Datasets:

shubhamg2208
/

lexicap

	WEBVTT

	00:00.000 --> 00:02.800
	The following is a conversation with Eric Schmidt.

	00:02.800 --> 00:07.640
	He was the CEO of Google for ten years and a chairman for six more, guiding the company

	00:07.640 --> 00:13.080
	through an incredible period of growth and a series of world changing innovations.

	00:13.080 --> 00:19.280
	He is one of the most impactful leaders in the era of the internet and the powerful voice

	00:19.280 --> 00:22.400
	for the promise of technology in our society.

	00:22.400 --> 00:27.760
	It was truly an honor to speak with him as part of the MIT course on artificial general

	00:27.760 --> 00:32.040
	intelligence and the artificial intelligence podcast.

	00:32.040 --> 00:37.120
	And now here's my conversation with Eric Schmidt.

	00:37.120 --> 00:40.840
	What was the first moment when you fell in love with technology?

	00:40.840 --> 00:47.080
	I grew up in the 1960s as a boy where every boy wanted to be an astronaut and part of

	00:47.080 --> 00:49.000
	the space program.

	00:49.000 --> 00:54.400
	So like everyone else of my age, we would go out to the cow pasture behind my house,

	00:54.400 --> 00:58.640
	which was literally a cow pasture, and we would shoot model rockets off.

	00:58.640 --> 01:00.200
	And that I think is the beginning.

	01:00.200 --> 01:05.680
	And of course, generationally, today, it would be video games and all the amazing things

	01:05.680 --> 01:09.280
	that you can do online with computers.

	01:09.280 --> 01:15.080
	There's a transformative, inspiring aspect of science and math that maybe rockets would

	01:15.080 --> 01:17.560
	bring, would instill in individuals.

	01:17.560 --> 01:21.720
	You've mentioned yesterday that eighth grade math is where the journey through Mathematical

	01:21.720 --> 01:24.520
	University diverges from many people.

	01:24.520 --> 01:27.080
	It's this fork in the roadway.

	01:27.080 --> 01:31.160
	There's a professor of math at Berkeley, Edward Franco.

	01:31.160 --> 01:32.720
	I'm not sure if you're familiar with him.

	01:32.720 --> 01:33.720
	I am.

	01:33.720 --> 01:35.400
	He has written this amazing book.

	01:35.400 --> 01:41.960
	I recommend to everybody called Love and Math, two of my favorite words.

	01:41.960 --> 01:49.680
	He says that if painting was taught like math, then students would be asked to paint a fence,

	01:49.680 --> 01:54.520
	which is his analogy of essentially how math is taught, and you never get a chance to discover

	01:54.520 --> 01:59.400
	the beauty of the art of painting or the beauty of the art of math.

	01:59.400 --> 02:05.240
	So how, when, and where did you discover that beauty?

	02:05.240 --> 02:12.040
	I think what happens with people like myself is that your math enabled pretty early, and

	02:12.040 --> 02:16.640
	all of a sudden you discover that you can use that to discover new insights.

	02:16.640 --> 02:22.120
	The great scientists will all tell a story, the men and women who are fantastic today,

	02:22.120 --> 02:25.560
	that somewhere when they were in high school or in college, they discovered that they could

	02:25.560 --> 02:30.760
	discover something themselves, and that sense of building something, of having an impact

	02:30.760 --> 02:35.520
	that you own drives knowledge, acquisition, and learning.

	02:35.520 --> 02:41.000
	In my case, it was programming, and the notion that I could build things that had not existed

	02:41.000 --> 02:46.560
	that I had built, that it had my name on it, and this was before open source, but you could

	02:46.560 --> 02:49.160
	think of it as open source contributions.

	02:49.160 --> 02:53.760
	So today, if I were a 16 or 17 year old boy, I'm sure that I would aspire as a computer

	02:53.760 --> 02:59.000
	scientist to make a contribution like the open source heroes of the world today.

	02:59.000 --> 03:03.760
	That would be what would be driving me, and I'd be trying and learning and making mistakes,

	03:03.760 --> 03:06.720
	and so forth, in the ways that it works.

	03:06.720 --> 03:12.360
	The repository that GitHub represents and that open source libraries represent is an

	03:12.360 --> 03:17.840
	enormous bank of knowledge of all of the people who are doing that, and one of the lessons

	03:17.840 --> 03:22.240
	that I learned at Google was that the world is a very big place, and there's an awful

	03:22.240 --> 03:26.360
	lot of smart people, and an awful lot of them are underutilized.

	03:26.360 --> 03:32.240
	So here's an opportunity, for example, building parts of programs, building new ideas to contribute

	03:32.240 --> 03:36.640
	to the greater of society.

	03:36.640 --> 03:41.000
	So in that moment in the 70s, the inspiring moment where there was nothing, and then you

	03:41.000 --> 03:44.800
	created something through programming, that magical moment.

	03:44.800 --> 03:50.360
	So in 1975, I think you've created a program called Lex, which I especially like because

	03:50.360 --> 03:51.560
	my name is Lex.

	03:51.560 --> 03:52.560
	So thank you.

	03:52.560 --> 03:58.240
	Thank you for creating a brand that established a reputation that's long lasting reliable

	03:58.240 --> 04:01.240
	and has a big impact on the world and still used today.

	04:01.240 --> 04:03.000
	So thank you for that.

	04:03.000 --> 04:11.880
	But more seriously, in that time, in the 70s, as an engineer, personal computers were being

	04:11.880 --> 04:12.880
	born.

