00:00hi everyone welcome to the a 6nc podcast
00:02I'm sonal so today we have a special
00:04episode we talk a lot about network
00:06effects is one of the most important
00:08dynamics especially in software based
00:09businesses you can see much of ours and
00:11others thinking on the topic at a 16
00:13c-calm slash network effects but today
00:15our special guest is w Brian Arthur he's
00:18widely credited for first describing
00:20network effects and beyond that has had
00:22a long and very influential career in
00:24economics especially as applied to the
00:26tech industry so I asked Marc Andreessen
00:27to co-host and add a little color
00:29commentary but first more about Brian
00:31Brian was formerly a professor of
00:33economics at Stanford is a visiting
00:35researcher at Parc formerly Xerox PARC
00:37and is also an external professor at the
00:39Santa Fe Institute because besides his
00:42foundational work in network effects
00:43he's also considered one of the fathers
00:44of complexity theory has written books
00:46on the nature of technology and how it
00:48evolves and has also written a number of
00:50pieces on AI and the autonomy economy
00:52all of which we'll touch upon in today's
00:54episode we also cover a lot of neat
00:56history in between and we end on the
00:58topic of innovation clusters around the
01:00world including Silicon Valley but first
01:02we begin briefly with where Brian's
01:04ideas came from you're a really
01:06influential economist whose and I can
01:09sometimes make fun of economic your work
01:17has really actually driven so much or
01:20described so much of what actually
01:22happens in technology and there seems to
01:23be a gap often between the worlds of
01:25economics and technology and you're
01:27really at the heart of that so why don't
01:29we start with some of your most seminal
01:30work starting with your famous classic
01:33paper around increasing returns and
01:34positive feedbacks if you were to just
01:36sort of distill and summarize some of
01:38the key concepts and how it contributed
01:40to the tech industry sure to go back a
01:42little bit I've been interested in
01:44technology for a long time I was trained
01:46as an engineer and then mathematician
01:49and operations research or basically
01:53so my basis is actually technology and
01:57then I added as a layer on top of that I
01:59fell into the wrong company and became
02:02an economist and I arrived in Stanford
02:06in 1982 at that time Silicon Valley was
02:10blossoming we said in 82 it was all
02:13Electronics then was but computation and
02:16the web then the cloud now it's about AI
02:20so Silicon Valley keeps morphing and
02:23changing I was enormous ly taken just by
02:27the sheer energy of the place in the
02:29early eighties and on through 26 years
02:33or so since it keeps recreating itself
02:37it's like looking into I don't know it's
02:39like looking into some cauldron of
02:41everything bubbling and changing all the
02:43time and it became very clear to me that
02:47there is a phenomenon going on in
02:49technology that you didn't see so much
02:52in the rest of the economy right the
02:54phenomenon of network effects which I
02:56should clarify in your papers is also
02:58named positive feedbacks and increasing
03:00returns yeah in standard economics if
03:04you get very large in the market
03:06everybody runs into some sort of
03:08diminishing returns and markets tend to
03:12balance the market fairly well a little
03:14shared that was the theory when I came
03:18along but it didn't seem to me the tech
03:21work that way go back to about 1980 to
03:2584 at the time we had VHS and we had
03:31beta those were the basic operating
03:35systems for video recorders and one of
03:38them happened be better Betamax was
03:41better and I started to wonder why VHS
03:43dominated the market I've always
03:45wondered this actually and then I
03:47realized that the hosts of small events
03:50early on had pushed VHS into a slight
03:54lead and if you were going down to your
03:57local movie rental store again this is
04:00back when blockbuster existed
04:02blockbuster you would tender will see
04:04more VHS movies that meant you've got a
04:07VHS recorder and that meant that they
04:10would stock more VHS aren't those
04:12compliments and economic terms
04:13oh yeah the two we're kind of
04:16interacting the more VHS is out there
04:18the more I buy the edges so I began to
04:21realize that I was seeing this in market
04:23after market there weren't diminishing
04:26turns if VHS got ahead it would further
04:28advantage the whole thing is quite
04:30unstable and small events tilted you
04:33towards beta or VHS my analogy was this
04:38was like bowling the ball perfectly done
04:40in the middle of an infinitely long
04:42bowling alley it could stay quite long
04:45in the middle but if it started to drift
04:47to one side it could further and then it
04:50would fall into the gutter at the side
04:53and that side would lock in the market
04:57so to speak and by the way you borrowed
04:59I think mmm you telling me that you
05:01borrow the lock in jargon from military
05:03like walking in on a target
05:04yeah the lock-in wasn't used heavily at
05:08the time I'm sure there are other people
05:10who who use the phrase but with fighter
05:13jet radar when you're going at very high
05:16speed and you're pursuing an enemy or
05:18something or maybe a radar station
05:21itself on the ground you lock into the
05:24target it's not just that you find the
05:26target but you want to lock on to that
05:28target and then you can release your
05:31weapons and the weapons will stay locked
05:33into that particular so I borrowed
05:38lock-in and since that's become very
05:41popular we were locked into this we're
05:43locked into that basically meaning that
05:45small chance events have landed you into
05:47something you can't get out of so what I
05:50realized were quite a few phenomena that
05:52have become famous since this was all
05:55very embryonic in my mind that there's
05:58sort of firms I was looking at one of
06:01them got a head out of half a dozen it
06:02could get further ahead you couldn't
06:05predict which one would get ahead it
06:07would start to get enough advantage that
06:10it could dominate the market and get
06:12still further ahead it would lock in it
06:15will have so much cost advantage or now
06:18we'd say it's so much user base that it
06:21would be hard to dislodge Microsoft got
06:24ahead with certain contracts very early
06:26in the game they locked in a lot of the
06:29personal software in the 1980s similarly
06:33other systems came along since
06:37there were search engines like Alta
06:39Vista as well as Google and others
06:41Google gets ahead and began to dominate
06:45that market and now has pretty well
06:47locked in you could say similar things
06:49for social media so it was a general
06:53phenomenon that anything that got ahead
06:56because you wanted to be with the
06:59majority of people could get further
07:01ahead we now call it network effects
07:04companies like that set up a network of
07:07users you want to be with the dominant
07:09network because your friends are with
07:11that or it's more valuable than where
07:12you yeah or they are you know more about
07:14it you hear more about it or you
07:16understand it better five generations
07:19ago none of our ancestors spoke English
07:22but we're all speaking English of the
07:25network effect I speak we speak English
07:27because we want to be understood by
07:29everybody small events that Gong
07:32otherwise in the seventeen hundreds it
07:34might have been French or if you were
07:36better than the fifteen hundred he could
07:38have been Latin or whatever so how is it
07:41received when you first put out this
07:43paper arguing against diminishing
07:44returns intact more towards positive
07:47feedbacks everything richer I wrote a
07:50paper on this in 1983 centered for
07:54leading economics journals not all the
07:57same time one after another I finally
08:00got it published six years later in 1989
08:03they didn't really they did not like it
08:05I kept getting reviewers saying we can't
08:08find fault in this but this isn't
08:09economic and in the meantime the idea
08:14was out there but there's no citation
08:18because no journal dared to publish this
08:21there's a good reason in those days what
08:24I was saying is that the economy could
08:26lock in two technologies or two products
08:31or even two ways of doing things that
08:34might be inferior because that came up
08:37maybe early on by chance and got locked
08:40in and during the Cold War in the mid
