00:04so I thought this was a young a good way
00:07of thinking about where we are today to
00:09talk about the end of the beginning
00:10because it struck me that we've been
00:12looking at charts like this for most of
00:13the last 20 years they always go up and
00:15to the right and now we're getting to
00:17the point that we're kind of finishing
00:18now the access story is reaching the end
00:20because we've connected three quarters
00:22of people on earth and we're going to
00:23connect all of the rest and as we go
00:25forward though the other part of the
00:27story isn't close to being finished the
00:30access story is finished but the use
00:31story is just starting to begin and it
00:34feels like that's what the next 20 years
00:35is about and so most people are online
00:38but most of the money is not and as we
00:40move towards addressing that we address
00:42any problems and we tackle harder
00:44markets and we probably change those
00:46markets a lot more than some of the
00:48markets we've addressed in the past and
00:49of course as we do that we have these
00:51two new fundamental layers to build
00:52things on so we spent the last ten years
00:54talking about and building stuff on
00:55social and on search now we think about
00:58machine learning and we think about
00:59building things on crypto currencies and
01:01so as we go into this I think it's
01:03useful just to start by thinking about
01:04market sizes and so the obvious one
01:08ecommerce looks big in dollar terms u.s.
01:11e-commerce now is about four hundred and
01:13fifty billion dollars a year but if you
01:15zoom out and look at retail spending
01:17it's still only about ten percent and
01:19then if you add bars and restaurants
01:21which of course we're now starting to
01:22address it looks smaller and then it's
01:24interesting to look in some of that and
01:26ask what that retail spending looks like
01:28so gas stations for example are going to
01:31transform and perhaps completely
01:33disappear over the next 20 years and
01:35that's as big or bigger than us
01:36e-commerce and then there's other
01:38segments inside that as well that are
01:39kind of interesting so we spend over a
01:41trillion dollars a year on cars and
01:43parts and maintenance all of that is
01:44going to go away or change or be
01:46completely transformed or relocated over
01:48the next 10 and 20 years but then retail
01:51of itself isn't the only story there's
01:53also all the rest of us consumer
01:54spending which is now over ten trillion
01:56dollars I used to do charts in billions
01:58of dollars this presentation is mostly
02:00trillions of dollars and so then if you
02:02zoom out and think about this the actual
02:04share we've taken so far of the
02:07opportunity is actually much smaller
02:08than it might look just if you look at
02:10commerce itself the same thing comes I
02:13think if you look at
02:13housing so Internet is now the largest
02:16single category in global advertising
02:17and much the same in the USA so now
02:20close to 40% of us have budgets and
02:23advertising of itself doesn't change
02:25very much over time it tends to stay
02:26stable as a share of GDP and so this
02:29leads to people saying well maybe Google
02:31and Facebook are topping out maybe
02:32there's not much grace left for those
02:33companies maybe they've got everything
02:35and there's no room for anybody else but
02:37I think you have to look both at
02:38advertising and at marketing which is
02:40this kind of other block we don't talk
02:41about so much and if you look at within
02:43that block you see things like Direct
02:45Mail and telemarketing which it seems
02:47kind of weird to me that we talk about a
02:49targeted ad on Facebook as advertising
02:51but we don't think about comparing and
02:52we compare that with TV advertising but
02:54we don't compare it with telemarketing
02:56or direct mail and the kind of a broader
02:58point to think about here if you buy
02:59space on Amazon if you buy placement in
03:02Amazon search results and we call that
03:03advertising if you buy exactly the same
03:05thing from Walmart then we call that
03:06marketing and somehow that's not the
03:08same thing but of course it is the same
03:10thing because what we're really talking
03:11about the job to be done is what you pay
03:13to reach a customer and then if you're
03:15thinking about what you pay to reach
03:16your customer there's other pools of
03:18capital to think about there as well so
03:19retail rent is money that you spend to
03:21reach your customer it's money you spend
03:23instead of marketing or instead of
03:24advertising the same thing indeed for
03:25shipping or for service there are whole
03:27pools of capital out there that we've
03:28really only just started to address if
03:30you go globally just sales and marketing
03:32just advertising and marketing in total
03:34is now close to a trillion dollars I