	04:12.880 --> 04:17.800
	Do you think you'd be able to predict the 80s, 90s, and the aughts of where computers

	04:17.800 --> 04:18.800
	would go?

	04:18.800 --> 04:23.160
	I'm sure I could not and would not have gotten it right.

	04:23.160 --> 04:27.960
	I was the beneficiary of the great work of many, many people who saw it clearer than I

	04:27.960 --> 04:29.160
	did.

	04:29.160 --> 04:34.760
	With Lex, I worked with a fellow named Michael Lesk, who was my supervisor, and he essentially

	04:34.760 --> 04:39.400
	helped me architect and deliver a system that's still in use today.

	04:39.400 --> 04:43.800
	After that, I worked at Xerox Palo Alto Research Center, where the Alto was invented, and the

	04:43.800 --> 04:50.600
	Alto is the predecessor of the modern personal computer, or Macintosh, and so forth.

	04:50.600 --> 04:55.360
	And the altos were very rare, and I had to drive an hour from Berkeley to go use them,

	04:55.360 --> 05:01.160
	but I made a point of skipping classes and doing whatever it took to have access to this

	05:01.160 --> 05:02.480
	extraordinary achievement.

	05:02.480 --> 05:05.000
	I knew that they were consequential.

	05:05.000 --> 05:08.240
	What I did not understand was scaling.

	05:08.240 --> 05:12.960
	I did not understand what would happen when you had 100 million as opposed to 100.

	05:12.960 --> 05:17.360
	And so since then, and I have learned the benefit of scale, I always look for things

	05:17.360 --> 05:23.160
	which are going to scale to platforms, so mobile phones, Android, all those things.

	05:23.160 --> 05:27.560
	There are many, many people in the world.

	05:27.560 --> 05:28.560
	People really have needs.

	05:28.560 --> 05:32.600
	They really will use these platforms, and you can build big businesses on top of them.

	05:32.600 --> 05:33.600
	So it's interesting.

	05:33.600 --> 05:36.880
	So when you see a piece of technology, now you think, what will this technology look

	05:36.880 --> 05:39.080
	like when it's in the hands of a billion people?

	05:39.080 --> 05:40.160
	That's right.

	05:40.160 --> 05:46.520
	So an example would be that the market is so competitive now that if you can't figure

	05:46.520 --> 05:51.880
	out a way for something to have a million users or a billion users, it probably is not

	05:51.880 --> 05:57.280
	going to be successful because something else will become the general platform, and your

	05:57.280 --> 06:04.360
	idea will become a lost idea or a specialized service with relatively few users.

	06:04.360 --> 06:06.000
	So it's a path to generality.

	06:06.000 --> 06:07.720
	It's a path to general platform use.

	06:07.720 --> 06:09.400
	It's a path to broad applicability.

	06:09.400 --> 06:15.000
	Now, there are plenty of good businesses that are tiny, so luxury goods, for example.

	06:15.000 --> 06:20.360
	But if you want to have an impact at scale, you have to look for things which are of

	06:20.360 --> 06:25.200
	common value, common pricing, common distribution, and solve common problems, the problems that

	06:25.200 --> 06:26.200
	everyone has.

	06:26.200 --> 06:30.440
	And by the way, people have lots of problems, information, medicine, health, education,

	06:30.440 --> 06:31.440
	and so forth.

	06:31.440 --> 06:32.440
	Work on those problems.

	06:32.440 --> 06:40.240
	Like you said, you're a big fan of the middle class because there's so many of them by definition.

	06:40.240 --> 06:46.600
	So any product, any thing that has a huge impact, it improves their lives is a great

	06:46.600 --> 06:48.960
	business decision and it's just good for society.

	06:48.960 --> 06:53.520
	And there's nothing wrong with starting off in the high end as long as you have a plan

	06:53.520 --> 06:55.520
	to get to the middle class.

	06:55.520 --> 06:59.280
	There's nothing wrong with starting with a specialized market in order to learn and to

	06:59.280 --> 07:01.080
	build and to fund things.

	07:01.080 --> 07:04.520
	So you start a luxury market to build a general purpose market.

	07:04.520 --> 07:09.640
	But if you define yourself as only a narrow market, someone else can come along with a

	07:09.640 --> 07:14.320
	general purpose market that can push you to the corner, can restrict the scale of operation,

	07:14.320 --> 07:17.320
	can force you to be a lesser impact than you might be.

	07:17.320 --> 07:22.800
	So it's very important to think in terms of broad businesses and broad impact, even if

	07:22.800 --> 07:26.360
	you start in a little corner somewhere.

	07:26.360 --> 07:33.200
	So as you look to the 70s, but also in the decades to come, and you saw computers, did

	07:33.200 --> 07:40.240
	you see them as tools or was there a little element of another entity?

	07:40.240 --> 07:46.240
	I remember a quote saying AI began with our dream to create the gods.

	07:46.240 --> 07:51.520
	Is there a feeling when you wrote that program that you were creating another entity, giving

	07:51.520 --> 07:52.800
	life to something?

	07:52.800 --> 07:58.880
	I wish I could say otherwise, but I simply found the technology platforms so exciting.

	07:58.880 --> 08:00.400
	That's what I was focused on.

	08:00.400 --> 08:04.640
	I think the majority of the people that I've worked with, and there are a few exceptions,

	08:04.640 --> 08:09.960
	Steve Jobs being an example, really saw this as a great technological play.

