12:58because you had these electronic keys with the little,
13:01you know, infrared things.
13:02Do you still have those?
13:04Or did they change 'em out?
13:06Still?
13:07- They're there.
13:08- Okay, so they were, the locks are not actually networked,
13:12so it actually trusts the key
13:15to tell it if it can open the lock,
13:17or at least back then they weren't networked.
13:18And I doubt they've changed since.
13:21But nevertheless, they were electronic and complicated.
13:25So just when we were moving into the building,
13:29there was a scaffolding outside.
13:31They were still finishing some stuff,
13:34but all the the doors, all the offices,
13:36they were locked with these electronic locks
13:38that I couldn't pick,
13:39except the balcony lock
13:43to the room that had the computer
13:44that would stamp out the keys.
13:47(people laughing)
13:49So I climbed out, I know the statute
13:51of limitations has passed.
13:52I hope so.
13:53I could tell the story.
13:54I climbed out, you know, from my office on the scaffolding.
13:59Like we, you know, I had to do this.
14:00I knew they were gonna take the scaffolding
14:01down like in the next week,
14:03- [Jennifer] Weren't we on the fourth floor?
14:04- Yeah. - [Jennifer] Okay.
14:06- But it was like a real, you know, scaffolding
14:08with like all the, I don't know, seems say I don't,
14:12I was a kid.
14:13That was the judgment I had.
14:14So anyway, I climbed over to the balcony,
14:17picked, that one was a physical lock,
14:19probably still is.
14:21Got into the computer there.
14:24I think I made a copy of all the software
14:27on the computer,
14:28like made myself a master key
14:30and then erased the copy.
14:32And then for a while the master key worked
14:34for kind of everything, but.
14:40- And I thought all you were doing was rollerblading up
14:42and down the hallways, which you were also doing.
14:44But yeah, okay.
14:45- But the scaffolding I didn't try with skates on,
14:48that would've been,
14:50that would've been too much.
14:52- Your career could have gone
14:54in so many different directions given the.
14:56(people laughing)
14:57- [Sergey] Yeah.
14:58- Education you were getting at the engineering school.
15:01But it didn't go in any, just any,
15:03the lock picking and career as a CIA agent or something,
15:10it didn't fully materialize.
15:13Tell us about how it did materialize.
15:17The, I mean.
15:18- I think what, you know, we worked on the ideas
15:22behind Google for a number of years,
15:24starting probably like in 95,
15:27and kudos to Larry to like really focusing
15:29on the link structure of the web.
15:32But at the time, the web was the new thing
15:35and everybody would do, you know,
15:38it was so easy to create some new idea.
15:42Like I think my first moneymaking idea
15:46was this pizza ordering, you know,
15:49it seemed crazy at the time
15:50that you could order food online.
15:51Nowadays we take it for granted.
15:54And as a joke, I like put a coke ad at the top.
15:57I thought it was ha, ha so funny.
15:58There'd be internet ads.
16:01But obviously, yeah.
16:03Turned out to be not that funny.
16:06(people laughing)
16:07Anyway, it failed quite profoundly
16:09because like the way it worked is you'd put in your order
16:13to the website and then, you know,
16:15pizza places were not online generally speaking,
16:18but I had this idea they had fax machines,
16:21so it would automatically send them a fax with the order.
16:24But then I realized they don't actually check their faxes
16:26very often and it flopped from there, so.
16:30That didn't particularly pan out,
16:32but it was, you know, at the time,
16:34like I guess all of us probably
16:38in the computer science department,
16:39like understood how the internet worked pretty well.
16:43How, you know, web servers worked.
16:44Like you could whip one of these out really quickly.
16:48So everybody was just drawing stuff on the web.
16:52It was just a very creative period.
16:54Anyway, so Larry was focused on the link structure.
16:56I was working on data mining at the time
17:01and we joined forces.
17:03And you know, soon enough we found we had something
17:07that was pretty useful for search,
17:11but we spent a while just experimenting with it at Stanford
17:15and like thinking to make an academic project or not.
17:19We tried to, you know, license it
17:21to various internet companies.
17:26One time we had pitched it to Excite
17:30and Vinod Khosla to his credit thought,
17:32hey, this is great,
17:33you guys should buy this.
17:34But Excite wasn't very interested.
17:36But we had an email back and forth with Vinod
17:39and we sent a note and said, okay, we'll you know,
17:44We'll license you the technology for $1.6 million.
17:47And we got a reply like 15 minutes later saying that,
17:51"Oh, that's a lot of moolah," but okay.
17:54And we're like excited.
17:55To graduate students, that was a lot of money.
17:58And then our friend Scott at the time,
17:59there were four us working on it.
18:02Scott and Alan were the other two.
18:04They went off to start their own companies.
18:06Anyway, Scott comes in laughing hysterically
18:09and it turns out he had faked the reply.
18:11'Cause back then you could,
18:12(people laughing)
18:14you couldn't send an email from anybody.
18:18So yeah, that deal didn't come to pass, obviously.
18:23But eventually Larry and I just decided that,
18:25hey, to scale this up,
18:27it would be really good to, you know, get actual money.
18:32And eventually we found some angels, which was very easy.
18:36And you know, my advisor Jeff,
18:39because for me, I was leaving a PhD program.
18:42My parents were disappointed,
18:44but he was like, well, why don't you just give it a try
18:46and if it doesn't work out, you come back.
18:48So I'm still on leave of absence technically,
18:50might still come back.
18:52We'll see how it goes.
18:53- I'm gonna come back to that later to see.
18:55- [Sergey] Yeah.
18:56- See if you might wanna think about coming
18:57back to get your degree.
19:00I love that part of the story.
19:03and it actually, it's interesting
19:04to think about in the context of today
19:06that entrepreneurship was sort of the last option to explore
19:09after you went out and tried to license and so forth.
19:14I mean you actually probably contributed
19:16a lot to changing that in some ways.
19:20- Yeah, I don't know.
19:21I mean, our journey was its own particular one.
19:26I mean, both Scott and Alan went off
19:28to run their own companies.
19:30That was part of the reason they left.