03:35suspect a few um point this out to Jeff
03:38bezel send that line about your margin
03:39is my opportunity starts coming to mind
03:41but this global point I think is also
03:43interesting because we really ought to
03:45think not just how big the opportunity
03:47but where else the opportunity is a good
03:49place to start in this is with another
03:50piece of global technology
03:52industrialization which much like the
03:54internet was sort of originated in one
03:56country and then spread to another big
03:58low-wage country on the other side of
04:00the ocean so this is the u.s. the UK
04:02inventing industrialization then the UK
04:05picking it up and the u.s. picking it up
04:06and taking it forward but then if we go
04:08forward essentially something else has
04:10happened that technology has died fused
04:11and he's gone to other places if we zoom
04:14in on that change where is that
04:16technology gone where is that change
04:18happened well unsurprisingly most of
04:20that balance has gone to China you can
04:22see much the same thing if you look at
04:23computing which was again invented in
04:25the US and in Europe but those countries
04:28to dominate it that was close to three
04:29quarters of computing capacity now it's
04:31much less than that and again of course
04:32a large part of that balance has gone to
04:34China which now has more smartphone
04:35users and the USA and Europe combined
04:37you can see the same picture if you look
04:39at usage this is global mobile in mobile
04:42internet traffic a pack is much bigger
04:44than North America and Western Europe
04:45combined and that of course was always
04:48reflected in the spread of spending
04:49power the global middle class is now a
04:51lot bigger than it used to be but also
04:52it's distributed in different places and
04:54that flows through to things like
04:55e-commerce so from almost nothing China
04:58again is now has more ecommerce in the
05:00US and Western Europe combined all of
05:03this kind of means that as we look at
05:05the position of the u.s. here
05:06penetration of e-commerce is actually
05:08quite a long way behind some other
05:09leading markets China obviously is
05:11leapfrogging there's a leap frogging
05:12story there South Korea in the UK of
05:14course are also well ahead of the USA
05:16although other parts of Western Europe
05:17on the other hand are quite far behind
05:19so there's an interesting picture of
05:21transition here from this being a US
05:23story to it really not been a u.s. story
05:25at all anymore we can see that again in
05:27company creation so these are the four
05:29big companies we spend all of our time
05:30talking about Google Apple Facebook
05:31Amazon in its I think it's kind of
05:33useful to get a sense of the relative
05:34spread of scale and growth in revenue in
05:37these companies because we tend to lump
05:39there's also another company people talk
05:40about in this context which is Netflix
05:42so that's kind of putting that on the
05:43chart to the same scale but I think much
05:45more interesting year to think about the
05:46three big Chinese Internet giant's which
05:48are smaller because they're not global
05:50in the same way but they are equivalent
05:52weight and certainly equivalent gross
05:54but those are just kind of the big
05:55giants the ones that people everybody
05:57talks about being scared of more
05:59interesting again I think to look at how
06:00many other billion five 10 20 50 billion
06:03dollar companies are being created
06:04outside the USA there's a lot of
06:06unicorns being created outside of the
06:08USA being created outside of Silicon
06:10Valley today and of course that's
06:12infected in rich capital which again was
06:13mostly invented in the USA and
06:15relatively recently dominant was the US
06:18was really the only place that this
06:19happened now again that whole business
06:21model that whole approach has been
06:22diffused and so company creation is also
06:26being diffused that gets us back to the
06:28question well what's the opportunity
06:29there's us ecommerce which is four five
06:31hundred billion dollars in that in the
06:33US retail but then there's global retail
06:36and then there's global consumer
06:37spending so we go from thinking about a
06:39couple of hundred billion dollars to 40
06:42potential opportunity that really means
06:44we've only just started scratching the
06:45surface of what we might be able to
06:47address how is it that we address that
06:49what kind of things do we do that are
06:51different does the curve just kind of
06:52keep going on up and to the right well
06:53it feels to me that the things are going
06:55to change a little bit so we think about
06:56the first 20 years of the internet we
06:58really did kind of the easy things
06:59didn't feel terribly easy at the time
07:02there's kind of entrepreneurs who didn't