08:44eighties this was not popular I gave the
08:47talk in Moscow in 1980
08:50for I was saying in a capitalist economy
08:53you can lock in to an inferior product
08:56hands went up you know professor we want
09:00to point out that in Soviet Union such a
09:03thing not possible because with
09:06socialist planning we do not make such
09:09mistakes the central planners will
09:11dictate the outcome I came back to
09:13Stanford go to PhD student I said figure
09:16this out I don't believe it he did he
09:19wrote a beautiful paper Robin Collins's
09:22name and he showed that even with the
09:26you can't foresee what's gonna happen
09:29and of course you can lock into the
09:31wrong thing economists hated this the
09:34whole idea was everybody's free to
09:36choose and that lands you in the right
09:38solution and I thought is that correct
09:42I'm free to choose who we always choose
09:45the best spouses social statistics might
09:49suggest otherwise but what it made for
09:52his very different game and Silicon
09:53Valley so speaking of it being a
09:55different game you know we have a lot of
09:57entrepreneurs that listen to our podcast
09:58how does it change the game because
10:00people always use the phrase game
10:01changing very freely well first of all
10:04entrepreneurs in Silicon Valley are
10:06really smart and they didn't exactly get
10:11all these ideas from me I'm not being
10:13modest I'm just being realistic
10:14when I brought out this theory it kind
10:17of corroborated their intuition so what
10:21that says this if you are thinking in
10:24standard terms and go back to brewing
10:27beer or company like General Foods if
10:31you want to make profits you're thinking
10:34of getting production up and running
10:36properly getting your costs down making
10:39sure everything's terribly efficient the
10:41game was different in tech the whole
10:45game was to try to early on grab as much
10:48advantage as you could and I remember
10:51that I wrote a paper in this the Harvard
10:53Business Review in 1996 and as that
10:57paper gets circulated very widely in
11:00I remember hearing one story that sun
11:04microsystems had developed java and
11:07naturally that cost a huge amount of
11:09money so the guys with the green shades
11:13accountants were saying naturally enough
11:16we should charge huge amount of money
11:18for anybody who buys this and the other
11:22people had read this theory say no no
11:26give it away give it away give it away
11:28for free and there was a tremendous
11:30hullabaloo over this and finally
11:35somebody took my article and just
11:37slammed it and McNeeley desk and it was
11:44game over he got the point immediately
11:46that what you do in increasing returns
11:49market is you try to build up your user
11:50base now that's become completely
11:53intuitive since there was a time it
11:56wasn't standard that the accountants
11:58were saying we need to amortize all the
12:01R&D money and we need to get that I play
12:04back as fast as we can so we'll charge
12:07arms and nights later we can drop the
12:10calls it requires deferral of
12:13gratification right it requires
12:14long-term thinking that's right it
12:16requires no other words not only
12:17strategic thinking but also long-term
12:18thinking long-term thinking you have to
12:20project forward to what the economics
12:22will be when you win yeah and again I
12:24think that that makes very different
12:26atmosphere in tech tech is not about
12:29making profits it's about positioning
12:32yourself in markets and trying to build
12:35up user base or network advantages
12:38trying to build on those positive
12:40feedbacks think of amazon.com for years
12:44and years they kept reinvesting and kept
12:46betting on the positive feedbacks and
12:49eventually they dominated that whole
12:51market now they can make huge profits
12:54and keep expanding but it gives you a
12:57very different way of thinking I called
12:59the standard way of doing things the
13:01halls of production you know these big
13:03factories but it seemed to me that what
13:06was happening in tech was not the halls
13:09of production I call that the casino of
13:13as if you had this huge marquee there
13:16are many tables with different games
13:19oh yeah we're doing a game on face
13:21recognition over here whatever and
13:24people come up to the table and it just
13:27says search engines say okay who's gonna
13:30be here we don't know the technology
13:32hasn't really started what's the
13:34technology gonna be like I have no idea
13:36miss you how much do I have to put up
13:38front well you know you could join the
13:41game masseur for maybe one balloon what
13:43are my chances of winning I have no idea
13:46perhaps if there are ten players whoever
13:49chances might be one in ten do you still
13:51want to play so this is a very different
13:54game and I don't want to make it sound
13:58like too much luck because the
13:59particular entrepreneurs who kind of
14:02knowing that their technology was right
14:05and they had a sort of instinctive idea
14:07of positioning the technology and
14:09building that user base early rather
14:13than saying we want to get profits out
14:16of this the game keeps changing but my
14:19point is that the basic game in tech is
14:23not the same as the basic game in
14:25standard production and every once in a
14:28while you see somebody taken from the
14:30standard production side of the economy
14:32some CEO brought into a tech firm and
14:36they don't quite get it
14:37the classic case was Apple the classic
14:41case was Scully that famous quote do you
14:43want to sell sugar water for the rest of
14:45your life to him to be enticed away from
14:47a beverage company to work purple and
14:49CEOs are very smart indeed but it's not
14:53just a matter of intelligence it's
14:55different way of thinking and it's so
14:58familiar to us now this new way of
15:00thinking in the valley in Silicon Valley
15:03that we take it for granted that we
15:06always thought that way but we didn't do
15:10you think that if you have a view on
15:11this or now do you think financial
15:12markets understand this to the degree
15:14that they should even after all this
15:15time no I'll give you two instances
15:18Warren Buffett very very famously said I
15:22don't dabble in height
15:23it says I I don't touch that simply
15:26because I don't understand it another
15:28friend of mine bill Miller of flag Mason
15:31I've known him for 20 or more years
15:34through the Sante fans yeah bill read
15:36this stuff got it understood it and did
15:40so the best answer I can give to that is
15:43it's not general knowledge among
15:45investors fully yet certainly wasn't 20
15:49years ago but there's an increasing
15:51number of people who get that the rules
15:53of the game are different in tech
15:54Furman's in standard business one of the
15:57smartest hedge fund managers I know he
15:58says they're still in financial markets
16:00is what he calls the New York Palo Alto
16:02arbitrage right and it's achill II he
16:04said a strategy spent half his time in
16:05New York and understand what all those
16:07assumptions are basically are the
16:08drivers New York is the driver of asset
16:10prices it's or most always my investors
16:12are at least in the US and then he says
16:14basically Palo Alto figure out all the
16:15ways they're wrong and then place the
16:16contrary bed and the theory I think that
16:19you're laying out underneath that is
16:20basically you might say that the New
16:21York mindset stereotypically might be
16:23the halls of reduction mindset yes even
16:25still yeah right it certainly is that
16:27way in Europe right I'm always amazed
16:30and slightly appalled that people think
16:33of technology in Europe as something
16:35that's done by very big companies that's
16:37pretty good technology but they don't
16:39get that this is a game of positioning
16:43of building user base and it's well
16:45understood in California it gets less
16:48well understood on the East Coast and
16:50they're not very well understood
16:52elsewhere so question another very smart
16:55Peter Thiel it takes it a step further
16:56he asserts that in the long run every
16:58kind of industry every kind of product
17:00either becomes a monopoly or a commodity
17:03in other words in the long run emergency
17:05either go to infinity or they're 100
17:06percent or they go to zero and it's just
17:08a question of time if you don't have
17:09increasing returns you're in a long-term
17:11downward slide to commodity yep and he
17:13asserts that the things we view as
17:14intermediate cases businesses today that