19:31Maybe they were impatient with us trying to take it out,
19:36but it wasn't like, you know, at the time a lot
19:39of people were starting companies, quite honestly.
19:43And like I said, they, Alan was kind of involved already
19:47with this weather company.
19:50I guess now known as Weather Underground,
19:51I think it got bought by,
19:53I can't remember, weather.com
19:54or one of the other ones recent, well, like in recent years.
19:57And Scott had this company that was archiving mailing lists
20:02and it was e-Groups
20:03and eventually got bought by Yahoo and so forth.
20:06But it was pretty common.
20:08But yeah, I don't know.
20:09We kind of, we probably took longer to make that kind
20:14of decision than a lot of other people would've.
20:19- So you, from that start, I mean, you look back now,
20:22of course, and Google's a $4 trillion company
20:26and you process 10 million searches every minute
20:32and are in a huge number of, I mean,
20:35just the enormous number of different product
20:38and so forth.
20:41At the time that you and Larry went to start the company.
20:47I mean, that none of that was obvious,
20:51at least to most people.
20:53And that's why I wasn't licensed and so forth.
20:56An you must have, you must have,
21:00you obviously made a lot of good decisions over the years
21:03to get from where you started to where the company ended up.
21:06Are there things that you did right at the beginning
21:09when you were creating Google that you look back
21:12and you think that was really important that we did
21:14that right from the start of the company?
21:24- I mean, I think early on,
21:28well, Larry was always very ambitious.
21:31He still is.
21:32In fact, there's almost no plan you can
21:34suggest to him that he won't say like,
21:37"Oh, that's not ambitious enough."
21:39You need, you know, not just the solar system,
21:41the galaxy, you know?
21:45So I think that's a little bit of sort of his passion.
21:51And, so yeah, we did have fairly early on the, you know,
21:56very ambitious mission statement.
21:59You know, to organize all the world's information
22:01and so forth.
22:05And, you know,
22:07I think that was a good kind
22:10of philosophy to start a company on.
22:14And also, we did start a fairly academic,
22:20I guess, minded company.
22:21I mean, we both came out of the PhD program,
22:25like a lot of the startups at the time
22:26were kind of out of college.
22:29I just do think that sort of shifts
22:32how you think about things a little bit.
22:36And you know, there are many phenomenal companies,
22:38just to be clear, that have come out of college.
22:43But you know, the investment in sort
22:46of foundational R&D and so forth,
22:48I do think was a part
22:49of the culture quite early on.
22:51- You hired a lot of PhDs as well,
22:53so it wasn't just the two of you.
22:56- Yeah, very much.
22:57I mean, I remember Urs Hölzle,
22:59who was one of our earliest guys.
23:02I knew him because I was on the Stanford, you know,
23:06the professor search committee.
23:08Like, like I'd already interviewed him.
23:11He got turned down for the Stanford job
23:14for, I don't know, whatever.
23:15It's complicated.
23:16You got a lot of good candidates.
23:17But like the moment he sent me a note, I was like,
23:21can you start tomorrow?
23:22I mean, because I already knew him
23:25and all of his qualifications.
23:30- I mean, you built, probably,
23:33I think there's a good argument
23:34that Google's been the most innovative company
23:38of the last 25 years in the whole world.
23:43And I mean, you can, and that by both in terms
23:45of product innovation,
23:48if you look,
23:48and a lot of great decisions went in there
23:51like building out video with YouTube
23:53and advertising with DoubleClick and Waymo
23:57and technical innovation going back to the beginning.
24:00But even now with chips, and and so forth.
24:06I'm curious about how you've,
24:08it's really hard for large companies
24:10to stay hugely innovative.
24:13Everyone has struggled with that
24:15and you've managed to do it.
24:17And many people, you know, attribute you personally
24:20with having a big impact there.
24:22How do you think about fostering a culture
24:25of innovation and your role in it?
24:28- Well, okay, thank you.
24:29You keep flattering me.
24:34I think that, well, first of all,
24:37we've definitely flopped on a bunch of things.
24:39We don't need to get into all of them right now,
24:42but we've had a long list of failures at the same time.
24:45So, you know, part of it is just trying.
24:49I think that because of the kind of academic roots,
24:58maybe we were more inclined to try hard things.
25:03And I think kind
25:07of coming into this last decade or so,
25:10especially the hard things have become more
25:12and more valuable.
25:16I guess I'm gonna, you know, if you look at AI,
25:19which obviously a huge trend,
25:21but like just, you know, the amount of computing that has
25:24to go into that,
25:26the amount of kind of deep math that has to go into that,
25:32those are all,
25:36you know, technically deep and challenging problems.
25:40And I guess it's just kind of a twist of fate
25:46that that turns out
25:48to be important at this stage in the world.
25:50I mean, there was a while where, you know, you could do,
25:52there was like pets.com, you remember?
25:54You can put anything on.com.
25:56It wasn't really that technically deep,
25:57marginal understanding of the web
25:59and you can do whatever.com.
26:02And you know, fortunately we were doing search,
26:05which did require some deeper technical skills,
26:12but the technical sophistication level has only gone up.
26:16And in fact, well now the people we hire are just,
26:21well, they're much more qualified than I am,
26:23or I was at the time.
26:25I was kind of a mathy computer science major
26:28because I had, like, during college,
26:31I did both math and computer science,
26:34which was somewhat unusual in my class.
26:40But nowadays, as we like hire people out of Stanford,
26:45and, you know, all the other top programs.
26:48Like these people are pretty sharp mathematically
26:53and computer science wise,
26:54and a bunch of 'em are physicists,
26:55because physicists kind of have to do the hard math
27:00and a lot of the stuff they do
27:02is like very computationally limited.
27:04So they need to kind of have some degree
27:06of the computation skills.
27:08So I just think somehow it has happened to be the case
27:12that some of the deep part tech
27:15has become increasingly important.
27:16And I think we just kind of lucked out
27:18on having set the bit early on in that direction.
27:22- That's an interesting observation
27:25that the technical problems have come to the fore again,
27:30as a competitive advantage for companies.
27:33So let's talk about AI for a minute.
27:35I mean, everyone's thinking about it.