07:03think it was terribly easy as they were
07:05doing it but what we did was we did
07:07stuff that would work with low
07:08penetration and would work with little
07:09capital it would work with consumers who
07:11weren't comfortable buying stuff online
07:12so they were kind of low touch goods
07:13things you didn't have to hold in your
07:14hand we tended to build tools rather
07:16than build entire stacks we tend to
07:18build information arbitrage we did sort
07:20of price comparison and plane tickets
07:21now we do things that need high
07:24penetration things that need lots of
07:25capital we invert all of these
07:27assumptions we build complete companies
07:28we build full stack companies we build
07:30businesses are based on information a
07:31good case study to think about this good
07:33illustration is to compare kind of a
07:35great company from the last 20 years
07:37like Yelp with businesses like doordash
07:39today Yelp is a low capital though
07:42penetration business it's information
07:44arbitrage business it's selling tools
07:45Jordache has to get your hot meal in
07:47half an hour that's a different kind of
07:49business with different assumptions
07:50about what the internet looks like what
07:51capital you need what consumers are
07:53willing to do comfortable and and
07:55comfortable to be willing to accept
07:56doing online so if we look at that forty
07:58trillion dollars of consumer spending
08:00then there's a whole bunch of things in
08:01here that we've only just addressed very
08:03peripherally or not really addressed at
08:05all and even where we have addressed
08:06them we're going to address them in the
08:08future in completely different ways so
08:10if you think at kind of the first 20
08:11years what did we do with housing well
08:13we did listings and we did price
08:14comparison now open door will buy your
08:15house if you look at transport we used
08:18to do plane tickets now we do cars and
08:19we do autonomous cars and we do Airbnb
08:21up at the top of the stack with Health
08:23well we used to give you a website that
08:24would let you persuade yourself you had
08:26the bubonic plague now we'll edit
08:27molecules and we'll edit genes and we'll
08:29islet cells and create completely
08:31different kinds of cures and all the way
08:33across the spectrum we move kind of to a
08:35different character of business and
08:37different way of addressing these
08:38problems the most visible place that you
08:40can see this happening today I think is
08:41in retail I think a good kind of
08:43conceptual model for thinking about
08:44retail is that there is retailers
08:46logistics and then there is retail as
08:48taste preference recommendation service
08:51and there's a spectrum between these two
08:53and the original Sears Roebuck
08:55hundred years ago did we tell us just
08:57excited as Walmart on the other hand
08:59we've never really managed to scale
09:00retailers taste outside big cities and
09:02outside small shops today we're
09:04addressing both of these so Amazon did
09:07retailer suggest expose successfully and
09:09now Amazon and other companies are
09:10doubling down on that so same day
09:12delivery free delivery above everything
09:14else groceries but at the same time we
09:16move across that spectrum we'd have more
09:17and more things that addressing the
09:18other kind of retail that are not just
09:20kind of a specialized version of FedEx
09:22but actually a question of
09:23recommendation that taste preference
09:25discovery all of this change of course
09:27affects the US probably more than
09:28anywhere else because the US has
09:29massively more retail square footage
09:31than anywhere else so that picture was
09:33likely to change quite dramatically in
09:34the next ten years now as we think about
09:36how we address those I'm actually old
09:38enough now to remember when he called us
09:39was asset light and buying crops was a
09:41really stupid idea in like kind of an
09:43exemplar of how everybody lost their
09:44minds in the late 1990s in fact the
09:46original Amazon business plan from 1994
09:49said no warehouses no stock no shipping
09:51that didn't last terribly long in fact I
09:54think it lasted about three weeks but
09:56then of course that Amazon has become
09:58well not really an asset light anymore
10:00amazon has become an infrastructure
10:02business and the thing that drives his
10:05change now more than anything else and
10:07into the next 20 years he seems like
10:08groceries the interesting thing about
10:10groceries is that it breaks the Amazon
10:12model in fact it breaks the whole
10:13ecommerce model because the whole point
10:14of Amazon is that everything is a
10:15commodity is one packet switching
10:17network and they don't have to know
10:19what's in the box they just ship it to
10:21you they just pull it off the shelf and
10:22ship it to you and so it's a commodity