17:16are like 20% margins are fated to
17:18decline to 0 over time is his view do
17:20you think too extreme or would you
17:21support even if you had kind of that
17:22start I like the idea I think is
17:25basically on target but there are
17:27perennial e commodity industries I'm
17:31thinking of Airlines where the margins
17:33are pretty low they're usually lower
17:37but still these persist and quite often
17:40governments intervene yeah I have a lot
17:42of sympathy for Peter Thiel's view I
17:44think that in the long long run things
17:47do tend to get dominated by only one or
17:51two players even in the standard
17:53businesses and the reason that's not
17:57completely and utterly true all the time
17:59is that there are new products getting
18:03launched all the time in standard
18:05product space and that keeps us in this
18:08more standard economic setup when you
18:11describe the work of increasing returns
18:13yeah you also mention the flipside of
18:15this sort of effect of increasing
18:17returns which is sometimes you might get
18:19to the point where the network can go
18:21back to a point where it goes to
18:23diminishing returns yes for example if
18:25there's too many listings on a market
18:26place or something do you've any
18:28thoughts on that or any new takeaways
18:30around that cuz if the network is more
18:32valuable as more people use it why would
18:34there be a diminishing return at a
18:35certain point if it gets too big is it
18:37like is there an ideal size no I don't
18:40think so I think it depends very much on
18:42the network itself some networks can
18:45eventually become commoditized and so if
18:49it's a commodity anybody can sort of
18:51come in it and offer the same thing but
18:54much more common pattern the pattern
18:56that I would expect is that there is a
18:59network go back to 1984 Microsoft moves
19:04in other companies move in Microsoft
19:06dominates but eventually what happens in
19:10increasing returns market is that the
19:13next invention comes along great and
19:15some other companies offering web
19:18services or something comes to dominate
19:20so you can dominate for a while in one
19:25large niche in the digital economy but
19:29then the next set of technology comes in
19:31the new players come in with that Google
19:34recognizes and Google's trying to stay
19:38by being in on the news well it reminds
19:40me when they tried to do like social
19:41networking when Facebook came along and
19:43now they're sort of just decided to
19:45become an AI first company yes that's
19:47they're gone but companies don't always
19:49make that transition from one technology
19:52to the next very well apples been very
19:56lucky but they've invented some of the
19:59technologies and then they're able to
20:01serve from that new board so to speak
20:03but the overall things that logins tend
20:08to last for a certain amount of time and
20:10then they become obsolete and some new
20:12game right or they become ubiquitous
20:16utility like a new game still comes
20:18along I would argue that Google's always
20:20gonna be around for oh yeah sure because
20:22they've sort of dominate that market but
20:24they may become like utility in that
20:26application that's right and then the
20:27advertisers may drift off to something
20:29halter you mentioned earlier that you
20:31don't think it's luck and this
20:33discussion makes it almost sound like
20:34it's an accident that there's a
20:35winner-take-all effect but is there some
20:37way of knowing early on the entrepreneur
20:39who maps out the future who knows the
20:41ecosystem how do we sort of know that
20:43these are the ones that we'll figure out
20:45how to tip the market in their favorite
20:47what are some of the indicators it's not
20:49an accident like they're pulling levers
20:51there McAfee owner she shows how hard
20:54this is to predict I remember sitting in
20:571991 I was invited to the Senate
21:01building to brief Al Gore who was the
21:03senator then it was an afternoon was
21:07quite hot and they're all sitting there
21:09everybody was a little bit sleepy and
21:12the gore says can you give me an example
21:14I can latch onto and I said yeah
21:17presidential primaries the phenomenon
21:23I'm talking about you know if something
21:25gets ahead of tense get further ahead
21:27it's true in presidential primaries that
21:31if some candidate pulls ahead they get
21:33more financial backing they can be more
21:36visible more visible they are the more
21:38likely it looks that they might win the
21:41presidency so they get further ahead and
21:44more backers you have to be quite away
21:48into the game before it's pretty clear
21:50that's the best I can do on that meaning
21:53sometimes this was very early tilt like
21:57within a few months it's pretty clear
21:59what's going to take
22:00/ but it can be very much like
22:03presidential primaries that's all the
22:05same mechanism and predicting exactly
22:10might look easy afterwards but on the
22:16spot it's very difficult to do well this
22:19goes to a shift in nature I think of how
22:20history is written right yes the way
22:22history gets written is the victor is
22:24imputed all kinds of positive qualities
22:28marvelous executor right and everybody
22:30new right everybody pretty everybody
22:32predicted and then of course the people
22:33who don't win it's like Oh idiot you
22:35know losers what were they thinking we
22:37experience this in venture capitalists
22:38like we basically get two kinds of press
22:40coverage one is what a bunch of geniuses
22:41we were for backing the successful
22:42company what a bunch of morons we were
22:44back in the failing company yeah and I
22:46keep pointing out we're the same people
22:47we don't wait between genius and
22:48we're somewhere in the middle yeah but
22:50your point is the nature of the
22:51technology casino the other thing I've
22:53observed is on this point is I don't
22:54know cynical sense maybe realistic sense
22:56in a sense the question of like what is
22:57the spark that causes one to jump ahead
22:59kind of doesn't matter less cynical way
23:04to put it might be there might be 20
23:06different ways somebody gets an initial
23:07jump it might be they start two months
23:08earlier it might be they raise a little
23:09more money it might even get a key
23:11distribution partnership whatever it
23:12kind of doesn't matter exactly what it
23:13is as long as yeah as long as there
23:15isn't and so there's a lot of
23:18idiosyncratic kind of history to these
23:20things yeah and my shorthand term for
23:23all that is luck of course there's no
23:25such thing there's just all small events
23:28who sat beside whom and a airplane and
23:31chatted up somebody or whatever or whose
23:34mother happen to be on the board of
23:35United Way as well as one example yeah
23:39yeah right all right this is item famous
23:42Bill Gates biography story where because
23:44his mom was on the board of United Way
23:46she met the CEO of IBM and then that
23:49meeting led to Microsoft and IBM
23:50striking a software deal that helped
23:52Microsoft in the early days right
23:55the other interesting kind of situation
23:57that we run into a lot on this when we
23:58try to figure this out is it's fairly
24:00often you'll have a scenario where
24:01you'll have to you might have 20 in the
24:03field but you'll have two companies that
24:04you kind of think are though to have
24:05those ability winning and one of them
24:08has is a little bit further ahead but
24:10has a somewhat less skilled or
24:13and then you'll have another company
24:14that maybe started a little bit later
24:15that will be further behind for the
24:16moment but has a much more experience to
24:18qualified yes founders yeah and if
24:21you're going entirely based on current
24:22trends you go with a less experienced
24:24less knowledgeable founder on the other
24:26hand you often have somebody very sharp
24:28was like oh yeah I know exactly what I'm
24:29doing he doesn't what he's doing I can
24:30take him out and like that's a real
24:32that's a conundrum that we face every
24:33and it really elevates this kind of
24:35question of like how important actually
24:36is skill I mean you've pretty much
24:38answered your own question I think skill
24:41is extremely important but it's not tech
24:43it's not even skill and raising money
24:46those are kind of necessary and but not
24:48sufficient what sort of skill is really
24:51really important is strategic skill it's
24:55feel for how to build here how to build
24:58up there basically it's I often thought
25:01of this as surfing you either get a wave