27:37You're back at Google working on it.
27:42You guys are at the forefront in a whole bunch of ways,
27:46and it's incredibly competitive.
27:47I mean, the amount of capital
27:50that's going into AI infrastructure
27:52is hundreds of billions of dollars,
27:54even at the level of individual companies.
27:57It's really extraordinary.
27:59How are you seeing the landscape right now
28:01for what's going on in AI?
28:05- Okay, let me think how to answer
28:06that without just pounding my own chest.
28:12I mean, yes, it's a huge amount of investment for sure.
28:18I guess I would say in some ways we
28:23for sure messed up in that we underinvested
28:26and sort of didn't take it as seriously as we should have,
28:30say eight years ago when we published the transformer paper.
28:34We actually didn't take it all that seriously
28:36and didn't necessarily invest in scaling the compute.
28:39And also we were too scared to bring it to people
28:41because chatbots say dumb things.
28:44And you know, OpenAI ran with it, which good for them.
28:48It was a super smart insight
28:49and it was also our people like Ilya
28:53who went there to do that.
28:58But I do think we still have benefited
29:03from that long history.
29:04So we had a lot of the research
29:06and development of neural networks kind
29:10of going back to Google Brain.
29:13That was also kind of lucky.
29:16It wasn't, you know, it wasn't luck that we hired Jeff Dean.
29:20I mean, we were lucky to get him,
29:22but we were in this sort of mindset
29:25that deep technical things mattered.
29:27And so we hired him.
29:28We hired a lot of people from Deck,
29:30honestly, because they had the top research labs
29:34at the time.
29:37But he was passionate about neural networks
29:40and it stemmed,
29:42I think from actually his college experiment.
29:45I don't know, he's like, he was like, whatever,
29:47curing third world disease
29:49and figuring out neural networks when he was like 16.
29:52And he's done crazy things,
29:53but he was passionate about it.
29:57He built up a whole effort.
30:00And actually in my division at the time in Google X,
30:04we had him, but I didn't, I was like,
30:07okay, Jeff, you do whatever you want.
30:09He's like, "Oh, we can tell cats from dogs."
30:10I'm like, oh, okay, cool.
30:13(people laughing)
30:17But you know,
30:18you also trust your technical people.
30:21And soon enough they were developing all these algorithms,
30:24these neural nets that were, you know,
30:26doing some of our search.
30:27And then, you know, Nome came up with a transformer
30:31and we were able to do more and more.
30:34But yeah, I mean we, so we had the underpinnings.
30:36We had the RD.
30:38We did underinvest for a number of years
30:40and didn't take it as seriously as we should have.
30:44But we also, at the time, we had developed the chips for it,
30:49like the TPUs go back, I don't know,
30:5112 years or something like that.
30:52Initially we were using GPUs
30:55or probably also among the earliest use GPUs.
30:58And then we used FPGAs,
31:01and then we tried to develop our own chips,
31:03which have now evolved through a bazillion generations.
31:08So I guess it was that
31:13the trust into going
31:14after the deep tech, getting the more computation out,
31:18developing the algorithms.
31:20And in parallel we were big investors
31:24in compute for a long time.
31:26So we've had the data centers for a long, long time,
31:29kind of on a scale that I don't think,
31:31well Amazon AWS also,
31:33it does have very sizable data centers,
31:35but very few have, you know, that scale
31:39of data center.
31:41You know, have their own semiconductors.
31:44Have, you know, the deep learning algorithms
31:48and so forth to kind of all the components of the stack
31:51to be able to perform at the forefront of modern AI.
31:57- How are you thinking about,
31:59I mean, the technology keeps getting better every year.
32:03There's a set of people who,
32:06and there's a lot of different visions
32:08for what artificial intelligence is gonna look like.
32:13Like are AI's really gonna be able
32:15to do everything humans can do,
32:17at least in front of a computer and maybe more broadly.
32:21What will that world look like?
32:22Do you have a view on that,
32:23on where the technology is going?
32:27- I mean, it is absolutely amazing just the rate
32:31of innovation and it's hugely competitive,
32:34now, obviously, as all of you see
32:36between the top US companies, the top Chinese companies,
32:43and it's, yeah, I mean, like, you know,
32:45if you skip the news in AI for a month,
32:47you're like way behind.
32:48Like, you know.
32:52So where is it going to go?
32:55I mean, I don't, you know, I think we just don't know.
32:57Is there a ceiling to intelligence?
33:00I guess in addition to the question that you raised,
33:02like, can it do anything a person can do?
33:05There's the question like what things can it do
33:07that a person cannot do?
33:09- [Jonathan] Yeah.
33:09- That's sort of a super intelligence question.
33:11- [Jonathan] Yeah.
33:12- And I think that's just not known,
33:14like how smart can a thing be?
33:19You know, we've had however many hundreds of thousands
33:22of years of human evolution
33:24and I don't know, whatever, millions of primate,
33:26but that's a pretty slow process compared
33:33to what's going on with AI.
33:35- Do you think we're ready for the speed
33:37at which the technology is advancing?
33:40- Are we ready for the speed the technology is advancing?
33:46I mean, look, so far I think people are getting,
33:49you know, definitely great use out of technology.
33:51I think even though there
33:54are doom and gloom forecasts here and there,
33:56like everybody's pretty well empowered.
34:00And the AI's, truth be told,
34:03are periodically dumb enough
34:06that you're always like supervising them anyway.
34:11But occasionally they're brilliant
34:12and give you a great idea.
34:14And occasionally, especially as a non-expert,
34:22well like, whatever,
34:23if I want to figure out how to create a new AI chip,
34:26I guess I could talk
34:27to our expert designers and stuff.
34:29But as a base case, I can at least,
34:31I can whip on my phone,
34:32I can talk to an AI about it.
34:34It'll probably give me a 90%, 80%, 90% decent sort
34:39of overview and understand it,
34:43or whatever, my health questions or whatnot.
34:47I mean, I do think it makes individuals very empowered,
34:51because generally speaking, you don't have
34:55like experts in X, Y, Z all around you all the time.
34:59And I think that empowerment can create a lot of potential,
35:04whether it's, you know, career or enterprise
35:08or health or living well.
