00:00welcome to the a 16c podcast I'm
00:02Benedict Evans and we have Tim O'Reilly
00:04on today here to talk about his new book
00:06among other things so your book is
00:08called WTF and it kind of it follows an
00:11arc talking about platforms in general
00:14but then instead of flowing through to
00:16kind of the broader impact have had
00:18these companies interact with us all and
00:19with the economy WTF can be an
00:22expression of delighted amazement or an
00:26expression of horror and when we're
00:30faced with technology today we really
00:31are hearing both the arc of the book is
00:33really it's really a tour through 35 40
00:38years in the technology industry looking
00:40at some of the great platforms and where
00:42they've gone wrong where they've gone
00:43right and what they do and in the course
00:47of that I look at some of the latest
00:49platforms they spend a lot of time on
00:50uber and lyft and on Airbnb kind of
00:53talking about how they have to make both
00:55parts of the market come together a lot
00:57of people don't understand for example
00:59of the need to make a thick market in a
01:02new city is is so this driving factor
01:04behind the high cost of these companies
01:06once that happens their economics are
01:08gonna be very different because they've
01:10got to recruit enough drivers they've
01:11got to get past they've got a little
01:13masculine base on critical mass on both
01:15sides and the thing that's interesting
01:16though is there's a lot when you
01:17understand the implication something
01:20like that you can really start to tease
01:21apart what people get wrong about
01:24company business models for example taxi
01:27companies you know traditional taxi
01:28companies think they thought oh it's
01:30this app if we have the app it will work
01:32you know we will be will be on you know
01:34we'll be able to compete but the app is
01:37only a tiny part of the business model
01:39because the fact that lyft can deliver
01:41three-minute pick up times is because
01:43all these part-time drivers showing up
01:46in their own car so that supply
01:48automatically Rises as there's more
01:51demand and it just doesn't work with
01:53their their business model which has a
01:54fixed number of cars which is sort of
01:57optimized for sort of middle of the day
01:59or whatever you know it's optimized for
02:00a certain level of service and you can't
02:03get there unless you take on all parts
02:06of the business model you'll see in some
02:07news stories fatuous statements like you
02:10know if if you know uber or lyft had a
02:13with vehicles they would get rid of all
02:15those pesky driver costs and the cars
02:18would be fully utilized and ago those
02:19two things don't go together
02:21yeah because if you own all the cars
02:23they're gonna be lightly utilized at the
02:25times when there's not lots of
02:27passengers I mean the one that really
02:28annoys me actually is when people say
02:30cars and aren't used for 98% of the day
02:32and therefore you could fill all of that
02:34up and you think well yeah okay so what
02:36you think there's going to be between 1
02:38and 5 o'clock in the morning then with
02:40this new app that you're going to deploy
02:41first I actually have to think about
02:43well what does rush hour utilization
02:45look like and rush vitalization is what
02:46like 80% 70% 90% whatever it is did you
02:50actually have to have enough cars for
02:51peak yeah not for those idle times I try
02:54to walk through an understanding of
02:56algorithmic systems because they're so
02:58central to platforms today and I try to
03:02tease apart this idea that every
03:06algorithm has something that's trying to
03:07optimize your Google optimizes for
03:09relevance Facebook optimizes for a kind
03:12of engagement and through the fake news
03:14controversy you can see how these
03:15algorithms can go wrong and she talked
03:17about why and praises automation or
03:20mechanization as a way of breaking the
03:23kind of the kind of the blinkers that
03:25people get when you say AI and suddenly
03:27people's brain shut down that it really
03:28place it that's right analyze well what
03:30are we actually building here and I
03:31think this is sort of a similar point
03:32here when we say algorithm people's
03:34brain shut down yeah and it's kind of
03:36one maybe once you just say like
03:38automation and I think this is
03:39particularly the thing you see with like
03:41you know this sort of you know the
03:42recent problems with Facebook
03:43advertising or Google advertising and so
03:46on or we're sort of extremist content on
03:49YouTube and people just say well just
03:51use the algorithm mm-hmm and it's as
03:55though you've kind of said robbed the
03:56magic lamp three times in the genie
03:58appears more understanding will lead to
04:00better better policy and better business
04:02decisions why do these platforms go
04:05wrong and start competing with their
04:07ecosystem one of the things that I was
04:08seeing was a repetition of patterns that
04:12I first observed in the you know the
04:14antitrust case against IBM and then
04:17Microsoft which is that companies that