10:23platform and of course you can't do
10:24groceries like that groceries needs
10:26almost a whole other company you need
10:27logistic different logistics different
10:29delivery different e-commerce platform
10:30and everything else but as the chart
10:31shows pretty clearly our groceries are
10:33so much bigger than e-commerce it's such
10:35a big opportunity that it's worth
10:36building almost a whole other company to
10:38do that for yourself here again of
10:40course the USA is behind this is a chart
10:43of the online grocery market and
10:45fairly unsurprising that China is bigger
10:48than the USA because China is apparently
10:49quite a big place what's more useful I
10:51think is to look at the revenue per
10:52capita so you see other large developed
10:55markets that are way ahead of the US
10:57here and so again you will see really
10:59substantial growth in how people do
11:02groceries in the US a kind of catching
11:03up with some of these other countries
11:04again I think what's more interesting
11:06though than just the war penetration is
11:08to think of how that change
11:09the industry so this chart shows you the
11:11number of items stopped in grocery
11:13stores in supermarkets over the last
11:15couple of decades and what you see here
11:17is that as the retail model changed the
11:19grocery business changed the number of
11:20SKUs there were change the way that you
11:22bought si and the really that the
11:23fundamental thing here is when you
11:25change how you buy things you change
11:26what people buy you change your
11:28purchasing journey you change what goes
11:29into the basket as we change that
11:31purchasing journey again whether that's
11:33kind of conventional ecommerce in any
11:35other category but in particular in
11:36groceries will change what kind of
11:38categories get bought and so that means
11:39money moves around categories change
11:41money moves from advertising to
11:43marketing as I mentioned earlier we
11:45won't just buy all the same stuff but on
11:46the web we'll buy different stuff that
11:48of course is a good way to think about
11:49what happens as we move from just the
11:51logistics part to the taste preference
11:53recommendation curation part now I think
11:56a good way of thinking about this is
11:57that the internet lets you buy anything
11:59that you could buy in New York in fact
12:00that's exactly what Sears Roebuck did
12:02100 years ago it let you buy anything
12:03that you could buy in New York but it
12:04doesn't let you shop the way you can
12:06shop in New York and that's now starting
12:08to change we have a bunch of building
12:09blocks and a bunch of new ideas for how
12:11you might build a consumer internet
12:12business they give you different ways of
12:14addressing these so we have social we
12:16have new model economic models like
12:18rental or subscription we have this sort
12:20of sense that you know actually if
12:21you're an online brand maybe you want to
12:23have a shop and maybe that's a
12:24substitute for spending money on
12:25Instagram and so we have different ways
12:26of requiring customers it feels like an
12:28awful lot of current consumer retail
12:30events are going to shift into
12:31advertising and a lot of advertising is
12:33going to shift into retail events and
12:35then of course we have machine learning
12:36changing the whole conception of what it
12:38is we might be able to recommend and
12:39suggest all of these building blocks
12:41give us things that you really shouldn't
12:43have work or things that would never
12:44have worked 20 years ago the idea that
12:46anybody would buy fashion online was
12:48about as insane as the idea that Bitcoin
12:50would replace fiat currency it was just
12:52kind of an obviously insane idea Warby
12:54Parker selling glasses online you're out
12:56of your mind cosmetics makeup that you
12:58haven't tried these are all kind of
13:00obviously stupid ideas these are all
13:02things that are now starting to work as
13:03we have these new building blocks as we
13:05have more penetration as consumers are
13:06willing to do effectively everything
13:09now this gets me back to the phrasing I
13:12used earlier as we think about new
13:13problems actually I think you're the way
13:15you could do this chart is what we did
13:17in the first twenty years was e-commerce
13:19and advertising what we do in the next
13:20twenty years actually is just everything
13:22what kind of everything else
13:23well what a stable GDP look like
13:25actually you can tell I don't like log
13:28scale charts pretty much so what does
13:29global GDP look like what pieces of that
13:31might need won't we address well how
13:33about cars Tesla this kind of little
13:35Californian company has been doing quite
13:36well they clean you may have read about
13:37this in fact Tesla is now outselling the
13:40other premium car manufacturers in the