35:12So look, I mean I don't think I have all the answers.
35:15I just do think it has a huge potential
35:17to improve individual capability.
35:20- Yeah, that's certainly the positive vision.
35:23That it could be an incredible augmenter
35:26of human capability.
35:28It's great that you're thinking about it that way.
35:31Let me ask a question, I think, which is always
35:34asked in the entrepreneurial thought leaders class,
35:39but is maybe particularly salient with the discussion of AI.
35:43Because one of the things I think every student at Stanford,
35:47and probably every college age student
35:49in the country is thinking about,
35:51is how will this technology affect
35:53their careers and their job opportunities
35:56and what they might go on to do.
35:59I'm curious if you have any advice for the students
36:01about what they ought to be studying
36:03or what they ought to be thinking about as they look at it,
36:06the job market and the future.
36:09- I mean, I think it's super hard
36:11to predict exactly what'll happen.
36:13I think, you know, if we look at from the advent
36:16of the web to cell phones and so forth,
36:19those have transformed our society profoundly,
36:22have transformed the kinds of jobs and careers
36:26and studies people do for sure.
36:28And AI will a hundred percent change that.
36:31But I think it's very hard right now
36:35in a rapidly shifting landscape to say exactly what.
36:39And also the AI we have today is very different
36:42from the AI that we well had five years ago,
36:45or the AI we are going to have in five years.
36:49So, yeah, I don't know.
36:53I think it's tough to really forecast.
36:56I mean, I would for sure use AI to your benefit.
36:58There are just so many things that you can do.
37:05Yeah, I mean, just myself as an individual,
37:07like I just, whether it's, you know,
37:11choosing a gift for my friends or family
37:14or brainstorming new ideas for products
37:17or what have you, or for art or something like that.
37:21Like I just turn to AI all the time now
37:23and it doesn't do it for me,
37:27because I always, you know, I typically will ask,
37:29give me five ideas, blah, blah, blah.
37:30And you know, probably three
37:32of them are gonna be junk in some way
37:33that I'll just be able to tell.
37:35But two will have some grain of brilliance,
37:41and you know, or possibly put it in perspective for me
37:46or something like that, that I'll be able to refine,
37:49think through my ideas.
37:51- Let me jump in with a really concrete question.
37:53So we have about 250 students out there.
37:56A lot of them are undergraduates.
37:58A great number of them have not selected their major yet
38:01because we give them a lot of flexibility here at Stanford
38:04for the undergraduates as well.
38:07A few years ago we could predict
38:09that a large number would choose computer science
38:12as their major.
38:13Are you recommending they pick,
38:14continue to pick computer science as their major
38:17and they're listening closely?
38:20- I mean, I chose computer science
38:23because I had a passion for it.
38:24So it was kind of a no-brainer for me.
38:30I guess you could say
38:31I was also lucky,
38:32because I was also in such a transformative field.
38:42I wouldn't like not choose computer science
38:45just because, you know,
38:48AI can be decent at coding nowadays.
38:54AI is pretty decent at a lot of things.
38:57Coding just happens to have like a lot of market value,
38:59which is why a lot of people pursue it.
39:01And furthermore, you know,
39:05better coding makes for better AI.
39:06So a lot of the companies like our own,
39:08that work on it care a lot about it.
39:10Like we use it a lot for our own coding
39:13and even for our algorithmic ideas and so forth.
39:18But that's because it's such an important thing.
39:22So I guess I wouldn't,
39:28I wouldn't go off and like switch to comparative literature
39:32because you think the AI is good at coding.
39:35The AI is probably even better comparative literature,
39:38just to be perfectly honest anyway.
39:39- [Jonathan] Yeah. - Like, I don't mean
39:43to disrespect comparative literature majors,
39:45but just like, you know, when you,
39:49when the AI writes the code,
39:51and just be honest, sometimes doesn't work,
39:52like it'll make a mistake that's pretty significant.
39:56Like, you know,
39:57getting sentence wrong in your essay about comparative
39:59literature isn't gonna like really have that consequence.
40:01So it's honestly easier for AI to do some
40:05of the, you know, creative things.
40:08- I think it's a very interesting observation
40:10about the technology.
40:11'Cause I think, you know, one inclination
40:13to say about AI is it's gonna be really good at solving
40:17these technical problems,
40:18but it won't necessarily do
40:19the things we associate with humans.
40:21Like being empathetic in a conversation.
40:24And if you ask one of these AI engines to say,
40:28simulate a conversation,
40:29it's actually pretty good at doing a lot of sort
40:32of giving you the structure for a complicated conversation.
40:35So I think that actually, I like that you're pointing
40:39to that uncertainty.
40:42One more question and then I wanna open up
40:44to the audience so that we give people a chance
40:46to ask questions.
40:47So this is the hundredth anniversary
40:49of the School of Engineering.
40:53If you were,
40:57if you were Jennifer
40:59and had to launch the school's second century,
41:02what would you be thinking about for the second century
41:05of the school of engineering?
41:07- Wow, okay.
41:08That's a big responsibility to like kind of plan.
41:13- [Jennifer] The dean job will come open.
41:15- It's a big responsibility.
41:16(people laughing)
41:25I mean, I guess I just would rethink what it means
41:31to have a university.
41:33I mean I know that sounds kind of annoying.
41:37(people laughing)
41:37That's the kind of thing Larry would say
41:39and I would be really annoyed with him.
41:41But you know,
41:43I mean we have this geographically concentrated thing
41:46and there's like the buildings
41:49and the fancy lecture halls.
41:54That really annoying blinking light, sorry, can't help it.
41:58You guys need to fix your,
42:01but realistically now, you know,
42:07information spreads very quickly.
42:09And many of the universities have, you know, obviously,
42:12whatever gone online, including Stanford.
42:14But, you know, MIT with the open courseware early on
42:17and all these startups that have gone this way, whatever,
42:21Coursera, Udacity, you name it.
42:23So the teaching is sort of getting spread
42:28and anybody can go online now and learn about it.
42:30You know, you can talk to an AI or take one of these classes
42:35and watch the YouTube videos.