04:21really are platform technologies that
04:23enable a whole ecosystem decide at some
04:27to start competing with their ecosystem
04:29because it's just not enough to go
04:31around and at that point they screw up
04:35the ecosystem starts to fall apart
04:37government turns against them people
04:39turn against them it's so amazing to see
04:42even idealistic companies forget that an
04:44ecosystem has to work for all of its
04:47participants you can't just think about
04:49yourself you can't just think about
04:50users you actually have to think about
04:52your suppliers as well and if you start
04:54competing with them eventually basically
04:57the the entrepreneurs who are making the
04:59the services that your ecosystem depends
05:01on go look for one that's more friendly
05:04it's interesting give them the google
05:06HTC deal number they spend a bit over a
05:08billion dollars to buy something it's
05:10not entirely clear what out of HUC it
05:12looks like they got some portion of
05:15HTC's phone engineers and some
05:17non-exclusive rights on IP and now you
05:20can see a parallel conversation with
05:22Microsoft in the surface that's right
05:24how far does Google go in competing with
05:27Android OEMs that you can see this like
05:29a clear strategic problem for them and
05:30that Apple has the high end and Google
05:32doesn't and there's reasons why Google
05:34is a reach business wants to be on those
05:35devices directly rather than just going
05:37through their Apple to get to those
05:39people but then if they go and build a
05:42business that takes on the high end of
05:44Android well that's where all the OMS
05:46make their profits so what does that
05:47look like you kind of see these
05:49interesting kind of tensions and
05:51dynamics that like this the high level
05:52strategy that we need to do a B and C
05:54but then you also have like all the
05:56mid-level people that's like the guy
05:57whose job it is to run that particular
05:59thing well of course that person's you
06:02know he's going to try and do X and y in
06:04order to grow their business and
06:05increase the margins and increase the
06:06profit because that's kind of the logic
06:08of the organization but sometimes you
06:10get all of these things these these
06:12things don't end up getting getting to
06:13be where you would want for the health
06:14of the border ecosystem right and you
06:16see how often you know great company
06:21decisions that end up creating enormous
06:23value don't follow that logic of profit
06:26maximization you know I still have think
06:28back to the origins of Google where they
06:30were you know we were living in the
06:32world of Alta Vista with literally
06:34blaring ads you know blinking at you on
06:38and Larry and Sergey were like we hate
06:40that we're not gonna do that and it was
06:43a moral rejection of that business model
06:47that eventually led them down the path
06:49of we want ads to be as effective and as
06:53useful as search results and that was
06:56the beginning of Google's dominance so
06:58you can have those strategic goals which
07:01are you know I think Google's in a hard
07:04place with this you know high-end
07:06low-end ecosystem play versus do we do
07:09it for ourselves but sometimes you
07:12actually have to make choices on behalf
07:15of the ecosystem because it's a
07:16longer-term smart choice mmm exactly and
07:19you have to work out the difference
07:20between the tactics and you know the
07:23platoons through advancing because
07:25there's territory in front of them so
07:26they'll advance and where you actually
07:28want the divisions to go and where you
07:29actually want the army to go yeah it
07:32seems to me that there are also really
07:34interesting questions about
07:36understanding what your real business
07:38model is and so for example this is
07:41shifting to a different competitive
07:44front that Google is facing you know the
07:47battle with Facebook it seems to me to
07:49the extent that Google says well we're
07:51gonna deliver this service directly with
07:53native content with Google native
07:55content versus bringing it from an
07:57ecosystem of outside providers they're
08:00losing touch with one of their key
08:02strategic advantages against Facebook
08:05where the logic of this is better for
08:06the user led them for example to much
08:09more to compete directly with Yelp and
08:11OSU visor people like that and you say
08:14well that actually you know it is often
08:16better for the user yeah I like having
08:18my weather results directly in in Google
08:21I like having my stock quotes directly
08:23in Google you know and we're but that
08:25logic ends up where again just as with
08:28the Android case you gave up where
08:31you've ended up taken more and more of
08:33the valuable properties and you end up
08:36becoming a you know a destination
08:38platform whereas in fact Google's
08:41original strategic strength was as a
08:43switchboard mmm it's interesting I've
08:45had sort of variants of this
08:46conversation with Steven Sinofsky
08:48on here he used to work most ran office
08:51and then ran Windows
08:52and this of course there's a the