	08:09.960 --> 08:15.520
	I think relatively few of the technical people understood the scale of its impact.

	08:15.520 --> 08:19.680
	So I used NCP, which is a predecessor to TCPIP.

	08:19.680 --> 08:21.240
	It just made sense to connect things.

	08:21.240 --> 08:26.240
	We didn't think of it in terms of the internet, and then companies, and then Facebook, and

	08:26.240 --> 08:29.200
	then Twitter, and then politics, and so forth.

	08:29.200 --> 08:30.800
	We never did that build.

	08:30.800 --> 08:32.920
	We didn't have that vision.

	08:32.920 --> 08:38.200
	And I think most people, it's a rare person who can see compounding at scale.

	08:38.200 --> 08:41.520
	Most people can see, if you ask people to predict the future, they'll say, they'll give

	08:41.520 --> 08:44.080
	you an answer of six to nine months or 12 months.

	08:44.080 --> 08:47.560
	Because that's about as far as people can imagine.

	08:47.560 --> 08:51.020
	But there's an old saying, which actually was attributed to a professor at MIT a long

	08:51.020 --> 08:58.120
	time ago, that we overestimate what can be done in one year, and we underestimate what

	08:58.120 --> 09:00.280
	can be done in a decade.

	09:00.280 --> 09:05.560
	And there's a great deal of evidence that these core platforms at hardware and software

	09:05.560 --> 09:07.800
	take a decade.

	09:07.800 --> 09:09.600
	So think about self driving cars.

	09:09.600 --> 09:12.160
	Self driving cars were thought about in the 90s.

	09:12.160 --> 09:17.160
	Over projects around them, the first DARPA Durand Challenge was roughly 2004.

	09:17.160 --> 09:19.760
	So that's roughly 15 years ago.

	09:19.760 --> 09:25.400
	And today we have self driving cars operating in a city in Arizona, right, so 15 years.

	09:25.400 --> 09:31.720
	And we still have a ways to go before they're more generally available.

	09:31.720 --> 09:33.840
	So you've spoken about the importance.

	09:33.840 --> 09:37.080
	You just talked about predicting into the future.

	09:37.080 --> 09:41.640
	You've spoken about the importance of thinking five years ahead and having a plan for those

	09:41.640 --> 09:42.640
	five years.

	09:42.640 --> 09:47.840
	And the way to say it is that almost everybody has a one year plan.

	09:47.840 --> 09:50.960
	Almost no one has a proper five year plan.

	09:50.960 --> 09:55.160
	And the key thing to having a five year plan is having a model for what's going to happen

	09:55.160 --> 09:57.000
	under the underlying platforms.

	09:57.000 --> 09:59.840
	So here's an example.

	09:59.840 --> 10:05.120
	Computer Moore's Law, as we know it, the thing that powered improvements in CPUs has largely

	10:05.120 --> 10:10.400
	halted in its traditional shrinking mechanism, because the costs have just gotten so high.

	10:10.400 --> 10:12.200
	It's getting harder and harder.

	10:12.200 --> 10:16.600
	But there's plenty of algorithmic improvements and specialized hardware improvements.

	10:16.600 --> 10:21.240
	So you need to understand the nature of those improvements and where they'll go in order

	10:21.240 --> 10:24.360
	to understand how it will change the platform.

	10:24.360 --> 10:28.000
	In the area of network connectivity, what are the gains that are going to be possible

	10:28.000 --> 10:29.480
	in wireless?

	10:29.480 --> 10:35.720
	It looks like there's an enormous expansion of wireless connectivity at many different

	10:35.720 --> 10:36.960
	bands, right?

	10:36.960 --> 10:40.520
	And that we will primarily, historically, I've always thought that we were primarily

	10:40.520 --> 10:45.040
	going to be using fiber, but now it looks like we're going to be using fiber plus very

	10:45.040 --> 10:51.560
	powerful high bandwidth sort of short distance connectivity to bridge the last mile, right?

	10:51.560 --> 10:53.100
	That's an amazing achievement.

	10:53.100 --> 10:56.880
	If you know that, then you're going to build your systems differently.

	10:56.880 --> 10:59.800
	By the way, those networks have different latency properties, right?

	10:59.800 --> 11:05.040
	Because they're more symmetric, the algorithms feel faster for that reason.

	11:05.040 --> 11:09.920
	And so when you think about whether it's a fiber or just technologies in general.

	11:09.920 --> 11:15.920
	So there's this barber, wooden poem or quote that I really like.

	11:15.920 --> 11:20.400
	It's from the champions of the impossible rather than the slaves of the possible that

	11:20.400 --> 11:23.280
	evolution draws its creative force.

	11:23.280 --> 11:27.840
	So in predicting the next five years, I'd like to talk about the impossible and the

	11:27.840 --> 11:28.840
	possible.

	11:28.840 --> 11:34.720
	Well, and again, one of the great things about humanity is that we produce dreamers, right?

	11:34.720 --> 11:37.760
	We literally have people who have a vision and a dream.

	11:37.760 --> 11:43.400
	They are, if you will, disagreeable in the sense that they disagree with the, they disagree

	11:43.400 --> 11:46.240
	with what the sort of zeitgeist is.

	11:46.240 --> 11:48.040
	They say there is another way.

	11:48.040 --> 11:49.040
	They have a belief.

	11:49.040 --> 11:50.320
	They have a vision.