25:04or you don't if you get the wave the
25:07whole momentum of that wave pulls you
25:09forward and then you've got to maneuver
25:10and stay in the green water oh yeah that
25:13was an analogy that Pete Broly used to
25:15use a lot at Parc for innovation because
25:17he's such an avid surfer he would
25:18compare the two and I remember reading
25:20an article years ago by Jay SB as well
25:22that compared executives to surfers but
25:25let's actually now shift gears and talk
25:26about like you know once you understand
25:28these concepts that we've been talking
25:29about so far once you had these building
25:31blocks like network effects and positive
25:33versus diminishing returns that you can
25:35essentially manipulate to pull levers
25:36and get the outcome you want maybe luck
25:38maybe not the bigger question is are
25:41there macro forces at play here too I
25:43don't mean macro economic I mean more
25:44around the nature of technology and how
25:46it evolves which coincidentally is the
25:48name of the book you published in 2009 I
25:50have a copy from you on my shelf
25:52anyway it surprised me that you once
25:53argued that tech evolution is not like
25:55evolution in the obvious sense so tell
25:57us about that well yeah about quite a
26:00while ago about 15 20 years ago I got
26:03really interested in where technology
26:05comes from and the idea around it we
26:09have is that there's some genius and an
26:12addict or something eg garage garage
26:15that's right cooking up technology and
26:18coming up with invention the sports
26:20started to become clear to me as having
26:22looked in detail at some inventions
26:26that technologies in a way come out of
26:29other technologies if you take any
26:31individual technology say like computer
26:36in 1940s it was made possible by having
26:40vacuum tubes by having relay systems by
26:44having very primitive memory systems
26:47maybe mercury delay tubes all of those
26:50things existed already and so it seemed
26:53to me that technologies evolved by
26:57people not so much discovering something
26:59newer inventing but by putting together
27:01different Lego blocks so to speak in a
27:04new way once something was put together
27:08like say a radio circuit for
27:12transmitting radio waves
27:14it could be thrown back on the Lego set
27:16and occasionally then some of the new
27:20combinations would get a name and be
27:22tossed back in things like gene
27:24sequencing were put together from
27:26existing molecular biology technologies
27:29and then that becomes a component right
27:32and yet others I mean CRISPR is a great
27:33example because you have CRISPR which
27:35itself is a gene editing tool the
27:37opening great that tool will be a
27:40component and future technologies and I
27:42began to realize this wasn't Darwinian
27:45wasn't darwin's mechanism its evolution
27:47but it wasn't that your very radio
27:51circuits are you very right that like a
27:53natural selection effect
27:55yeah you can't vary radio circuits and
27:58then suddenly get a computer out of that
28:00or radar you can't very air piston
28:04engines in 1930 and get a jet engine out
28:07of that these things come along is
28:09completely new combinations using new
28:11principles and that keeps adding to your
28:14lego set and that it starts to explain
28:17why there's a controversy or a question
28:20say in the 1920s anthropologists were
28:23asking why don't you have trams and
28:27steam engines in the Trobriand islands
28:30and they began to say well it's not
28:32because the Islanders are stupid it's
28:35because they don't have these
28:37building blocks to build it out of and
28:40that in turn as many implications one of
28:43them is if you get a region like Silicon
28:45Valley with an enormous number of these
28:48building blocks and more important it
28:50has the people who understand the craft
28:53to put all this together not just the
28:55science but what parameters then it can
28:59very quickly keep coming up with new
29:01combinations what's the implication for
29:04adoption though for industry the
29:06implication is that if you have a new
29:08collection of technologies let me just
29:11mention AI artificial intelligence those
29:14are all building blocks industry doesn't
29:17adopt AI a is a slew of technologies
29:22it's a new Lego set industries using its
29:26own technologies and what really happens
29:28is that industries the medical industry
29:31the healthcare industry the aircraft
29:35industry the financial industry they
29:38encounter this new Lego set of AI and
29:42they pick and choose components to
29:44create their own combined one of the
29:47interesting sort of aspects that I find
29:48is there's a consequence of what you're
29:49describing there it seems to me is a
29:51long prehistory of almost any mess
29:54analogy unquote right a couple favorite
29:56examples I have of that the French had
29:58optical telegraphy working I think 40
30:01years before other people figured out
30:03electromechanical telegraphy so
30:04literally des glass Paris with light
30:07pulses going through this is like the
30:081820s or 1830s yeah some sort of I mean
30:13they were like relay stations but it was
30:15little flashes of light like lanterns
30:17for glass tubes and so there's a sort of
30:18fiber optics 160 years yeah prematurely
30:22the other my other favorite example is
30:23MIT published a great book called tube
30:25years ago so at the prehistory of
30:27television and we think of television as
30:29being like 1930s 1940s Philo Farnsworth
30:31all these guys it turns out the idea for
30:33television immersion mmediately upon the
30:35idea for radio and there was a Scottish
30:37inventor named John Logie Baird and in
30:41he invented mechanical television yes
30:43because he couldn't do the electric you
30:45do light you did mechanical and so he
30:46literally had spinning wooden blocks
30:48yeah so he had pixels there's almost
30:50like a computer display but made out of
30:51literally wooden blocks and the pixels
30:53would basically spin the wooden blocks
30:55that spin to form pictures and he
30:57famously it's one of the funniest scenes
30:58in the book is he takes it to the board
30:59of governors of the BBC in 1912 or
31:02something and they're like you are so
31:03completely out of your mind yeah he's
31:06just like oh just let me prove it let me
31:08prove it let me get some sets out there
31:09I'll prove that people want to do this
31:10and they finally gave him a programming
31:12they gave him access to the radio
31:14frequency a Thursday night midnight for
31:1615 minutes and he was broadcasting for
31:17months nobody ever saw yeah right and
31:21then 30 years later right people picked
31:22it up and actually made the version that
31:23works and so I just so then I go throw
31:26this just kind of say so then what we
31:27could project forward is that all of the
31:28breakthrough technologies of the next 30
31:3040 50 years yes they already in a sense
31:33exist yes that's right in some form yes
31:35is that pretty much to get a new
31:37technology you need two things you need
31:39to sort of have a principle meaning a
31:41way of doing things early television
31:43worked on this idea that you could pick
31:46up pixels or little snippets of light or
31:49darkness in the image you're looking at
31:52and then transmit those by radio very
31:55high frequency decoded at the other end
31:57and reproduce on some screen or another
32:00so yeah you need principle and you need
32:03the components there's a famous example
32:05Stanford was involved in in the very
32:08early 1900s the US Navy was very
32:12interested in telegraphy or Telegraph
32:16what they had at the time was spark
32:18radio so it sort of you know send these
32:21Morse code things because the whole
32:24spectrum anybody could pick it up so
32:27they were looking for a perfect sine
32:29wave it's continuous time radio a
32:32continuous wave not just a spark wave at
32:36a single frequency there was a company
32:38formed Pacific Telegraph sometime around
32:411906 1907 they managed to get the guy
32:46who invented the triode vacuum tube for
32:48no divorced yeah lee deforest came out
32:51from yale kind of on the run
32:53from predators DeForest and federal
32:57Telegraph spent several years trying to
33:00get perfect sine wave so they could
33:03transmit radio waves on a single
33:05frequency offshore to naval vessels they
33:09couldn't really do it and in 1912 AT&T
33:14put out a call for inventions their idea
33:18was to be able to telephone from New
33:20York to Chicago but you needed to have
33:23some sort of repeating circuitry needed