08:55inexorable logic of the organization but
08:58you need to sort of balance well are you
08:59pushing back against that or is it just
09:01something that this is clearly what that
09:03product was always going to be like you
09:04know an X or Excel was always going to
09:06have charts your example yeah
09:09you know and I think that's absolutely
09:10true initiative historical analogy is
09:13that a sort of empires always tended to
09:16grow so this mechanism that they would
09:17kind of subdue the rebellious tribe on
09:19the for the Troublesome tribe on the
09:21frontier that was raiding over the
09:22frontier and killing their farmers and
09:23so you go out and you kind of subdue
09:25this tribe and then they get kind of
09:26civilized and settled and you build some
09:28fortifications and guess what there's
09:30another you know violent violent
09:32aggressive drive over the next mountain
09:35range so okay so nothing job on that
09:40though there you know the Romans
09:42basically when they conquered a tribe
09:44they made them citizens yes they gave
09:46them a path to citizenship but the
09:47problem is in the tribe but it is so
09:52true though that if you consider ways to
09:55bring people into the value that you're
09:57creating that actually is a really
10:00interesting way and obviously that's
10:01what Silicon Valley Giants do when they
10:04acquire companies as opposed to you know
10:05killing them well this is the line this
10:07is the two ways to make money in
10:08software is bundling and unbundling yeah
10:10um I mean we're kind of at an
10:11interesting breakpoint at the moment
10:13years ago clearly the thing to talk
10:15about with mobile and mobile is talking
10:17about mobile now is a bit like talking
10:19about pcs in 2000 that you sort of you
10:22know the ecosystem always done and we
10:23know what happened and we know like the
10:25primary question and the question is now
10:26what are we build on top of that yeah
10:28you know there's another thing I would
10:29say in that's different than what do we
10:32build on top of it which is what are the
10:36capabilities of this device that we have
10:38not realized yet you know you know when
10:40you think about uber and lyft they were
10:43a realization of capabilities that were
10:45there in the smartphone but you said we
10:47find new white space yeah you know and
10:50that that that you know just like how
10:52could we not see you know there were
10:54connected taxi cabs and they had a
10:56screen in the back to show ads they had
10:58a connected credit card reader but
11:00nobody put her all together oh wait we
11:02did some in the car we I can show you I
11:04research reports from 99 to 2000 when I
11:06was a telecoms analyst where people made
11:07all these lists of all these things that
11:09were going to happen and everything
11:10falls had an M at the beginning so M
11:12back he was at M banking and M this and
11:14M that and M the other and no one had
11:15like m taxis you know there wasn't it
11:17wasn't anybody's list um I mean it's
11:20interesting you say that though it's
11:21like so there's one the conversation I
11:22was sort of getting at is that you get
11:25the platform and then there's white
11:26space on top of the platform you built
11:28the but to your point actually also
11:30sometimes you discover whole a whole new
11:31piece of white space and I think you
11:33could argue that's happening now with AR
11:34on smartphones there so that's like a
11:36thing that wasn't there before now
11:38suddenly that sort of become possible
11:40right and there's this new combinatorial
11:42innovation you know that comes from
11:44multiple inventions this is almost like
11:47somebody's just dumping out pieces of a
11:50puzzle onto the table yeah you couldn't
11:52solve the puzzle so somebody brought the
11:54rest of the pieces one of the ways I
11:55described what's sort of happening in
11:56hardware components are so much smaller
11:58and cheaper more power efficient than
12:00their components you've had in pcs um
12:02it's like someone took your shipping
12:03container full of Lego and dumped it out
12:04onto the table or dumped it out onto
12:06Sand Hill Road and everyone's kind of
12:07picking up the bits of Legos well what
12:09would we do with these things and now I
12:10think if one was looking at new curves
12:12one would be looking at cars which are
12:15not something that appends the computing
12:17ecosystem but they're of kind of
12:18enormous and radical change in the world
12:20as we get autonomous cars and then
12:23machine learning which is a sort of new
12:25foundational technology um but it's not
12:28like it's a new platform machine
12:30learning is like this new foundational
12:33technology that in a sense will enable a
12:35bunch of new stuff and new companies but
12:37it makes Google better Google hmm and it
12:39makes Facebook stronger Facebook and
12:41Amazon may be stronger Amazon it doesn't
12:43unlock it doesn't it doesn't it's not a
12:44shift that the changes the balance of
12:46power that's an interesting perspective
12:48although it is certainly possible that
12:51we're just too early to see the way that