	11:50.320 --> 11:56.560
	If you look at science, science is always marked by such people who, who went against

	11:56.560 --> 12:01.360
	some conventional wisdom, collected the knowledge at the time and assembled it in a way that

	12:01.360 --> 12:03.760
	produced a powerful platform.

	12:03.760 --> 12:11.120
	And you've been amazingly honest about in an inspiring way about things you've been wrong

	12:11.120 --> 12:14.800
	about predicting and you've obviously been right about a lot of things.

	12:14.800 --> 12:23.880
	But in this kind of tension, how do you balance as a company in predicting the next five years,

	12:23.880 --> 12:26.520
	the impossible, planning for the impossible.

	12:26.520 --> 12:32.720
	So listening to those crazy dreamers, letting them do, letting them run away and make the

	12:32.720 --> 12:38.760
	impossible real, make it happen and slow, you know, that's how programmers often think

	12:38.760 --> 12:44.800
	and slowing things down and saying, well, this is the rational, this is the possible,

	12:44.800 --> 12:49.160
	the pragmatic, the, the dreamer versus the pragmatist.

	12:49.160 --> 12:56.680
	So it's helpful to have a model which encourages a predictable revenue stream as well as the

	12:56.680 --> 12:58.720
	ability to do new things.

	12:58.720 --> 13:03.120
	So in Google's case, we're big enough and well enough managed and so forth that we have

	13:03.120 --> 13:06.600
	a pretty good sense of what our revenue will be for the next year or two, at least for

	13:06.600 --> 13:07.960
	a while.

	13:07.960 --> 13:13.720
	And so we have enough cash generation that we can make bets.

	13:13.720 --> 13:18.760
	And indeed, Google has become alphabet so the corporation is organized around these

	13:18.760 --> 13:19.760
	bets.

	13:19.760 --> 13:25.560
	And these bets are in areas of fundamental importance to, to the world, whether it's

	13:25.560 --> 13:31.920
	digital intelligence, medical technology, self driving cars, connectivity through balloons,

	13:31.920 --> 13:33.440
	on and on and on.

	13:33.440 --> 13:36.080
	And there's more coming and more coming.

	13:36.080 --> 13:41.480
	So one way you could express this is that the current business is successful enough

	13:41.480 --> 13:44.720
	that we have the luxury of making bets.

	13:44.720 --> 13:48.960
	And another one that you could say is that we have the, the wisdom of being able to see

	13:48.960 --> 13:53.920
	that a corporate structure needs to be created to enhance the likelihood of the success of

	13:53.920 --> 13:55.320
	those bets.

	13:55.320 --> 13:59.760
	So we essentially turned ourselves into a conglomerate of bets.

	13:59.760 --> 14:04.360
	And then this underlying corporation Google, which is itself innovative.

	14:04.360 --> 14:08.160
	So in order to pull this off, you have to have a bunch of belief systems.

	14:08.160 --> 14:12.080
	And one of them is that you have to have bottoms up and tops down the bottoms up.

	14:12.080 --> 14:13.600
	We call 20% time.

	14:13.600 --> 14:17.040
	And the idea is that people can spend 20% of the time on whatever they want.

	14:17.040 --> 14:21.960
	And the top down is that our founders in particular have a keen eye on technology and they're

	14:21.960 --> 14:24.000
	reviewing things constantly.

	14:24.000 --> 14:27.520
	So an example would be they'll, they'll hear about an idea or I'll hear about something

	14:27.520 --> 14:28.880
	and it sounds interesting.

	14:28.880 --> 14:34.920
	Let's go visit them and then let's begin to assemble the pieces to see if that's possible.

	14:34.920 --> 14:39.920
	And if you do this long enough, you get pretty good at predicting what's likely to work.

	14:39.920 --> 14:42.120
	So that's, that's a beautiful balance that struck.

	14:42.120 --> 14:44.560
	Is this something that applies at all scale?

	14:44.560 --> 14:53.960
	So seems seems to be that the Sergei, again, 15 years ago, came up with a concept that

	14:53.960 --> 14:59.040
	called 10% of the budget should be on things that are unrelated.

	14:59.040 --> 15:05.040
	It was called 70, 20, 10, 70% of our time on core business, 20% on adjacent business

	15:05.040 --> 15:06.920
	and 10% on other.

	15:06.920 --> 15:11.200
	And he proved mathematically, of course, he's a brilliant mathematician, that you needed

	15:11.200 --> 15:18.800
	that 10% right to make the sum of the growth work and it turns out he was right.

	15:18.800 --> 15:24.320
	So getting into the world of artificial intelligence, you've, you've talked quite extensively and

	15:24.320 --> 15:32.160
	effectively to the impact in the near term, the positive impact of artificial intelligence,

	15:32.160 --> 15:38.720
	whether it's machine, especially machine learning in medical applications and education

	15:38.720 --> 15:44.160
	and just making information more accessible, right in the AI community, there is a kind

	15:44.160 --> 15:49.520
	of debate, so there's this shroud of uncertainty as we face this new world with artificial

	15:49.520 --> 15:50.800
	intelligence in it.

	15:50.800 --> 15:57.560
	And there is some people like Elon Musk, you've disagreed on at least on the degree of emphasis

	15:57.560 --> 16:00.800
	he places on the existential threat of AI.

	16:00.800 --> 16:07.120
	So I've spoken with Stuart Russell, Max Tagmark, who share Elon Musk's view, and Yoshio Benjio,

	16:07.120 --> 16:09.240
	Steven Pinker, who do not.