13:41USA but if we actually think about the
13:44total opportunity for cars for software
13:47for electric then again we get another
13:49of these charts that really ought to be
13:51a log scale chart but actually tells you
13:53a clearer story if it isn't what is a
13:54total opportunity here and again this is
13:56just see USA not globally which is even
13:58bigger more interesting though than just
14:00looking at numbers of cars I think is
14:02looking at what's inside the cars so
14:04this chart shows you the component cost
14:06of a conventional car on the right which
14:08is the OEMs and the tier ones and then
14:10how much companies from the electronics
14:12industry are going to eat that because
14:14as we move to electric as we move to
14:16batteries and then again as we move to
14:17autonomy you have a completely different
14:19mix of components and mix of suppliers
14:20inside the cart software eats the car
14:22more interesting again though software
14:23changes what the car is so as we go to
14:25autonomy we change what we think a car
14:28is it won't just look like today's car
14:29but without a steering wheel it will
14:30look different it will do different
14:32things it will go in different places
14:33will unbundle the car journey entirely
14:35into different sorts of vehicles
14:36different modalities what are you going
14:38to be doing inside that car if you're
14:40not staring at the wheel well you should
14:41probably invest in alcohol companies in
14:43TV and TV is another interesting to
14:45think about here software each TV well
14:47Netflix and Amazon are now top-tier
14:50content producers they have equivalent
14:51budgets to the other major US
14:52entertainment companies in fact this is
14:54a great quote from back in 2010 that
14:57Jeff books at Time Warner said this was
14:59like the Albanian aren't being worried
15:00about the Albanian army I hope somebody
15:02is keeping an eye on the Albanian army
15:03now I think more interesting though is
15:06to ask what do we actually mean by TV so
15:09watching people play games is now bigger
15:12than Netflix or bigger than HBO about
15:14bigger than ESPN the latest um triple-a
15:17console game had a big opening weekend
15:19than any of the recent Hollywood
15:20franchises watching people play games is
15:23a big deal eSports is now an equivalent
15:25size to any of the local US sports not
15:28equivalent to any of the global sports
15:30but the US ones will do for now but then
15:32if you look at the revenue here this has
15:34yet to catch up and so that's another
15:35place where you'd expect to see
15:37an awful lot of kind of interesting
15:38Grayson is interesting businesses being
15:40created going up to the high level TV
15:43viewing is going down internet
15:44consumption people looking at their
15:46phones is going up tech industry spent
15:4820 years trying to get into the living
15:49room and it did get into the living room
15:51but it didn't do it with TVs it did it
15:52with mobile phones and so we have this
15:54fundamental shift in where the attention
15:56is and what kind of things you're going
15:58to be doing with that I've talked a lot
16:00software weeks money is kind of an
16:02interesting topic to talk we're hearing
16:03from Angela later on today talking about
16:05this we had this this is interesting
16:07idea that all the money being put into
16:08FinTech is is enormous and that's
16:10clearly too much in the market won't be
16:12able to absorb this and yes there's a
16:14lot of money going in it's just kind of
16:15starting to address some of the most
16:17obvious pain points but again if we
16:19think about the overall opportunity
16:20there are trillion dollar pools of
16:21capital here that are going to be
16:23addressed by software and going to be
16:24addressed by technology in sterling it's
16:26worth adding housing from an earlier
16:27side onto this just to get a sense of
16:28them how much money there is in money
16:30but then as we look at this as an awful
16:32lot of people still untouched no matter
16:34how big those segments are portion of
16:36the u.s. population is unbanked much
16:38larger population of the global
16:39population is unbanked and those people
16:41will not be addressed by giving them all
16:42a paper check and a credit card they'll
16:44be addressed in totally different ways
16:45with software with the internet with
16:46mobile and in the u.s. incidentally
16:48there's a lot of people who are sort of
16:49notionally banked but actually aren't
16:51banked at all there's a common saying
16:52here that the u.s. basically has two
16:54banking systems as a banking system for
16:55the people in this woman there's a
16:56banking system for the other people and
16:58they get a much worse deal and that
16:59again is something that a lot of people
17:01are going to think about changing in the
17:02next 10 and 20 years time as we look at
17:04how some of those tools again the USA is