12:54one of those players may well turn in
12:57their particular machine learning
12:59platform into the open platform in the
13:01way that that Amazon turned you know
13:03effectively cloud computing into a
13:05platform when other people could have
13:08done it but weren't really ready to one
13:10of the things that happened was it was
13:12sort of non-obvious it happened when we
13:13went to smartphones was that on the
13:16desktop it was really a pain
13:18to have more than one social network
13:19you'd have to open another browser tab
13:21go to another website you like you get
13:23email notifications or something
13:25and so you wanted to have everything on
13:26one site and you had this nice big
13:28screen so they could just keep adding
13:29features and they could kill any
13:30competitor by adding another tab whereas
13:32you go to mobile and the screen is much
13:34smaller and you have a home button so
13:36it's actually easier to get to another
13:37feature by tapping home and going to
13:38another app that it is to navigate and
13:40inside that app yeah and every app has
13:42your address book and so you get all of
13:44those things kind of come together Amen
13:46is really easy it means even now it's
13:47really easy to have like fo five or six
13:49social apps on your phone it was like
13:50there's a phone app nobody's using all
13:53of that it's what everyone is using like
13:54two or three of them and that kind of
13:57rolled out in like two thousand eight
13:59nine ten and that gave us Instagram and
14:02whatsapp and like a hundred other
14:03smaller ones and Facebook bought the to
14:06that one but they might you know that
14:07wasn't like inherent in the model in
14:09create the model created it like an
14:11inherently plural model and Facebook was
14:13able to adapt to that yeah but they
14:15might not have done I think a lot about
14:16the know screen model which i think is
14:20an important part of our future you know
14:22amazon showed it with the echo I think
14:25the reason that they got it right where
14:27other people weren't was they they
14:28basically built a device that literally
14:31assumed there was no screen whereas you
14:33know in the first generation of Google
14:35Apps and even now sometimes you know you
14:37asked for certain functions and it just
14:39opens the smartphone app and that's
14:41backwards from just building it in I was
14:43kind of calling this like frictionless
14:45computing yeah you don't have to turn it
14:47on it's not even the screen I mean I
14:48think there's a lot of overlap for me
14:50between the Amazon echo and the under
14:52SmartWatch yeah because they don't do
14:53anything that you couldn't do on your
14:54phone but they do it with less friction
14:56verse exactly if you remember the thing
14:58that you can do that's right and the
15:00other thing that's gonna be really
15:01interesting is who creates the
15:02horizontal platform across multiple
15:04devices and when you talk about what's
15:06that next platform I think it may be in
15:08the level of what is the software that's
15:11running in your car and in your watch
15:14and in your you know home speaker and
15:16then your phone ended your computer and
15:19I can just talk to this agent for
15:22example or I can interact with it in a
15:24variety of ways wherever I am on
15:26whatever device I am and it's
15:29a continuous experience I feel like that
15:31might be the second step I think like
15:33the immediate step is you buy is you get
15:36Venn diagrams yeah so while the light
15:39you know electrifying your home so you
15:41get a washing machine and a dishwasher
15:42and the common standard is AC power
15:44that's it but you didn't like go to the
15:46store and say why I'm not going to
15:48you bought a washing machine and the
15:50same like in Warren Ohio t that you'll
15:53buy a connected TV and you'll buy
15:54connected speaker and maybe you'll have
15:57a connected door lock and a connected
15:58alarm and like the door lock will talk
16:00to the alarm but it won't talk to the TV
16:01and the TV will talk to the speaker but
16:03it won't turn eyes none of them will
16:04talk to the smart meter yeah I remember
16:07actually seeing a Jeff Bezos in 2003
16:09gave a great talks and it was the
16:11history of electricity basically there
16:13was there were lights and people were
16:16plugging in all kinds of new electric
16:18motor motorized devices in just massive
16:20nests of wires coming down from that
16:22from the light socket and you know
16:25people would get caught in there you
16:26know automatic your laundry rain machine
16:28and have their hair ripped out and it's
16:32just so just all the things that had to
16:34come together for electricity to
16:36actually work mmm there's a lot like all
16:38these incompatible devices so the device
16:40is sure they did come in but it wasn't
16:42till we got standardized power and we
16:45got standardized interfaces for
16:47interacting with that power that's the
16:50question is well what how far back does
16:52the standardization need to go it may be
16:54the standardization is Wi-Fi and or
16:57low-power radio and AC power DC power