	16:09.240 --> 16:13.320
	And so there's a, there's a, there's a lot of very smart people who are thinking about

	16:13.320 --> 16:17.280
	this stuff, disagreeing, which is really healthy, of course.

	16:17.280 --> 16:22.800
	So what do you think is the healthiest way for the AI community to, and really for the

	16:22.800 --> 16:30.880
	general public to think about AI and the concern of the technology being mismanaged

	16:30.880 --> 16:33.000
	in some, in some kind of way.

	16:33.000 --> 16:37.560
	So the source of education for the general public has been robot killer movies.

	16:37.560 --> 16:38.760
	Right.

	16:38.760 --> 16:44.840
	And Terminator, et cetera, and the one thing I can assure you we're not building are those

	16:44.840 --> 16:46.200
	kinds of solutions.

	16:46.200 --> 16:51.240
	Furthermore, if they were to show up, someone would notice and unplug them, right?

	16:51.240 --> 16:57.760
	So as exciting as those movies are, and they're great movies, where the killer robots to start,

	16:57.760 --> 17:00.520
	we would find a way to stop them, right?

	17:00.520 --> 17:04.320
	So I'm not concerned about that.

	17:04.320 --> 17:08.680
	And much of this has to do with the timeframe of conversation.

	17:08.680 --> 17:16.120
	So you can imagine a situation a hundred years from now, when the human brain is fully understood,

	17:16.120 --> 17:19.880
	and the next generation and next generation of brilliant MIT scientists have figured

	17:19.880 --> 17:25.960
	all this out, we're going to have a large number of ethics questions, right, around science

	17:25.960 --> 17:29.760
	and thinking and robots and computers and so forth and so on.

	17:29.760 --> 17:32.360
	So it depends on the question of the timeframe.

	17:32.360 --> 17:37.280
	In the next five to 10 years, we're not facing those questions.

	17:37.280 --> 17:42.000
	What we're facing in the next five to 10 years is how do we spread this disruptive technology

	17:42.000 --> 17:46.520
	as broadly as possible to gain the maximum benefit of it?

	17:46.520 --> 17:51.880
	The primary benefit should be in healthcare and in education, healthcare because it's

	17:51.880 --> 17:52.880
	obvious.

	17:52.880 --> 17:55.840
	We're all the same, even though we somehow believe we're not.

	17:55.840 --> 18:00.400
	As a medical matter, the fact that we have big data about our health will save lives,

	18:00.400 --> 18:05.520
	allow us to deal with skin cancer and other cancers, ophthalmological problems.

	18:05.520 --> 18:10.080
	There's people working on psychological diseases and so forth using these techniques.

	18:10.080 --> 18:11.680
	I go on and on.

	18:11.680 --> 18:15.840
	The promise of AI in medicine is extraordinary.

	18:15.840 --> 18:20.360
	There are many, many companies and startups and funds and solutions and we will all live

	18:20.360 --> 18:22.120
	much better for that.

	18:22.120 --> 18:25.680
	The same argument in education.

	18:25.680 --> 18:31.760
	Can you imagine that for each generation of child and even adult, you have a tutor educator

	18:31.760 --> 18:37.320
	that's AI based, that's not a human but is properly trained, that helps you get smarter,

	18:37.320 --> 18:41.440
	helps you address your language difficulties or your math difficulties or what have you.

	18:41.440 --> 18:43.400
	Why don't we focus on those two?

	18:43.400 --> 18:49.240
	The gains societally of making humans smarter and healthier are enormous.

	18:49.240 --> 18:54.000
	Those translate for decades and decades and will all benefit from them.

	18:54.000 --> 18:58.560
	There are people who are working on AI safety, which is the issue that you're describing.

	18:58.560 --> 19:02.920
	There are conversations in the community that should there be such problems, what should

	19:02.920 --> 19:04.360
	the rules be like?

	19:04.360 --> 19:09.360
	Google, for example, has announced its policies with respect to AI safety, which I certainly

	19:09.360 --> 19:14.320
	support and I think most everybody would support and they make sense.

	19:14.320 --> 19:19.760
	It helps guide the research but the killer robots are not arriving this year and they're

	19:19.760 --> 19:23.840
	not even being built.

	19:23.840 --> 19:31.040
	On that line of thinking, you said the timescale, in this topic or other topics, have you found

	19:31.040 --> 19:37.920
	a useful, on the business side or the intellectual side, to think beyond 5, 10 years, to think

	19:37.920 --> 19:39.480
	50 years out?

	19:39.480 --> 19:42.000
	Has it ever been useful or productive?

	19:42.000 --> 19:49.040
	In our industry, there are essentially no examples of 50 year predictions that have been correct.

	19:49.040 --> 19:54.320
	Let's review AI, which was largely invented here at MIT and a couple of other universities

	19:54.320 --> 19:57.840
	in 1956, 1957, 1958.

	19:57.840 --> 20:01.800
	The original claims were a decade or two.

	20:01.800 --> 20:08.040
	When I was a PhD student, I studied AI a bit and it entered during my looking at it a period

	20:08.040 --> 20:13.880
	which is known as AI winter, which went on for about 30 years, which is a whole generation

	20:13.880 --> 20:18.800
	of scientists and a whole group of people who didn't make a lot of progress because the

	20:18.800 --> 20:22.160
	algorithms had not improved and the computers had not improved.