42:40So I guess, yeah, what does it mean to have a university?
42:45Are you trying to maximize the impact in that case?
42:50You know, probably just limiting the geographically
42:53is not gonna be so effective.
42:55To be fair, I guess, you know,
43:00in the Bay area is kind of a special place.
43:02But yeah, I mean, I know I'm kind of rambling here
43:07and thinking through, but yeah,
43:11I just, I don't know that
43:12for the coming century that, you know,
43:15the idea of school of engineering
43:17and the university is gonna mean the same thing
43:21as it used to.
43:22I mean, people move around, work remotely,
43:26collaborate across.
43:30It's a little bit at odds,
43:30'cause we're trying to get people actually into the office
43:32and I think they do work better in person together,
43:35but that's at a certain particular scale.
43:40Like at some level if you have a hundred people together
43:43over there, it's like kind of fine.
43:45They don't have to be at the same place
43:46as these other a hundred people.
43:54And increasingly I do see sort of individuals kind
43:57of who create new things sort of regardless of degree.
44:04I mean, in as much as we've hired a lot of academic stars,
44:08we've hired tons of people who don't have bachelor's degrees
44:11or anything like that and they just figure things out,
44:16you know, on their own in some weird corner.
44:21I don't know. I, I think it's really a hard question.
44:25Yeah, I guess I don't feel like I'm gonna magically deliver
44:28you like the new recipe,
44:29but I just think this format is likely to be
44:35the one for the next a hundred years.
44:41- You took that in a deeper direction than I was.
44:44- [Sergey] Oh, sorry.
44:45- No, it was great actually.
44:46It was a bit deeper.
44:47- It sounded more presidential than Dean like.
44:49I think he's talking to you.
44:52- I agree.
44:53- It applies to the whole university.
44:54- You actually surface
44:56the most fundamental questions about the university.
44:58Which is that that the, you know, part
45:00of the university is about the creation
45:03and transmission of knowledge.
45:04That's the fundamental mission.
45:06Those can be done in different ways as technology advances.
45:10And then there's a question about the model of having kind
45:13of a density of talent all in one place,
45:15bumping into each other, sort of, you know,
45:17which, of course, was what led to you creating Google
45:20and has led to a lot of great things.
45:21And will there be substitutes for that
45:26kind of ecosystem that gets created on a university campus
45:30and, or, you know, how fundamental is that?
45:35And will it continue to be?
45:36So I actually thought that was,
45:38I appreciate that you surface such a deep question
45:40in this session.
45:42All right, I wanna make sure we give some questions
45:44for other folks out in the audience.
45:46So Jennifer, I'm gonna turn to you
45:47to take some questions from the folks out here.
45:51- Yes, so the students
45:52in the Entrepreneurial Thought Leaders class
45:54submitted questions in advance
45:56and a number of those were selected.
45:59And so with the time we have left, we're going
46:01to have a few questions from our students.
46:04I think the first one is over here.
46:07- Dean Widom, President Levin and Sergey Brin,
46:10thank you for your time.
46:11My name is Rasha Barve from Kansas City
46:13studying MSENIR.
46:15And my first question goes out to Sergey.
46:17It actually just touches on what we were discussing.
46:20Google largely grew out
46:21of the academic work you authored on PageRank,
46:24and with industry now driving so much of today's innovation,
46:28do you still feel that the academia
46:30to industry pipeline is crucial?
46:32And if so, how might you strengthen it?
46:36- Wow, that's a great question.
46:40Is the academia to industry pipeline crucial?
46:46Yeah, I'm gonna give you an I don't know on that,
46:50because I guess,
46:55you know, when I was a grad student,
46:59the sort of time from some new idea
47:06to it being maybe commercially valuable was many decades.
47:15If that compresses, then it no longer makes
47:18as much sense to do.
47:21I mean, in academia you have freedom
47:22to think about it for a while.
47:25Whatever, you apply for grants, you do this and that
47:27and you can kind of spend couple decades thinking about it
47:31and then it percolates.
47:32And then, you know, eventually maybe there's
47:35some big company or your startup kind of pursues it.
47:39The question is, does that make sense
47:43if that timeline shrinks a lot?
47:48I think there are,
47:51I think there are certain things
47:54that for sure make sense
47:55and I definitely, you know, even within AI,
48:01you know, periodically keep up with the Stanford research
48:03and other universities.
48:06Occasionally we, whatever, hire those folks
48:10and collaborate with 'em and whatnot.
48:15But I guess I don't know that they needed to have
48:20that sort of period of time
48:24that they were trying, whatever,
48:25some new attention thing let's say.
48:28And they spent a couple years experiment
48:31and then, you know, they took it
48:32to industry in one form or another.
48:35I mean obviously industry is also doing all those things.
48:38So probably not a huge argument for that.
48:42Radical sort of new architectures and things, maybe.
48:50But it's sort of, you know, the time
48:52that industry will scale, it will be much faster.
48:57I guess, you know, quantum computing comes to mind.
49:00There was sort of first brainstormed, I dunno,
49:04when did Definement, like in the eighties or something
49:06kind of postulate this idea
49:07of quantum computing
49:10and now there are a bunch of companies
49:11that are included.
49:13They're sort of doing it.
49:15There are also university labs
49:16that try like new ways to do it.
49:20That's kind of like on the fence maybe.
49:24I guess I would say if you have some completely new idea,
49:28like you're not doing superconducting qubits like we are,
49:30or whatever, the trapped ions like a bunch
49:35of startups are,
49:38but you have some new way.
49:42Maybe you need to let it marinate in university
49:45for some number of years.
49:46Those things are kind of hard.
49:48It could make sense,
49:51but then at some point if you decide it's really compelling,
49:53you're probably gonna go ahead
49:54and take it commercial in one way or another.
50:00Yeah, I don't know.
50:01I want to give you a clear cut answer,
50:03because you know,
50:06the top companies now do invest
50:07in much more fundamental research
50:10and I think it's sort of with AI starting
50:13to pay off that, you know, those investments are paying off.
50:20So I guess it would shift the proportion
50:24of endeavors that you would do.