17:01and it's not necessarily that they all
17:03know Isis better I suspect that it's
17:06going to be the ability to understand
17:08commands in some common way yeah it
17:13could be I don't know I've had the like
17:15I've had the demo of the full connect at
17:17home and you know I devices and I like
17:21buying this stuff and I like removing
17:22friction and I you go there and you walk
17:24in and you say no hey Google hammers and
17:27hey Siri your movie night tell or turn
17:29the lights on or lights on in the
17:31kitchen and it just happens and I kind
17:33of look at this and I think I can like
17:35shut my eyes and imagine I want this
17:38and I can sort of write the script that
17:41says yes there's like the easy answer is
17:42like come on it dum-dum alsina just
17:45press the sweat light switch but it
17:46actually is not one light switches ten
17:48light switch is because there's a side
17:49lights and there's low although it's
17:50actually is not it is easier than just
17:52pressing the switch but then I just rent
17:54not got a new apartment and like am I
17:56going to go and buy all these plugs and
17:58connected light bulbs too well but
18:00that's not really all that kind of
18:02connected home vision isn't the way I
18:04think about it I think it's more that
18:06just the way it's building up gradually
18:11that you know my calendar shows up in my
18:14car you know and I can select something
18:16from my calendar to get directions in
18:18the map I can talk to my phone or I can
18:22talk to my car and they have a common
18:24data substrate that that knows about me
18:26that's centered on me and and that is
18:29starting to be built you know and one of
18:31the big reasons you know like I have an
18:33iPhone and I have a an Android phone and
18:37I tend to live in Android because I have
18:39a suite of apps with my data you know
18:42there is this gravitational attraction
18:43of who knows the most about me and who
18:47has more of my data who has built
18:49services that I already rely on there's
18:51an interesting kind of conversation
18:53around kind of privacy and permission
18:54there um I mean I feel like it's as much
18:58I mean it but it's as much to do with
19:00well how you think about that company
19:03and what that companies well is as what
19:05that company wants to do you accept that
19:06your bank knows how much money you've
19:07got because that's kind of how it works
19:09and you know you accept that your phone
19:11company knows where your phone is
19:12because that's how it works
19:14and I think there's a lock in is feels
19:16like there's an interesting distinction
19:17here between Google and Facebook that
19:19people accept that Google knows a
19:20certain amount because that's kind of
19:21how Google works whereas people just
19:24presume that anything they put on
19:25Facebook is public and don't trust any
19:26assurances about ever seek if they just
19:28assume that they can't believe it and
19:30it's not reliable mm-hmm and so people
19:32kind of get this sense of what is in
19:34place or the permission that that
19:35company has and so there they thought
19:39you can like one could like imagine the
19:40same product from two different
19:41companies where for one and that would
19:43be entirely normal and okay and no one
19:45would be waste an eyelid and where it
19:47raised an eyebrow and for another it
19:48would be completely nobody so I think I
19:50mean I think Apple the face ID that
19:51Apple announced actually
19:52be a great example of that you know
19:54Apple has spent enough time like talking
19:56consistently incredibly about we don't
19:58want your data that they say well care
20:00takes feature it models your phone your
20:02face and that stored on the phone and we
20:03can't see and we can't give it to anyone
20:04it's just your phone unlocking your
20:05phone and if Google or Facebook had
20:08announced that we'd be having a
20:09different completely different
20:10conversation well I had that
20:11conversation with Tony Fadell about why
20:14Google didn't do something like the echo
20:17and he said could you imagine if it were
20:18Google who would first come out with a
20:20device that was always listening the
20:22blow back would have been enormous and
20:24in fact that it was Amazon that didn't
20:26have that Big Brother story already they
20:30were a retailer you know an e-commerce
20:32site it and found a computing site but
20:35they weren't kind of this omnipresent
20:37Big Brother you narrative about them it
20:40it gave them more room to innovate in
20:43that yeah I use the analogy of you know
20:47the thinking of these systems as really
20:51thinking of the corporation itself as a
20:52kind of artificial intelligence and just
20:55as Google has an objective function
20:58we've actually told our corporations
21:00that they should optimize for share
21:03price and you know well that for
21:07economic return and of course the theory
21:09behind that was that it would be good
21:11for society but just as we've sort of
21:13seen with problems with you know spam or
21:16you know fake news these algorithmic
21:19systems don't always turn out the way we