33:26to clean up the wave every 20 miles or
33:30so and then retransmitted it turned out
33:33that within about six months three
33:35inventors to forest among and come up
33:37with a triode vacuum tube early
33:40amplifier that amplifier was fed back
33:44that becomes an oscillator kind of like
33:46a microphone shrieking the oscillator
33:49gives you a perfect sine wave you could
33:52modulate that and send that out as a
33:56radio message to ships offshore or to
33:59anything and if I recall right this is
34:03quite early in the game these radio guys
34:06or the headphones they were always
34:08called sparks the radio officers and
34:11ships they were listening to Morse code
34:14one time not very far from here and
34:17suddenly somebody transmitted music
34:27music and it's a bit of a long story but
34:36the point is that individual inventions
34:40like the triode vacuum tube when put
34:42together in clever ways with other
34:45components give you an oscillator which
34:48is the basis of radio transmission they
34:51give you radio receivers etc and that
34:54builds up the broadcasting industry
34:56which in turn parts of that are used to
34:59give you television and then in real a
35:02form on or off switches these things
35:07to give you logic circuits and in turn
35:11that gives you early so the technologies
35:16don't come out of nowhere they come out
35:18of a very deep understanding of what's
35:21in the Lego box and how to put those
35:23things together well so the pessimistic
35:26view on that would be boy that means by
35:28implication that really aren't the kind
35:29of Eureka moments that people think
35:31about and the pessimistic view and that
35:32is then therefore there's really not
35:34going to be anybody sitting around in
35:35the next 20 years is gonna say I want to
35:36build warp drive and their professor
35:38than light travel and they're just going
35:39to come up with it or immortality or
35:41whatever you know these and so in a
35:43sense they're it's an argument against
35:44kind of dramatic innovation let's just
35:46say determine innovation on the other
35:47hand it's an optimistic argument because
35:49it says the number of combinations of
35:50the literally exactly infinite over time
35:54well it did argue that there are
35:56breakthroughs you know there are Eureka
36:00moments they tend to work that I'm
36:03sitting here wondering how I could get
36:05some effect how could I transmit images
36:07by radio wave and I could be sitting
36:11there thinking for months well I could
36:12use this combination that combination
36:14and other combination and then suddenly
36:17I realize if I can get this in place and
36:20that in place the other thing in place
36:22that's gonna work and the interesting
36:26thing is and I've read individual
36:28accounts either dozen from inventors
36:32even lab books you see this again and
36:34again can't do it can't do it can't do
36:37it and then oh oh oh one of my favorite
36:42stories is that a steam engine already
36:44existed way before James Watt and James
36:48Watt in the 1760s I think it was in
36:53Glasgow was brought in to see if he
36:55could improve it so what thinks it over
36:59and he thinks oh well you know you're
37:01heating the steam you're expanding it in
37:03a cylinder then you're suddenly cooling
37:05it again and all of this is pretty slow
37:08what if I allowed the steam to expand
37:12the cylinder and then that steam is
37:15ejected into a second cylinder
37:18kept at very low temperature suddenly
37:20the steam collapses as a vacuum etc so
37:24he invented an independent cold cylinder
37:27he thought of it passing the village
37:30green on a Sunday Sabbath day he was
37:34properly Scottish it nearly killed him
37:37Isis and there I was they season me
37:40knows it's gonna work but he can't get
37:42into his workshop until Monday you can
37:45just read this stuff and see this half
37:48kilogram that he can't prove the concept
37:52he's a machinist and he got it to work
37:54fairly readily he's using the building
37:56blocks to basically people are using
37:58existing building blocks to do this sort
38:00of combinatorial situation combinatorial
38:02yeah the point I'm making is that new
38:06technologies don't build up as just pure
38:09inventions there's plenty of
38:10breakthrough insights but they build out
38:14of what's already there the components
38:16and quite often then new things come
38:19along some key breakthrough technologies
38:21deep learning this one CRISPR is another
38:24right these aren't just isolated
38:25components they themselves are tools and
38:27literally recombine or create other
38:29technologies and by the way in that
38:30sense I think it is very much like
38:32evolution I mean we had you've all her
38:33re on the podcast too and basically in
38:36his book sapiens he argues that tech
38:37helps mankind leapfrog natural evolution
38:40and only in that context we were talking
38:41about it across a much larger time scale
38:44but in this context I do think of it as
38:46a primordial soup for the next phase yes
38:48on that know you mentioned deep learning
38:49which we think of it as basically
38:51machine learning distributed computing
38:52artificial intelligence I mean just for
38:54this purpose we can broadly clump that
38:56into one category and I remember a big
38:58piece you did for McKinsey quarterly
38:59right before I left Park it was around
39:002011 and it was on the second economy
39:03basically an autonomy economy and
39:05actually you should surprise this
39:06because then I'd like to talk to you
39:07about how you might update that today
39:08given all the advances in AI sense sure
39:12what I was pointing out was that there's
39:14a familiar physical economy the one we
39:17all know about it has to do with retail
39:19stores and factories and banks all the
39:23stuff that we see in the physical world
39:25I was checking into a flight in San Jose
39:28airports sometime around 2011 and when I
39:33put my frequent flyer card in that
39:36suddenly it was triggering a lot of
39:38processes certainly the flight was being
39:42alerted that I was now there maybe TSA
39:46was being alerted so I began to realize
39:49that some others a huge second economy
39:52out there of machines talking to
39:54machines I was thinking of it as a very
39:58large underground unseen invisible
40:02economy could be in the cloud of servers
40:07talking to servers of software and
40:10algorithms talking the servers talking
40:12to other servers all being transmitted
40:15and in conversation always on and
40:19occasionally then putting out shoots up
40:22into the physical world and it reminded
40:24me as a metaphor of aspen trees aspen
40:28trees apparently are one huge organism
40:31that is they're all connected
40:33underground with the same root system
40:35and what you see on the surface is the
40:38trees themselves but there is a very
40:42very large underground root system
40:45that's all connected these roots are all
40:47talking to each other and this would be
40:50like the second economy I now think I
40:53should have chosen the term virtual
40:56economy or better still the autonomous
40:59economy because all of this is happening
41:01without our knowing it's an autonomous
41:04it's things talking to things so I don't
41:07emphasize an internet of things it's
41:10more like an internet of conversations
41:13things triggering things things
41:15switching off things things querying I
41:18mean just to give it a quick picture if
41:20you have that image of you putting the
41:21card in the kiosk at the airport and you
41:24have all these machines talking each
41:25other if you were the light up all those
41:27machines at once they'd be all around
41:28the world goes there be servers and
41:30amazon's cloud there would be something
41:32local the local printer there'd be
41:34something else like a processing payment
41:36being maybe in Palo Alto there could be
41:38all these different pieces kind of
41:40not just and yes and not just a few
41:43dozen computers or servers lighting up
41:45because those servers would be lighting
41:47up other server right and so there in
41:50the end there could be hundreds of
41:51thousands of servers that were lighting
41:54up very briefly maybe only for a few
41:56fractions of a second and then shutting
41:59down again and then passing messages so
42:02I was interested in this autonomous
42:04economy there was a general conversation
42:06about automation and robots and 3d
42:10printing I thought no they're missing
42:12the point I tend to think that the
42:15digital revolution I believe there is