50:27But I do think there are still some things that do,
50:31that do take, you know, like the decade of kind
50:33of more pure research
50:36that maybe companies are gonna be more reluctant to pursue
50:41because that's just too long in a time to market.
50:45- All right, next question I think is over here.
50:49- Hi everyone, my name is Arnov
50:51and I'm a freshman studying computer science and math.
50:54My question is for Sergey Brin.
50:57As AI accelerates at this unprecedented rate,
51:01what mindset should young aspiring entrepreneurs
51:05like myself adopt to avoid repeating earlier mistakes?
51:11- Oh, what mindset should you adopt
51:13to avoid repeating earlier mistakes?
51:26Yeah, when you have it like your cool
51:27new wearable device idea, really fully bake it
51:31before you have a cool stunt involving
51:34skydiving and airships.
51:38That's one tip.
51:44No, I actually, I like the sort of what we were doing back
51:47in the day for Google Glass that's like, you know,
51:49an example of prior mistakes.
51:52But I think I tried to commercialize it
51:54too quickly before, you know,
51:58we could make it more, you know,
51:59as cost-effectively as we needed to
52:01and as polished as we needed to
52:02from consumer standpoint and so forth.
52:05I sort of, you know, jumped the gun
52:07and I thought, oh, I'm the next Steve Jobs,
52:09I can make this thing, ta, da.
52:14That's probably one.
52:15I guess, yeah, if I encapsulate it,
52:19yeah, everybody thinks they're the next Steve Jobs.
52:21I've definitely made that mistake.
52:23But you know, he was a pretty unique kind of guy.
52:28So, yeah,
52:32I guess I would say, you know,
52:34make sure you've baked your idea long and developed it
52:38sort of long enough to far enough point before,
52:44there's sort of a treadmill you get onto
52:46where you're sort of,
52:47outside expectations increase, the expenses increase,
52:51and you're sort of then,
52:52you kind of have to deliver by a certain time.
52:57And you might not have that, you know,
52:58you might not be able
52:59to do everything you to do in that amount of time.
53:01You kind of get this snowball, I guess, of expectations
53:06that happens and you don't give yourself all the time
53:09you need to process them.
53:10That's the mistake I would've tried to avoid.
53:13- All right. I think we're gonna go over to this side.
53:18- I thank you for the talk.
53:20This question is, so my name is Esha Bargetag.
53:22I'm an undergraduate freshman at Stanford University.
53:25This question is for Sergey Brin and Jennifer.
53:28So we see a lot of AI companies
53:30improving large language models
53:31via scaling data and scaling compute.
53:34My question is, once we do run out of data
53:37and once we do run out of compute,
53:39what do you think will be the next direction?
53:41Would it be in newer architecture,
53:43something an alternative to transformers?
53:45Or would it be a better learning method,
53:47something better than like supervised learning
53:50or RL that we use to train these large language models?
53:53Or is it a completely different direction
53:55that you have thought of before?
53:57Thank you.
53:59- Yeah, I take it from my point of view.
54:00I mean all of the things that you listed,
54:03I would say have already been bigger factors
54:07than scaling computer, scaling data.
54:11I think that's sort of what people notice the scaling,
54:15because you're like building data centers
54:17and buying chips and like,
54:21well there were all the publications from OpenAI
54:25and Anthropic about like different
54:27kinds of scaling laws.
54:28So I think that attracts a lot of attention.
54:31But I think if you carefully line things up,
54:34you'll see that actually
54:36the algorithmic progress as outpaced,
54:39even the scaling, over the last decade or something.
54:43At some point, actually a while ago, many,
54:45while I was in grad school, I think I saw this kind of plot
54:48for like the N-body problem.
54:51You know, like if you have gravitation,
54:52they're all flying around and it actually, you know,
54:57there's been huge, you know,
54:59Moore's law increase in compute over since people started
55:02worrying about that in the fifties to, I don't know,
55:04by the time I read about in the nineties,
55:06but actually the algorithms to do the N-body problem
55:10far outpaced that compute scale up.
55:14So I think you're gonna find that, you know,
55:17companies like ours are never gonna
55:19turn down being at the frontier of compute.
55:24But that's, yeah, that's just sort of an,
55:29that's the dessert after, you know, your main course
55:33and the veggies of actually having
55:35done your algorithmic work.
55:37- I guess I'll jump in
55:39and say that in terms of running out of compute
55:43or running out of data,
55:44or specifically running out of compute,
55:45we're very familiar with that here already.
55:48It's actually a issue that it's difficult for a university
55:52to have the type of compute that the companies have.
55:56We don't even come close.
55:57But that does lead us to do quite a bit
56:00of innovative work in
56:01what happens when you have less compute
56:03and how to make, you know, more of less.
56:06So we do do a lot of that work here already.
56:11Next question I think also on this side,
56:15- Hi everyone, my name is Andy Zivortsy.
56:18I'm a second year graduate student in chemical engineering.
56:21My question is to all the speakers,
56:26which emergent technology do you think
56:30is seriously being underestimated in terms
56:34of its long term impact?
56:37Thank you.
56:41- Whew, okay.
56:42What emergent technology is being seriously underestimated?
56:45Wow.
56:50Okay, I probably, obviously can't say AI
56:52because it's hard to argue.
56:53But it could be underestimated.
56:55- [Jonathan] It could be underestimated.
56:56- It could be underestimated,
56:57but probably not emergent at this point.
56:59We couldn't use that one.
57:01I mean a lot of people do wonder about quantum computing.
57:05What it will bring.
57:08It's probably not what I would hang my hat on
57:10to answer that question.
57:12Although I definitely support
57:14sort of our efforts in quantum computing and so forth.
57:18But there are many unknowns.
57:20I mean, technically speaking we don't even know
57:23if p is not equal to np.
57:25Like on the computation front,
57:27there's just so many unanswered questions
57:31And the quantum algorithms are, you know,
57:33specific for particular,
57:36very particular structured problems.
57:41That side I'm a big proponent,
57:43but it's hard to put my finger on that.
57:50I mean, perhaps the applications of both AI,
57:53and for that at quantum computing,
57:55to materials science,
57:57because what could we do with different kinds of materials
58:00that are better in a whole host of ways?