21:21thought and we then go back and we have
21:23to try to debug them and I think you
21:25know the point that I kind of come to in
21:27the book is we have to actually come
21:29back and effectively debug are the rules
21:33by which we're running our society
21:34because just like these platforms it is
21:37not an unfettered you know natural free
21:41market it's a constructed thing it's
21:43kind of interesting I mean you know the
21:44old joke that you know communism is
21:47never failed because the Communist will
21:48always say it's never been tried
21:49properly and I always feel like the sort
21:51of free market system is kind of the
21:52reverse in that you know almost
21:54everybody says ok you free market needs
21:56regulation and government intervention
21:58and the arguments are about how much and
21:59so you need lots and you need
22:01anti-monopoly but you don't just need
22:03the kind of the Nightwatchman state but
22:06do you know health insurance and you
22:07need food where health and safety
22:11regulations and so on and so there's a
22:13list of the arguments to sort of about
22:14where you put that my point is you have
22:16this algorithm if you like you have this
22:18mechanism of the market price system and
22:20then you you put your fingers on the
22:22scale every now and then you adjust it
22:23you know conscious that when you do that
22:25you're going to have unintended
22:25consequences but you still feel like a
22:27there here are these cases where we're
22:29going to make a decision that we're
22:30going to adjust the scales and you could
22:33argue certainly that people look at
22:34these sort of fully automated systems
22:36like Google search or Facebook and so on
22:38and say well it's just fully automatic
22:39and you just have to be fully automatic
22:41and say don't touch it say well yes it
22:43is fully automatic in it has to be fully
22:44automatic we don't have price controls
22:45you know we don't have central planning
22:47but that doesn't mean you don't change
22:48stuff yeah exactly and I think that
22:50that's why I think we need to change the
22:53metaphor that shapes our thinking
22:56economists are kind of moving away from
22:58this consensus but for many years there
23:00was this this physics adulation here
23:03where economics was going to be you know
23:05a hard science and I really think it's
23:08much more like game design you know I
23:10mean yes there are rules of physics in
23:13say a game like basketball you know you
23:15can't get the ball through a hoop if
23:17it's too small you know you can only
23:18humans can only throw the ball so far we
23:21have this incredible opportunity to
23:23rethink the rules of our game or to play
23:27it in different ways with these
23:29unexpected moves I kind of wonder can
23:32you know AI you know with all the dark
23:36prognostications from people like Elon
23:38there's a wonderful prognostication that
23:41and I could help us see ourselves more
23:43clearly that you know that if you can as
23:46a thought experiment imagine you know a
23:48far far more powerful AI looking at some
23:51of our economic problems and coming up
23:53with creative solutions that we haven't
23:55tried yet for adjusting the rules of our
23:57game in a sense what go did was do the
24:02massive permissionless experimentation
24:05of moves that the free market does
24:08that's right you know free market you
24:10know you want to start a business
24:11selling flavored water you want to start
24:13a business selling cookies you want to
24:15start a business doing that you can do
24:16it it might work it might not is one
24:17creative it set of creativity in one set
24:20parameters that's right now Joe brings
24:21kind of massive radical experimentation
24:24which is sort of the point of a
24:26free-market economy that you get that
24:27radical experimentation and of course
24:30it's also the point of the kind of the
24:31startup ecosystem in Silicon Valley that
24:32you have experimentation by the hundred
24:34experiments by startup instead of two
24:36experiments inside a big company all of
24:37these things are goal seeking behavior
24:39and there was a time when the goal of a
24:44start-up was to build a real business
24:45you know we think about you know the
24:48founders of Silicon Valley they they
24:50didn't have any idea that they were
24:51going to get acquired they had the idea
24:54that they were going to build a business
24:55doing this new thing and now you you
24:59have an awful lot of companies where the
25:01goal from day one is the exit and where
25:04the behavior is targeted towards the
25:06exit so that makes them to me financial
25:09products rather than companies yes that
25:11happens and you get waves of tourists
25:13and there have been waves of tourists in
25:15the past you know three MBAs who did are
25:19Harvey balls matrix to work out what
25:20company was good what prospect was going
25:22to have the highest and quickest
25:23acquisition turn and come in and say
25:26well we're going to sell it in two years
25:27to Google and once you funders will hire
25:29an engineer to build it we had a wave of
25:31that around mobile a few years ago and