42:16such a thing I believe it keeps morphing
42:19or changing about every twenty years the
42:21digital revolution gets a new theme and
42:24the latest revolution comes almost by
42:28accident that in the 2010 xur so we
42:33started get huge numbers of sensors
42:36sensing chemical sensing visual pixels
42:41sensing images sensing temperatures hunt
42:44by the hundreds and dozens and hundreds
42:47of thousands and all these sensors out
42:50there and they were maybe feeding back
42:52from smartphones or from your car and
42:55huge amounts of data about the same time
42:58this was no coincidence along comes a
43:01new generation of neural networks
43:04powered by deep learning but more than
43:08powered by all the data that the sensors
43:10were bringing us and these algorithms
43:13started to be able to do one thing very
43:16well and that was pattern recognition
43:18could recognize your voice much better
43:21than before because of all the data all
43:24the training it could recognize phases
43:26so suddenly we got the ability of
43:30algorithms to do things that we thought
43:33only humans could do
43:34as recently as twenty years ago or ten
43:37years ago we would have said oh yeah
43:39computers are great but they'll never be
43:41good at what you humans are good at in
43:43this what are humans good and we're good
43:45at recognizing things were good at fast
43:48Association computers they can do
43:53we're not much good at logic so it
43:55seemed that the whole world was nicely
43:56divided but now but now computers have
44:00learned to associate of thinking these
44:03patterns mean such and such and so
44:07suddenly were in an area that we thought
44:09only human beings were going to be good
44:11at and we're seeing industry after
44:15industry change as a result it's not
44:18just automation it's much more than that
44:21it's redoing or restructuring whole
44:24areas of the economy so I was looking
44:27for an analogy what in history that even
44:31vaguely resembles what's happening the
44:34printing revolution starting around the
44:361450s suddenly information went from
44:40being very closely guarded by
44:42monasteries and Abbey's and libraries
44:45these big vellum books chained desks and
44:49with printing that became publicly
44:51available so printing made information
44:55externally available and that changed
44:58everything it very much changes the way
45:02Copernicus for example how does his
45:05disposal data that he could not have got
45:08hold of if they just existed in
45:10monasteries it made a huge difference it
45:13brought in modern science it helped the
45:15Renaissance and this brought us our
45:18modern world I mean I would agree but is
45:21that the big transformation now that we
45:22have the modern tech equivalent of the
45:24printing press what's gone external now
45:27is not information what's gone external
45:31is intelligence I may be driving in a
45:34convoy of 50 driverless cars and the
45:38whole idea of the car adjusting the car
45:41is talking to roadside sensors and
45:45servers it's talking to other cars is
45:48talking to the Highway Patrol servers
45:51and so on as basically farming out its
45:54intelligence into this other economy and
45:58then getting back intelligent actions in
46:00return so it's a bit like phone-a-friend
46:04only the friend is incredibly smart and
46:07the friend consists of again these
46:09hundreds of thousands of servers talking
46:13to each other and then adjusting what
46:15you do so suddenly intelligence doesn't
46:18just exist on human beings
46:21suddenly intelligence exists in the
46:24cloud or in this autonomous economy and
46:27we conform at not just getting
46:30information but getting smart moves back
46:33and this is making all the difference
46:36it's not about the form intelligence
46:38takes it's that intelligence is no
46:40longer housed internally in the brains
46:41post role of human workers because it's
46:43moved outward into the virtual economy
46:46so when intelligence is not just
46:48information but sort of decision-making
46:50or being able to external eyes a lot of
46:52this I mean one of the things you
46:53mentioned earlier is about these
46:54building blocks of technology what
46:56happens when all of these things are
46:58available to everybody equally like is
47:00there not like how to serve a Red Queen
47:01effect where everyone's accessing the
47:04same building blocks and tools so how do
47:06companies how do industries find
47:08competitive advantage in that kind of a
47:11world ah I think the answer to that
47:13question is timing if I'm a retail bank
47:16whatever that might be I might be quite
47:19a large bank and I'm saying all these
47:22externally intelligent technologies and
47:25the algorithm is suddenly available how
47:28can I make use of that and how can I
47:31bring those into my operations and
47:34combine them with what I'm doing them
47:35making mortgage loans I'm acting as
47:39escrow or something you know all these
47:41various different types of financial
47:45operations I can make a lot of them
47:48automatic and autonomous and get an
47:52advantage the trouble is that that can
47:55be rapidly commoditized so what does
47:58that mean for jobs and there's bad guys
47:59we talk a lot about how whenever
48:01industries are changed in this way you
48:02know through tech and other shifts that
48:04other new jobs classic examples include
48:06more designers in the age of Adobe
48:08design that new jobs never existed
48:10before like social media managers I can
48:12only exist today what's your take here
48:15what I'm seeing is about 90 years ago so
48:19John Maynard Keynes pointed out that he
48:23thought by a hundred years time 2030
48:26we'd be in an economy where the
48:28production problem was largely solved
48:31there'd be enough in principle to go
48:33around for everyone there might be
48:35plenty in principle goods and services
48:38around but getting access to them meant
48:41you needed wages which you needed a job
48:43for and that was not possible I think
48:47that's what Keynes said in that regard
48:49is becoming true in other words the
48:52trough is full but how do the piggies
48:54get their share of the trough so we're
48:58now in a new distributive era what
49:00counting is not how much is produced but
49:03who gets what the whole question of
49:07growth and getting more economic product
49:11out there physical product and services
49:14that's a job for entrepreneurs and job
49:17for engineers who gets what is much more
49:21a political issue and that's not quite a
49:24job just for politicians but it's a job
49:26for society to solve and we haven't
49:29solved it in Europe or anywhere else so
49:31it's a new era so the problem with that
49:35theory is the same problem is that
49:36they're in in Keynes era right which is
49:38sort of Milton Friedman's observation in
49:39the nineteen fifties nineteen sixties
49:41when that issue came up again which is
49:42that human Watson needs are infinite
49:44right yeah we are one of the things we
49:46are best at as a species is coming up
49:47with new things and then the things that
49:50we want one generation become the things
49:52that we need in the next generation air
49:53conditioning goes from being a luxury to
49:55being you know something that we
49:55expected everything else and you know he
49:59speculated as a thought experiment he
50:01said look you know we have no way of
50:03envisioning that wants and needs of what
50:04people will have in the future is no
50:05they'll be there he said look maybe
50:06it'll be a bit like you know right now
50:08psychiatry is it looked very good and
50:09maybe in the future it'll be a basic
50:11human right to have access to a
50:12psychiatrist and we'll employ half the
50:13population basic high interest to the
50:15other half and just as one example
50:21okay exactly and so and then in economic
50:24terms of course the problem with Keynes
50:25analysis was it overlooks the role of
50:26productivity growth right that's just
50:28the scenario that you're describing
50:29there's a scenario of like rapidly
50:32increasing productivity growth and in a
50:33world of rapidly increase in
50:34productivity growth you have gigantic
50:37gains in economic welfare you have
50:39gigantic growth market lying industries
50:40right you have gigantic amounts of
50:42entrepreneurial activity that come out
50:43of that and that then generates a
50:45fountain of new jobs to satisfy all
50:46those new wants and needs and then
50:48finally I can't resist pointing out that
50:49you're making this argument on a day
50:50when the unemployment rate in the u.s.