17:06not in the lead so on the Left we have
17:08mobile penetrant penetration on the
17:10right we have mobile payment value it's
17:12fairly obvious that something
17:13interesting is happening in China here
17:14again and that's something again that's
17:16that's going to start spreading around
17:17the world so we have here a bunch of
17:19building blocks a bunch of new ways of
17:21building FinTech new ways of building
17:23financial services applications we have
17:25vastly more data we have completely
17:26different kinds of data we have totally
17:28different consumer expectations of what
17:30it is you ought to be able to do and
17:31what a service ought to look like we
17:33have the capability to unbundle things
17:35from the big traditional conglomerates
17:36and aggregators like retail banks and
17:39again we have machine learning is a new
17:40way of bringing understanding to that
17:41information I'm talking about money the
17:44only thing left is to talk about death
17:46we have a death fund sorry a bio Fund
17:48and drug discovery is 75 billion dollars
17:51feels like a big market it's an
17:52interesting market is one that venture
17:54capitalists have addressed for a long
17:55time but it's now being addressed in new
17:57kinds of ways so machine learning again
17:59a recurrent theme changes discovery but
18:02things like in more interesting things
18:03like social and wearables change trials
18:05if you actually want to get 10,000
18:07people who have a condition social means
18:08a completely different way of getting
18:10these people wearables mean a completely
18:11different way of monitoring those trials
18:13and of course now we now now have ways
18:14of actually programming genes and
18:16programing cells which my colleague
18:17Holger will talk about later
18:18that means we can do totally different
18:20kinds of cures but the 75 billion dollar
18:22number but then there's other numbers to
18:24think about so this drug discovery and
18:25then there's a drug market and then
18:27there's health insurance but then
18:28there's a global market for health care
18:30what actually gets spent spent on you
18:32when you're sick and that's seven over
18:34seven trillion dollars how do we address
18:36some of those well seven trillion
18:38dollars is spent on health care but
18:39that's kind of the cost of being sick if
18:41I did any part of the cost of being sick
18:42but it's more interesting you think well
18:44what's the cost of not being sick what
18:46hat is the value of not getting sick
18:47what's the value of not dying what's the
18:49value of never having to cure any of
18:50these things in the first place and
18:51those are things that we can now start
18:52thinking about addressing a good way of
18:54sort of thinking about what's happening
18:56here is that we've sort of gone through
18:57all of the easy monsters and now we're
18:59getting to the death tell each other to
19:00the god mode we kind of killed the foot
19:02soldiers now we've got the giant robotic
19:03brain with the Gatling gun and so we're
19:05kind of addressing different character
19:07of problem a harder problem harder
19:09markets to address than we addressed in
19:11the past what tools do we have to do
19:12that well we have kind of interesting
19:14new fundamental layers if you think
19:16about what happened in the first twenty
19:18years we went from these kind of blind
19:19alleys of a well in the information
19:21superhighway and interactive TV and then
19:23the Internet and the web came along and
19:25this gave us kind of decentralized
19:26permissionless innovation you didn't
19:28have to ask AT&T or Time Warner for
19:30permission to put up a website and that
19:32was what enabled the explosion then it
19:34turned out we needed some kind of
19:35organizing layer on top of that and we
19:36got Google and we got Facebook and we've
19:38spent the last ten years thinking about
19:39Google and Facebook and how they
19:40organize things and those platforms are
19:42centralized by design and they capture
19:46meaning and intent and their form can
19:48capture value they're also kind of
19:50highly abstracted they don't really know
19:52what a thing is itself they can only
19:53look at it a kind of tools we removes
19:56now we have two new fundamental layers
19:58we have crypto and we have machine
19:59learning which among other things are
20:01swinging away from centralization what
20:03do we mean by that what would those
20:05well what is this well how would Google
20:08or Facebook or Amazon answer that
20:09question what is this thing
20:11well Google will say it's on pages that
20:13say wishbone and Hanna's Wagner Facebook
20:15will say people who like furniture
20:16shared links to this Amazon will say
20:18it's a SKU I don't understand the
20:20question people who bought this Scoob or
20:21other SCU's and not really very much