58:03I mean, kind of the sky's the limit.
58:08- I was thinking of materials as well actually,
58:10but partly 'cause a lot of,
58:13the underestimated is sort of interesting.
58:15There's so much attention right now on, you know,
58:18like what are the opportunities
58:20for technological innovation?
58:22So many technologies that aren't there yet,
58:24like fusion energy or quantum,
58:27they probably, it'd be hard to say
58:30that people are missing them
58:31and not paying any attention to them right now and AI,
58:33but I think materials would be,
58:36in my mind, would be one of them.
58:39And probably some of the opportunities
58:43in biology and health of which there are many
58:48in molecular science that that's a.
58:52It's probably getting less attention than AI right now.
58:54But there's also a huge revolution in molecular science.
58:59- Yeah, I was gonna say exactly the same thing.
59:00I kind of watched the spotlight move around
59:03and the spotlight is very large on AI right now,
59:07but it was shining on biology
59:10and it shouldn't stop shining on it.
59:12There's all kinds of things going on in synthetic biology,
59:15very exciting things.
59:16So I think we need to broaden that spotlight a little bit.
59:21Okay, over here.
59:24- Hi, my name is Drew Mi
59:25and I'm a student coming from Singapore.
59:28My question today is for Sergey
59:29and it's a bit more personal.
59:32So we all grew up having limiting beliefs
59:34and I was curious of what limiting beliefs
59:37or deeply held beliefs you had while building Google
59:40that you had to change
59:42and how did that affect your decision-making?
59:44Thank you.
59:46- Huh, limiting beliefs.
59:51Yeah, I guess like, I, I had a very,
1:00:00like my life expanded pretty dramatically
1:00:03at a much of stages.
1:00:04Like I was born in Moscow in the Soviet Union
1:00:08and it was very different.
1:00:09You know, very poor, where everybody was very poor.
1:00:13And I lived in a little 400 square foot apartment
1:00:18with my parents and my grandmother
1:00:19and had to walk up five flights of stairs.
1:00:25I dunno, didn't really think about the world outside.
1:00:29I guess I was lucky that my father kind of got a hint
1:00:33of the world outside.
1:00:34He, I guess went to some conference in Poland
1:00:39where they told him what the western world was like
1:00:41and he decided to move us,
1:00:43which was very controversial at the time in the family.
1:00:49But eventually we got to the US
1:00:50and we're still, you know, very poor
1:00:53and had to make our way out of having nothing.
1:00:58And, you know, at time I had to learn a new language,
1:01:01give up, you know, I had to meet, make all new friends.
1:01:07So it was sort of challenging transition, but awakening.
1:01:11And I think when I came to grad school at Stanford,
1:01:13it was sort of a similar,
1:01:15like now I had sort of this,
1:01:19all this freedom in the way the professors entrusted me
1:01:24and just something about California
1:01:26that was very freeing and liberating in thought,
1:01:31given the tradition of the state.
1:01:35One that we're a little bit getting away from in California,
1:01:37if I'm being honest.
1:01:38But I'm not gonna complain about that.
1:01:45But I guess it's this experience, I guess.
1:01:54I guess I'm answering question backwards,
1:01:56not like really a limiting belief.
1:01:58I guess I had had the experience of expanding
1:02:03my world in ways that seemed very painful at the times,
1:02:07but later paid off just because of my personal history.
1:02:13And I guess, you know,
1:02:16those challenging transitions can pay off.
1:02:26- Right, next question.
1:02:28- Hello, thank you to all of you for being here.
1:02:32My name is Lu Baba,
1:02:33I'm a second year master student
1:02:35in management, science and engineering.
1:02:37Originally from Casablanca, Morocco.
1:02:40My question is also for you Sergey,
1:02:41it's also more on the personal side.
1:02:44So you've achieved success at a scale
1:02:48most people never experience.
1:02:51Looking at your life now,
1:02:54what is your definition of a good life?
1:02:56What does it mean to you beyond all these accomplishments?
1:03:01Thank you.
1:03:02- Okay, thanks, what is the definition of a good life?
1:03:10Well I guess that's, you know,
1:03:13being able to enjoy your life, you know,
1:03:15whatever you build.
1:03:17I like to have family.
1:03:19Have one of my kiddos here.
1:03:20My girlfriend is here.
1:03:23You know, I feel grateful to be able
1:03:26to spend quality time with them.
1:03:32I do feel quite grateful to be able
1:03:35to be intellectually challenged sort of at this stage.
1:03:39I actually retired like a month before COVID hit
1:03:44and it was like the worst decision.
1:03:48I had this vision that I was gonna sit in cafes
1:03:52and study physics, which was my passion at the time.
1:03:56And yeah, that didn't work 'cause there were no more cafes.
1:04:01(people laughing)
1:04:03And yeah, I don't know, I was just kind of stewing
1:04:08and kind of felt myself spiraling,
1:04:12kind of not being sharp.
1:04:14And then I was like, oh I gotta get back to the office,
1:04:17which at the time was closed.
1:04:18But you know, after a number of months we started
1:04:20to have some folks going to the office
1:04:22and I started to do that occasionally
1:04:25and then started spending more and more time
1:04:28on what later became called Gemini,
1:04:31which is super exciting.
1:04:32And to be able to have that
1:04:35technical creative outlet,
1:04:37I think that's very rewarding,
1:04:40as opposed to if I'd like stayed retired.
1:04:42I think that would've been a big mistake.
1:04:46- All right, I think we have time for one or two more.
1:04:48I believe over on this side.
1:04:51- Hello. thank you guys all so much for being here.
1:04:53My name's Stanley Leo.
1:04:53I'm a freshman planning on studying
1:04:55management, science and engineering.
1:04:56And I had a question for all three of you.
1:04:58So for some context, like before arriving here,
1:05:00I was absolutely terrified.
1:05:01'Cause everyone here is like super talented.
1:05:02I'm like, what is going on?
1:05:04Like I have no clue like why I'm here.
1:05:06And everyone just seems way too smart for me.
1:05:08But after getting to know people,
1:05:09I realized they're all just really
1:05:10relatable and normal people.