50:51drop below 4% there's certainly no trace
50:53and remember once a day in the American
50:55economy you actually have very low
50:56productivity growth not very high
50:57productivity growth yeah which is
50:59counters against the argument that
51:00there's some level of unprecedented
51:02technological disruption that's
51:03happening because you could certainly
51:04can't see it in the numbers and then and
51:05then you have unprecedented levels of
51:07job growth and employment sure so the
51:09facts seem to be on the other side of
51:10well let me both agree and disagree here
51:12I certainly agree that there would be
51:15whole new categories of jobs I very much
51:17like the idea that we can swap a
51:27I think there'd be plenty of new jobs
51:29invented at the same time though not
51:33just through automation and not just
51:35through algorithms but over the last 20
51:38or 30 years we've had a huge amount of
51:40globalization jobs have been offshored
51:44and that's not just due to the rise of
51:47China it's due to the rise of
51:49telecommunications of I can keep track
51:51of all the suppliers in China all the
51:54factories in China the inventories and
51:56so on in real time couldn't have done
51:59that much in the 1980s because the
52:02technology wasn't there and that
52:05hollowed out an enormous amount of
52:08traditional workers in the middle of
52:10America and certainly in Britain and in
52:13other countries so where I would come
52:15out on this question
52:17I like your observation I agree yes we
52:20will get new jobs but quite often
52:21there's a big lag in between the
52:26original happening of hollowed-out
52:31something taking its place an analogy
52:34that I like said in Britain in the 1850s
52:37the economy was going gangbusters a new
52:41textile companies the railways just
52:44starting to kick in there's all kinds of
52:46possibilities steel works everything
52:49that's suddenly very serious and at the
52:52same time so there are people getting
52:54very rich but at the same time there
52:57were there's child labor there are the
53:00 finian world the whole
53:01Dickensian world of people almost being
53:04worked to death both are true the
53:07economy is going gangbusters some people
53:10are not doing well out of this it took
53:12about 30 to 60 years before the whole
53:15thing equalized and workers had safe
53:18conditions at much better conditions and
53:20eventually the rebel to partake in a
53:23decent way in all this wealth creation
53:26so what I would say is that the digital
53:28economy through globalization and now
53:32through algorithms is pressing us into a
53:37scramble to invent new categories of
53:40jobs I'm optimistic I think eventually
53:42we'll get on top of this and I'm hoping
53:45we do it in a good way where we have
53:47creative pursuits not just rote jobs
53:51like we might have had a hundred years
53:54ago I think things are going quite well
53:56good so it is a global world now and it
53:59depends on what your frame of references
54:00for me my frame of reference is I have
54:02relatives in India and we're now
54:04increasing in their middle class if your
54:06frame of reference is global you see
54:08this as a very different kind of shift
54:10it really depends on where you sort of
54:11put the square the rectangle of the
54:12frame and rows min because there is
54:15Africa another great example Cambodia
54:17you have all these countries there's
54:19something interesting happening there so
54:21speaking of that I'd love to hear
54:22because you spent a lot of time in
54:23Singapore yeah I'd love to hear your
54:25thoughts on sort of the evolution of
54:27that because we've often made the
54:28argument that this kind of form top-down
54:30government planned innovation cluster
54:33never works out and Singapore is a rare
54:36how would you distill it having been on
54:38well I'm a watcher of countries that
54:42look as if they're in trouble and then
54:44make their way out of trouble
54:46Finland's a good example because the
54:49Cold War shuts down Finn Andrews broker
54:51a bit like Hong Kong in between the west
54:53and the East then around 1990 suddenly
54:57the bridges are but the river ceases to
54:59exist and so then they invented their
55:02way out of that with Nokia and other
55:04companies they're backwards to the wall
55:07and I could say the same thing in
55:09Singapore when the country was set up
55:12about 51 or 52 years ago they felt very
55:15much as if there had been set adrift so
55:18like a little rowing boat it was being
55:20towed behind Malaysia and then somebody
55:23cut the rope so I think again it was a
55:26matter of desperation very good planning
55:28people like Lee Kuan Yew who led the
55:31government and what they did was they
55:34decided that they would go into what was
55:37then tech manufacturing they had
55:40inherited shipyards from the British etc
55:43so they're able to station themselves as
55:45a very early manufacturer but like Hong
55:48Kong or Taiwan produced cheap goods and
55:53may take great advantage that the oil
55:55tankers had to stop that and become a
55:59commercial and brokerage hub for
56:01shipping since that they've moved into
56:04finance what I'm finding and let me
56:07broaden into Asian countries including
56:11China we tend to think of as recently as
56:1410 years ago we would have thought of
56:16China as being not fully developed
56:19not at all like Japan which is developed
56:22Singapore's quite developed what we're
56:25now seeing in Asia is that a lot of
56:28countries in Asia including China their
56:31digital revolution is not much more than
56:34two to three years behind what's
56:36happening in California or in the West
56:38they're extremely well advanced they're
56:41paying a huge amount of attention to
56:45and it's not just that they're following
56:47in China they're not just following say
56:50genomics or AI they're inventing their
56:55own Singapore by dint of strong will and
56:59going techie has managed to do that
57:03already what I do notice in Singapore is
57:06that they tend to not so much initiate
57:10perfectly new technologies but they're
57:13very quick to take them up China though
57:16is able to initiate initiate them
57:18especially in things like genomics do
57:20you think the initiation thing matters
57:22because the part of your thesis around
57:24there being these building blocks that
57:25are widely available which leads to this
57:27combinatorial innovation combinatorial
57:29notion as you describe it I wonder if
57:31that even matters so much anymore
57:33because give these building blocks
57:34open-source API is all are available
57:37like application programming interface
57:39that's people can combine into entirely
57:41new companies it seems like you can
57:43actually draw on the best of the best
57:44expertise I think so that's been a long
57:48debate actually in economics why put all
57:52the effort into initiating something
57:54when you can just position yourself to
57:56learn the technology quickly the other
57:59case is good to be first
58:01I think it's debatable what I would say
58:04though in China is that when it comes to
58:07a country digitizing everything China
58:10isn't going to be far behind
58:12it's especially true of AI actually yes
58:14especially in artificial intelligence
58:17and in genomics and probably in several
58:20other industries no mix is particularly
58:21interesting because they're the first to