	20:22.160 --> 20:26.160
	It took some brilliant mathematicians, starting with a fellow named Jeff Hinton at Toronto

	20:26.160 --> 20:33.120
	in Montreal, who basically invented this deep learning model which empowers us today.

	20:33.120 --> 20:40.400
	The seminal work there was 20 years ago and in the last 10 years, it's become popularized.

	20:40.400 --> 20:43.520
	Think about the time frames for that level of discovery.

	20:43.520 --> 20:46.080
	It's very hard to predict.

	20:46.080 --> 20:50.240
	Many people think that we'll be flying around in the equivalent of flying cars.

	20:50.240 --> 20:51.240
	Who knows?

	20:51.240 --> 20:56.680
	My own view, if I want to go out on a limb, is to say that we know a couple of things

	20:56.680 --> 20:58.000
	about 50 years from now.

	20:58.000 --> 21:00.480
	We know that there'll be more people alive.

	21:00.480 --> 21:04.000
	We know that we'll have to have platforms that are more sustainable because the earth

	21:04.000 --> 21:09.440
	is limited in the ways we all know and that the kind of platforms that are going to get

	21:09.440 --> 21:13.000
	billed will be consistent with the principles that I've described.

	21:13.000 --> 21:17.560
	They will be much more empowering of individuals, they'll be much more sensitive to the ecology

	21:17.560 --> 21:20.440
	because they have to be, they just have to be.

	21:20.440 --> 21:24.160
	I also think that humans are going to be a great deal smarter and I think they're going

	21:24.160 --> 21:28.320
	to be a lot smarter because of the tools that I've discussed with you and of course people

	21:28.320 --> 21:29.320
	will live longer.

	21:29.320 --> 21:32.080
	Life extension is continuing apace.

	21:32.080 --> 21:36.840
	A baby born today has a reasonable chance of living to 100, which is pretty exciting.

	21:36.840 --> 21:40.760
	It's well past the 21st century, so we better take care of them.

	21:40.760 --> 21:46.160
	You mentioned interesting statistic on some very large percentage, 60, 70% of people may

	21:46.160 --> 21:48.360
	live in cities.

	21:48.360 --> 21:53.880
	Today more than half the world lives in cities and one of the great stories of humanity in

	21:53.880 --> 21:57.560
	the last 20 years has been the rural to urban migration.

	21:57.560 --> 22:02.720
	This has occurred in the United States, it's occurred in Europe, it's occurring in Asia

	22:02.720 --> 22:04.760
	and it's occurring in Africa.

	22:04.760 --> 22:09.280
	When people move to cities, the cities get more crowded, but believe it or not their health

	22:09.280 --> 22:15.560
	gets better, their productivity gets better, their IQ and educational capabilities improve,

	22:15.560 --> 22:19.840
	so it's good news that people are moving to cities, but we have to make them livable

	22:19.840 --> 22:22.800
	and safe.

	22:22.800 --> 22:28.240
	So you, first of all, you are, but you've also worked with some of the greatest leaders

	22:28.240 --> 22:30.180
	in the history of tech.

	22:30.180 --> 22:37.080
	What insights do you draw from the difference in leadership styles of yourself, Steve Jobs,

	22:37.080 --> 22:45.320
	Elon Musk, Larry Page, now the new CEO, Sandra Pichai and others from the, I would say, calm

	22:45.320 --> 22:49.600
	sages to the mad geniuses?

	22:49.600 --> 22:53.880
	One of the things that I learned as a young executive is that there's no single formula

	22:53.880 --> 22:56.200
	for leadership.

	22:56.200 --> 23:00.080
	They try to teach one, but that's not how it really works.

	23:00.080 --> 23:04.360
	There are people who just understand what they need to do and they need to do it quickly.

	23:04.360 --> 23:06.800
	Those people are often entrepreneurs.

	23:06.800 --> 23:09.080
	They just know and they move fast.

	23:09.080 --> 23:13.400
	There are other people who are systems thinkers and planners, that's more who I am, somewhat

	23:13.400 --> 23:18.760
	more conservative, more thorough in execution, a little bit more risk averse.

	23:18.760 --> 23:24.120
	There's also people who are sort of slightly insane, right, in the sense that they are

	23:24.120 --> 23:29.040
	emphatic and charismatic and they feel it and they drive it and so forth.

	23:29.040 --> 23:31.440
	There's no single formula to success.

	23:31.440 --> 23:35.320
	There is one thing that unifies all of the people that you named, which is very high

	23:35.320 --> 23:41.240
	intelligence, at the end of the day, the thing that characterizes all of them is that they

	23:41.240 --> 23:46.360
	saw the world quicker, faster, they processed information faster, they didn't necessarily

	23:46.360 --> 23:50.160
	make the right decisions all the time, but they were on top of it.

	23:50.160 --> 23:54.600
	The other thing that's interesting about all those people is they all started young.

	23:54.600 --> 23:58.560
	Think about Steve Jobs starting Apple roughly at 18 or 19.

	23:58.560 --> 24:01.840
	Think about Bill Gates starting at roughly 2021.

	24:01.840 --> 24:07.040
	Think about by the time they were 30, Mark Zuckerberg, a more good example at 1920.

	24:07.040 --> 24:13.720
	By the time they were 30, they had 10 years, at 30 years old, they had 10 years of experience

	24:13.720 --> 24:19.920
	of dealing with people and products and shipments and the press and business and so forth.