1:05:12So for all three of you, you guys are viewed as like some
1:05:14of the best leaders, innovators in the world.
1:05:17But if there's one thing you'd like to share
1:05:18that is reassuringly relatable
1:05:20and human about yourself, what would that be?
1:05:22Yeah.
1:05:26- You wanna start Sergey?
1:05:29(people laughing)
1:05:31Okay, I'm gonna share it
1:05:32and then I'm going to try to undo it, but okay.
1:05:38I realize that sometimes I'm embarrassed to ask things.
1:05:41I don't know, but I will go ahead.
1:05:44Wait, what is management, science and engineering like?
1:05:48(audience laughing)
1:05:50Is it like a Dilbert kind of like,
1:05:53I'm gonna manage, how does that work?
1:05:57- [Stanley] It's the class. - It's a class.
1:06:00- [Jennifer] It's a major - [Sergey] Wait, this class?
1:06:02- [Jennifer] It's a department.
1:06:02- It's a class of management, science.
1:06:03- Is that this class?
1:06:04I guess I should have read the details more.
1:06:06- [Stanley] It's a major.
1:06:07- It's called management science and.
1:06:09It's the department?
1:06:10- It's a department, yes. - But what do you study?
1:06:14(people laughing)
1:06:17Like what are the classes?
1:06:19- So management, science and operate,
1:06:22and engineering. I'm going to just.
1:06:24- Okay.
1:06:25- Say they just had their 25th anniversary,
1:06:29but they were the merger of three departments,
1:06:32industrial engineering, operations research,
1:06:35and engineering economic systems.
1:06:36So I think that describes.
1:06:39- [Sergey] Okay, okay.
1:06:40- That sort gives you a little triangle there
1:06:41of what they do.
1:06:42So some universities will have an industrial
1:06:44engineering or operations research.
1:06:46We have this all bundled together here
1:06:48in Management Science and Engineering,
1:06:50which is the department that sponsors
1:06:52the Entrepreneurial Thought Leaders seminar,
1:06:55which is what we are.
1:06:56- [Sergey] Okay. - Conducting right now.
1:07:00- All right, well I guess I didn't really know that,
1:07:02so that's my embarrassing truth, but I'm glad I asked.
1:07:05- What makes me relatable is that I can explain things
1:07:08to Sergey Brin, pay attention to them.
1:07:11(people clapping)
1:07:16I'll let you off the hook, John,
1:07:17and we'll go to our last question.
1:07:19Do we have one more question?
1:07:22I think we do, yeah.
1:07:23- We can ask one more.
1:07:24Hi, my name's Zena.
1:07:26I'm actually the course assistant for the class.
1:07:28So thank you for being here
1:07:29and it's a great thing that we can have this last class.
1:07:32I'm gonna ask you something that we ask a lot
1:07:34of our speakers usually is to give a recommendation
1:07:37to the students as to what do you do with your time
1:07:41to stay on top of things.
1:07:42And you just said you really like staying sharp
1:07:44and being on top of what's happening in AI and whatnot.
1:07:46So what books do you read?
1:07:48What podcasts do you listen to in your car?
1:07:52- Okay, I'm gonna try to do this without advertising or so.
1:07:55Well, okay, so the thing I like to do,
1:07:59but you shouldn't do it now
1:08:00because we have like way better version coming,
1:08:03but I do talk to Gemini live in the car often and I ask,
1:08:08but the publicly available version right now
1:08:10is not our good version.
1:08:12So like, you shouldn't do it today,
1:08:14but give me a few weeks to actually ship
1:08:18what I have access to,
1:08:20because we have like an ancient model
1:08:22behind it in the publicly research version right now.
1:08:24It's a little embarrassing, but I do like ask it like,
1:08:29you know, whatever, I want to develop a data center,
1:08:33you know, whatever, I need how many hundreds of megawatts
1:08:37of this kind of power, that kind of power,
1:08:38how much it's gonna cost.
1:08:40And I just talk to it about stuff on my drive.
1:08:45Okay, that does seem kind of self advertising with Gemini.
1:08:52I mean, I do periodically listen
1:08:53to a whole bunch of podcasts.
1:08:56The "All In Guys," are actually one of my favorites
1:09:01and they're great hosts.
1:09:05We just visited Ben Shapiro, another podcaster down in,
1:09:08we were in Florida, got to see his studio.
1:09:12I mean a bunch of these podcasters are actually pretty
1:09:15fun to meet in person.
1:09:18But yeah, I guess, okay,
1:09:19that's not how you're gonna learn about it.
1:09:22You're gonna,
1:09:23but I do just listen to them, see what's up.
1:09:26But I do prefer to have an interactive discussion
1:09:31on my drive, so that's why I talk to the AI,
1:09:33as embarrassing as that sounds.
1:09:37- [Jennifer] Okay.
1:09:38- [Jonathan] Sort of a glimpse of the future,
1:09:39I think actually.
1:09:40That's a good way to end.
1:09:42We'll probably all be doing it.
1:09:43- [Jennifer] So - Soon.
1:09:44- Might, yes.
1:09:46So thank you John.
1:09:47Thank you, Sergey.
1:09:49I also wanted to thank Emily Ma.
1:09:51Emily is a Stanford adjunct lecturer.
1:09:53(people clapping)
1:10:01Emily is a co-instructor of the course.
1:10:04She's also a Google employee,
1:10:06and she saw the potential
1:10:07for this event and partnered with us.
1:10:09So thank you very much.
1:10:10Thank you all for being here,
1:10:12for celebrating the School of Engineering's 100th year.
1:10:16This was a perfect way to close out our first century
1:10:20and let's see what happens next.
1:10:22Thank you.
1:10:23- [Sergey] Thank you. - [Jonathan] Thank you.
1:10:26(people clapping)
1:10:27- [Jonathan] Congratulations.

About this transcript

This page contains the full transcript of Big ideas begin here: Sergey Brin at Stanford by Stanford University School of Engineering, generated from the public captions YouTube serves with the video. The transcript has 10,436 words across 1,467 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to translate the transcript into 100+ languages.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.