58:23do human scale studies of CRISPR yes
58:25because we regulatorily
58:26rightly so we may not be able to or
58:28maybe not so rightly so I don't have an
58:31opinion on that what I see is this sort
58:33of technology expanding rapidly into the
58:36rest of the world and the other country
58:39of course to mention is India for
58:42technological education India has been
58:45IIT places like that Bangalore and India
58:51is not very far behind
58:52China is in a better position because
58:54China is top-down hierarchical they can
58:59change their economy we go back and
59:01forth around that's all the time but
59:02every past industrial planning top-down
59:05centralized model of coordination has
59:07eventually eaten its own yes I'm falling
59:09on its own like hoisted on its own
59:11petard to use that expression another
59:12which is kind of the thing that I
59:14inevitably seems to happen but what
59:16it'll inevitably happen with China well
59:19it may never be happen in the United
59:21States - that's a good point I do think
59:24that occasionally the economy's got a
59:27bit tired people get complacent etc I
59:30was in India I've been there several
59:33times but a long time ago like 1975 and
59:37there were old English cars Morris
59:40minors Rea driving around taxis that you
59:44wouldn't have seen since the 1950s in
59:47London and the Indian economy has gone
59:51light-years beyond that well I was
59:54saying that one of the other shifts
59:55there which is important to note here
59:57for this part of the conversation is
59:58that India China Singapore they've moved
01:00:02away well India went through an
01:00:03outsourcing phase as you describe this
01:00:05to being originators of their own
01:00:07destruction they're not just a copycat
01:00:09narrative and we've written about this
01:00:10when it comes to China as well I mean
01:00:12just yesterday Walmart announced this
01:00:13buying Flipkart which that's kind of an
01:00:16inversion of the typical model that
01:00:17would have happened before so anyway I
01:00:19think that's an important shift that
01:00:21this is playing out yeah the rest of the
01:00:23world is very rapidly catching up I
01:00:26still think that the US economy is going
01:00:28to do extremely well that's great mystic
01:00:31well it's not just optimism I things
01:00:34pretty well in a vertical let me restate
01:00:38this I think what's going to happen the
01:00:40next decade or two the story in the US
01:00:44economy is simply going to be that huge
01:00:47industries are going to reorganize
01:00:49themselves along the lines of autonomous
01:00:54intelligence when you describe that the
01:00:56economy has these sort of 20 years beans
01:00:59that you go and you've described them as
01:01:01more things than you're writing like
01:01:03sort of fundamental sea changes and you
01:01:05described integrated circuits already in
01:01:07fast computation as first we talked
01:01:09about the connection of digital
01:01:10processes and now you men
01:01:11and these sensors though cheap and
01:01:14ubiquitous sensors my question for you
01:01:17as someone who's long studies this is
01:01:19how do you know when you're seeing the
01:01:21beginning of one of these revolutions
01:01:23that it's a morphing in the making is
01:01:24this third of a hindsight view because
01:01:26you are sort of seeing it early with
01:01:27everything else what are the signs that
01:01:29tell you this is a morphing this is a
01:01:32big theme that's emerging that gives you
01:01:34the confidence to say that about say
01:01:35deep learning or CRISPR even I think
01:01:37that a change is usually quite well
01:01:40underway before people pick it up you
01:01:44wake up one day and you said oh my god
01:01:46the game has changed in the case of
01:01:48sensors I remember in 2010 or so sitting
01:01:52down with the CTO of Intel and I asked
01:01:55him can you tell me when the average
01:01:58sensor is for example at a parking meter
01:02:01that might sense a car being at the
01:02:04the average sensor is going to drop
01:02:07below by 10 cents per unit and he said
01:02:10yeah that'll be around 2013-2015 he knew
01:02:14pretty well exactly and so I thought
01:02:16that's gonna be a game-changer because
01:02:18we will now know what's happening
01:02:21everywhere what I didn't see at the time
01:02:23was that the ubiquity of sensors would
01:02:27bring in big data some of us saw that in
01:02:31advance but the big data didn't see
01:02:34would bring in all these smart
01:02:36algorithms right and so it's the
01:02:39combination is there a way to see these
01:02:42new things coming along yeah if you're
01:02:45waiting for them this reminds me of a
01:02:47story that Alvy Ray Smith tells he's a
01:02:48park alum as well
01:02:49he co-founded Pixar back in the day and
01:02:51he did a piece for me at Wired about how
01:02:53they knew very early on they had John
01:02:55Lasseter they had this creative vision
01:02:56they knew very early on the kinds of
01:02:58things that they wanted to do and they
01:03:00later mapped out like a trajectory of
01:03:02their movies based on Moore's law but it
01:03:04was like a tool for them so they saw it
01:03:06but yeah they had to wait but usually
01:03:08it's hard to see the best I can hope for
01:03:11at least in my own cases that within two
01:03:13or three years you just go oh no the
01:03:16games change right
01:03:17and when the game changes you realize
01:03:20you're in a slightly different era and
01:03:22when you're in that era you realize that
01:03:26it's not going to last that in 10 years
01:03:29time 20 years time or 30 or some of the
01:03:31beer a different version want to make
01:03:34this comment very quickly I've been
01:03:36physically in the Silicon Valley
01:03:38if you can't Berkeley I've been in
01:03:40navigation yeah okay I've been in the
01:03:43Bay Area now for very close to 50 years
01:03:46I was a grad student in Berkeley and
01:03:49then in Stanford I've been here since
01:03:521982 and then all that time when the
01:03:55game gets a little tired at times people
01:03:58say oh the valleys over but it doesn't
01:04:01it discovers new technologies and then
01:04:03reinvents itself that's the way
01:04:05capitalism works here in Silicon Valley
01:04:07but in other countries where it's more
01:04:11planned it's may have stopped and places
01:04:16like that can come to a halt as a result
01:04:18that's the perfect note to end on I'm
01:04:20gonna quote a piece of one of your
01:04:21middle early papers you talk about
01:04:23whether there's any hope in the
01:04:25complexity essentially and you say it
01:04:27shows us an economy perpetually
01:04:29inventing itself perpetually creating
01:04:32possibilities for exploitation
01:04:33perpetually open to response an economy
01:04:37static timeless and perfect but one that
01:04:40is alive ever-changing organic and full
01:04:44of messy vitality it's not a coincidence
01:04:46that I wrote that because that's where
01:04:50Silicon Valley operates inventing and
01:04:53reinventing itself and morphing and
01:04:55changing in a way you can't quite
01:04:57predict and in a way that I think is
01:05:00delightfully messy but ordered at the
01:05:03same time fabulous a messy ordered
01:05:06vitality thank you for joining the a6
01:05:10and Z podcast yeah thank you Brian that
01:05:12was really tremendous and thank you very
01:05:13much for having me I'm delighted thank