20:24else what are we going to be able to do
20:26with machine money well we have more in
20:28more levels of meaning and more levels
20:29of understanding so we will know it's a
20:31chair that would be a good step we can
20:32kind of do that now we could never do
20:33that before now we can do that we can
20:35say it's a chair and then we'll say well
20:36it's a modern Scandinavian design
20:38classic but actually you go another
20:39level of meaning in and you say right
20:41well if you like that chair you're not
20:42going to like Disney cruises so we won't
20:43show you Disney cruises ads will show
20:45you something else that's to say we get
20:46to successive levels of meaning of what
20:49a thing is rather than kind of guessing
20:51at it's through multiple levels of
20:53levels of abstraction from who linked to
20:55it or from or page it we shared on this
20:57of course is not just a consumer story
20:58if we go back and think about by oh well
21:00machine learning means I can look at an
21:02x-ray and tell you whether you've got
21:03cancer and do that at massive scale and
21:05massive automation but I can also say it
21:07looks like you're going to have a heart
21:08attack in the next couple of days or I
21:09can say well change your behavior or
21:11you're likely to have a heart attack in
21:13the next day or two that's to say we get
21:14to successive levels of meaning which
21:16successive kinds of value in successive
21:18kind of kinds of addressability
21:19the fundamental framework for thinking
21:21about machine learning it's a tool to
21:23scale people but to scale people in two
21:24very different ways one of them is I
21:26have a million pictures in the basement
21:28and I now have a million interns that I
21:29can send down to look at all of those
21:30that's one kind of machine learning it
21:32automates a very basic human task but
21:34something we couldn't automate before
21:35but the other is I have a million
21:37pictures in the basement now I can send
21:38one intern to go and look at all of them
21:40and come back in half an hour and say
21:41well when I looked at the third hundred
21:43thousand this interesting picture
21:44started coming out and so that's to say
21:46we get a superpower in a sense we get
21:47things we could never see before so we
21:49think about machine learning we also
21:50think about crypto this is a different
21:52kind of crypto why does crypto matter
21:55well imagine in 1993 asking why does the
21:57internet matter first you'd have a long
21:59argument about whether you should say
22:00Internet or internet or web or world
22:02wide webs and then you would say well
22:04this is a distributed decentralized
22:06permissionless network anyone can build
22:08anything you don't need to get
22:10permission from the phone company to
22:11build it so it's a great story that
22:12Steve Jobs said hey have you seen this
22:14internet thing we should buy it it's
22:16country and work like that anymore
22:17changed how this stuff worked
22:18now how do we think about what those
22:20applications were going to be well we
22:21had no idea so this is web service in
22:23ninety and Internet traffic in the early
22:2590s Mac had more share of web service in
22:27Linux and the world wide web was 3% of
22:30these were applications built on that
22:32network these were not the important
22:33applications everything we see now is in
22:35that little purple box at the top right
22:37so you have the network in the you had
22:38the things built on the network you
22:40didn't know what the things were going
22:41to be that were importantly what was
22:42important was the capability the same
22:44thing with cryptocurrency we have a
22:45distributed decentralized network anyone
22:47can build applications with native trust
22:49native value exchange payment and
22:51mechanisms for understanding meaning
22:53built into it so we will come back to
22:55this some this pendulum from closed to
22:58open to close to open we think about
23:01what these things can do we have new
23:03ways of finding meaning in intent new
23:04ways to understand what a thing is and
23:06what it might mean that you're
23:07interested in a thing new ways to build
23:09networks on that building networks on
23:11that that a decentralized sort of
23:12permissionless and also of course a
23:13market reset whenever you have a
23:14fundamental change in technology that
23:16just kind of shakes the trees and brings
23:17new companies out of that so a final
23:19thing to think about here I called this
23:21presentation the end of the beginning
23:22this is what the beginning looked like
23:23this adorable young man is called
23:25Geoffrey Bezos and he's just entered the
23:27books business this is Geoffrey bezels
23:30today he is not in the books business
23:32anymore and this is Geoffrey bellows in
23:35the next 20 years and what other things
23:37he might be building so thank you