	24:19.920 --> 24:24.480
	It's incredible how much experience they had compared to the rest of us who were busy getting

	24:24.480 --> 24:25.480
	our PhDs.

	24:25.480 --> 24:26.480
	Yes, exactly.

	24:26.480 --> 24:32.760
	We should celebrate these people because they've just had more life experience and that helps

	24:32.760 --> 24:34.520
	inform the judgment.

	24:34.520 --> 24:41.360
	At the end of the day, when you're at the top of these organizations, all the easy questions

	24:41.360 --> 24:43.680
	have been dealt with.

	24:43.680 --> 24:45.840
	How should we design the buildings?

	24:45.840 --> 24:48.400
	Where should we put the colors on our product?

	24:48.400 --> 24:51.440
	What should the box look like?

	24:51.440 --> 24:55.520
	The problems, that's why it's so interesting to be in these rooms, the problems that they

	24:55.520 --> 25:00.200
	face in terms of the way they operate, the way they deal with their employees, their

	25:00.200 --> 25:04.160
	customers, their innovation are profoundly challenging.

	25:04.160 --> 25:09.360
	Each of the companies is demonstrably different culturally.

	25:09.360 --> 25:11.800
	They are not, in fact, cut of the same.

	25:11.800 --> 25:16.680
	They behave differently based on input, their internal cultures are different, their compensation

	25:16.680 --> 25:24.920
	schemes are different, their values are different, so there's proof that diversity works.

	25:24.920 --> 25:33.440
	So when faced with a tough decision, in need of advice, it's been said that the best thing

	25:33.440 --> 25:39.840
	one can do is to find the best person in the world who can give that advice and find a

	25:39.840 --> 25:44.880
	way to be in a room with them, one on one and ask.

	25:44.880 --> 25:51.920
	So here we are, and let me ask in a long winded way, I wrote this down, in 1998 there were

	25:51.920 --> 26:01.960
	many good search engines, Lycos, Excite, Altavista, Infoseek, AskJeeves maybe, Yahoo even.

	26:01.960 --> 26:07.040
	So Google stepped in and disrupted everything, they disrupted the nature of search, the nature

	26:07.040 --> 26:12.040
	of our access to information, the way we discover new knowledge.

	26:12.040 --> 26:19.120
	So now it's 2018, actually 20 years later, there are many good personal AI assistants,

	26:19.120 --> 26:22.360
	including of course the best from Google.

	26:22.360 --> 26:28.720
	So you've spoken in medical and education, the impact of such an AI assistant could bring.

	26:28.720 --> 26:34.920
	So we arrive at this question, so it's a personal one for me, but I hope my situation represents

	26:34.920 --> 26:41.200
	that of many other, as we said, dreamers and the crazy engineers.

	26:41.200 --> 26:46.680
	So my whole life, I've dreamed of creating such an AI assistant.

	26:46.680 --> 26:50.800
	Every step I've taken has been towards that goal, now I'm a research scientist in Human

	26:50.800 --> 26:58.920
	Senate AI here at MIT, so the next step for me as I sit here, facing my passion, is to

	26:58.920 --> 27:04.880
	do what Larry and Sergey did in 1998, this simple start up.

	27:04.880 --> 27:10.640
	And so here's my simple question, given the low odds of success, the timing and luck required,

	27:10.640 --> 27:14.280
	the countless other factors that can't be controlled or predicted, which is all the

	27:14.280 --> 27:16.560
	things that Larry and Sergey faced.

	27:16.560 --> 27:23.080
	Is there some calculation, some strategy to follow in this step, or do you simply follow

	27:23.080 --> 27:26.560
	the passion just because there's no other choice?

	27:26.560 --> 27:32.880
	I think the people who are in universities are always trying to study the extraordinarily

	27:32.880 --> 27:37.360
	chaotic nature of innovation and entrepreneurship.

	27:37.360 --> 27:42.880
	My answer is that they didn't have that conversation, they just did it.

	27:42.880 --> 27:48.840
	They sensed a moment when, in the case of Google, there was all of this data that needed

	27:48.840 --> 27:53.940
	to be organized and they had a better algorithm, they had invented a better way.

	27:53.940 --> 28:01.040
	So today with Human Senate AI, which is your area of research, there must be new approaches.

	28:01.040 --> 28:07.320
	It's such a big field, there must be new approaches, different from what we and others are doing.

	28:07.320 --> 28:12.320
	There must be startups to fund, there must be research projects to try, there must be

	28:12.320 --> 28:15.200
	graduate students to work on new approaches.

	28:15.200 --> 28:19.120
	Here at MIT, there are people who are looking at learning from the standpoint of looking

	28:19.120 --> 28:23.840
	at child learning, how do children learn starting at each one?

	28:23.840 --> 28:25.560
	And the work is fantastic.

	28:25.560 --> 28:30.120
	Those approaches are different from the approach that most people are taking.

	28:30.120 --> 28:33.980
	Perhaps that's a bet that you should make, or perhaps there's another one.

	28:33.980 --> 28:40.200
	But at the end of the day, the successful entrepreneurs are not as crazy as they sound.

	28:40.200 --> 28:43.200
	They see an opportunity based on what's happened.

	28:43.200 --> 28:45.400
	Let's use Uber as an example.

	28:45.400 --> 28:49.840
	As Travis sells the story, he and his cofounder were sitting in Paris and they had this idea

	28:49.840 --> 28:52.160
	because they couldn't get a cab.

	28:52.160 --> 28:56.800
	And they said, we have smartphones and the rest is history.

	28:56.800 --> 29:04.040
	So what's the equivalent of that Travis Eiffel Tower, where is a cab moment that you could,

	29:04.040 --> 29:08.800
	as an entrepreneur, take advantage of, whether it's in Human Senate AI or something else?

	29:08.800 --> 29:11.480
	That's the next great start up.

	29:11.480 --> 29:13.760
	And the psychology of that moment.

	29:13.760 --> 29:20.120
	So when Sergey and Larry talk about, and listen to a few interviews, it's very nonchalant.

	29:20.120 --> 29:25.280
	Well, here's the very fascinating web data.

	29:25.280 --> 29:29.080
	And here's an algorithm we have for, you know, we just kind of want to play around with that

	29:29.080 --> 29:30.080
	data.

	29:30.080 --> 29:32.520
	And it seems like that's a really nice way to organize this data.

	29:32.520 --> 29:38.000
	Well, I should say what happened, remember, is that they were graduate students at Stanford

	29:38.000 --> 29:41.320
	and they thought this is interesting, so they built a search engine and they kept it in

	29:41.320 --> 29:43.400
	their room.

	29:43.400 --> 29:47.520
	And they had to get power from the room next door because they were using too much power

	29:47.520 --> 29:48.520
	in the room.

	29:48.520 --> 29:51.640
	So they ran an extension cord over.

	29:51.640 --> 29:55.360
	And then they went and they found a house and they had Google World headquarters of

	29:55.360 --> 29:57.600
	five people to start the company.

	29:57.600 --> 30:02.560
	And they raised $100,000 from Andy Bechtelstein, who was the sun founder to do this, and Dave

	30:02.560 --> 30:04.520
	Chariton and a few others.

	30:04.520 --> 30:11.960
	The point is their beginnings were very simple, but they were based on a powerful insight.

	30:11.960 --> 30:14.320
	That is a replicable model for any startup.

	30:14.320 --> 30:16.520
	It has to be a powerful insight.

	30:16.520 --> 30:17.680
	The beginnings are simple.

	30:17.680 --> 30:19.960
	And there has to be an innovation.

	30:19.960 --> 30:24.280
	In Larry and Sergey's case, it was PageRank, which was a brilliant idea, one of the most

	30:24.280 --> 30:26.880
	cited papers in the world today.

	30:26.880 --> 30:29.880
	What's the next one?

	30:29.880 --> 30:37.280
	So you're one of, if I may say, richest people in the world, and yet it seems that money

	30:37.280 --> 30:43.200
	is simply a side effect of your passions and not an inherent goal.

	30:43.200 --> 30:48.360
	But it's a, you're a fascinating person to ask.

	30:48.360 --> 30:55.080
	So much of our society at the individual level and at the company level and its nations is

	30:55.080 --> 30:58.920
	driven by the desire for wealth.

	30:58.920 --> 31:01.280
	What do you think about this drive?

	31:01.280 --> 31:07.000
	And what have you learned about, if I may romanticize the notion, the meaning of life,

	31:07.000 --> 31:10.520
	having achieved success on so many dimensions?

	31:10.520 --> 31:16.960
	There have been many studies of human happiness and above some threshold, which is typically

	31:16.960 --> 31:23.600
	relatively low for this conversation, there's no difference in happiness about money.

	31:23.600 --> 31:30.120
	The happiness is correlated with meaning and purpose, a sense of family, a sense of impact.

	31:30.120 --> 31:34.440
	So if you organize your life, assuming you have enough to get around and have a nice

	31:34.440 --> 31:40.400
	home and so forth, you'll be far happier if you figure out what you care about and work

	31:40.400 --> 31:41.800
	on that.

	31:41.800 --> 31:44.120
	It's often being in service to others.

	31:44.120 --> 31:47.840
	It's a great deal of evidence that people are happiest when they're serving others

	31:47.840 --> 31:49.640
	and not themselves.

	31:49.640 --> 31:57.480
	This goes directly against the sort of press induced excitement about powerful and wealthy

	31:57.480 --> 32:01.840
	leaders of one kind and indeed, these are consequential people.

	32:01.840 --> 32:06.720
	But if you are in a situation where you've been very fortunate as I have, you also have

	32:06.720 --> 32:12.160
	to take that as a responsibility and you have to basically work both to educate others and

	32:12.160 --> 32:16.760
	give them that opportunity, but also use that wealth to advance human society.

	32:16.760 --> 32:20.440
	In my case, I'm particularly interested in using the tools of artificial intelligence

	32:20.440 --> 32:22.800
	and machine learning to make society better.

	32:22.800 --> 32:24.000
	I've mentioned education.

	32:24.000 --> 32:29.040
	I've mentioned inequality and middle class and things like this, all of which are a passion

	32:29.040 --> 32:30.160
	of mine.

	32:30.160 --> 32:31.920
	It doesn't matter what you do.

	32:31.920 --> 32:36.560
	It matters that you believe in it, that it's important to you and that your life will be

	32:36.560 --> 32:40.480
	far more satisfying if you spend your life doing that.

	32:40.480 --> 32:45.320
	I think there's no better place to end than a discussion of the meaning of life.

	32:45.320 --> 32:46.320
	Eric, thank you so much.

	32:46.320 --> 32:47.320
	Thank you very much.

	32:47.320 --> 33:16.320
	Thank you.