25:33we've had a wave of that now around
25:35machine learning as I've sort of a war
25:37for talent and to some extent around
25:39cars as well um but I think that's kind
25:43of froth I mean the order for now is not
25:45gonna come in and say and if this all
25:46goes well I'll sell it to Google in two
25:49years whether they say it or not I think
25:51we're you know just in a period where
25:53there's more froth than beer you could
25:56certainly argue I think that Google
25:59Facebook are the big platform companies
26:02now because they're 10 times or 20 times
26:05larger than such companies were in
26:07previous cycles employ vastly more
26:10people and can have pay them much much
26:13more and give them much much much more
26:15freedom so the drive for those people to
26:19leave and do their own thing is smaller
26:22your incentive to leave to quit and live
26:25on ramen noodles and go and swing for
26:28the lottery ticket to mix my metaphors
26:29it's different so that which it might
26:32have been if you're at Microsoft
26:33IBM or at Cisco or so on in the past
26:36this is a question around you know how
26:37these companies evolve the vigor and
26:40aggression and self-awareness of how
26:43Google Apple Facebook Amazon of one now
26:45yeah are kind of different from some of
26:47these leading companies in the past so I
26:49statistics and calls them like hyper or
26:51super evolved organisms like they've all
26:53read the books they will read cloquet
26:54Christensen they all saw what happened
26:56to IBM and Microsoft and MySpace and
26:58Twitter and they all except for Apple
27:00the founders control company and so you
27:03could argue actually it's not that the
27:05problem is not so much people making
27:07stuff because they want to get hired
27:08it's a fact that Google just go up in
27:10vacuum everybody in belief that's a
27:13really good point and the other thing
27:14that's really different today if you
27:16think about the days when Microsoft was
27:18so dominant there was really one yeah
27:21now this for four or five or six that
27:24are serious competitors and well okay so
27:29you kind of go for this blitzscaling
27:31where you're going to get the scale how
27:32many companies are gonna win and so then
27:35you know if the obligations are in a win
27:38be acquired or die you know a lot more
27:42people are going to die and then are
27:45acquired and a lot more people are going
27:46to acquired then then win but isn't that
27:49just kind of saying do great do okay
27:51fail and if you do okay okay maybe you
27:53know maybe someone buys you but if the
27:57platform's take more and more of the
28:00capability the possibility it seems to
28:02me that they have an obligation just for
28:05the sake of the economy and for the sake
28:07of the ecosystem that they depend on to
28:11actually think about how do they create
28:13opportunity for an ecosystem I think if
28:16this is a profound metaphor for our
28:18overall society because we have built an
28:21economy where a small number of people
28:23own and control platforms and many of
28:26them are using those platforms
28:27extractive lis and competing with the
28:29ecosystem as a whole that's why we see
28:31this vast income inequality why we see
28:34you know the hollowing out of the
28:36economy I don't think it's technology
28:38you told Pullman to be less greedy and
28:40look what look at Microsoft now nodes
28:42they've rediscovered their soul perhaps
28:45they are not sonido dominant
28:47tech industry you could argue about
28:48their relevance to the future tech
28:50industries and ISM there's a
28:51machine-learning story there that will
28:53play out but you know the wheel of
28:55fortune turned and then we have 14 turn
28:57for the wheel or the cycle turn for IBM
28:59cycle turn for Microsoft it's not
29:02apparent how that cycle would turn for
29:04Google or Apple or Facebook or MSN but
29:05it will turn oh absolutely
29:07in one sense there's a good creative
29:09destruction aspect of that story but
29:12when I look around at and I think
29:16particular target at financial markets
29:18as a platform for our society because
29:20they too are a platform technology you
29:22know this is the lifeblood of of how
29:24businesses get funded how they grow and
29:29increasingly in financial markets we see
29:33a distortion of that original function
29:34you know Apple for example here's Apple
29:37saying well we can't afford to make
29:39these phones in the United States why
29:41aren't they saying what we actually
29:42could in fact afford to make these
29:45products in America where they cost more
29:46and we take money out of that pot that
29:49the shareholders are claiming and we say
29:50we're going to give it to society and
29:52the form of investment in our country
29:54there's a bunch of issues without one of
29:56them is of course that they cluster for
29:58making these devices in isn't chance and
30:00it's not that you go there because you
30:01get cheap labor it's an awful lot more
30:02than that yeah that's better that's
30:04where everything has to be in a sense
30:06because that's where it gets done yeah
30:08there's a sort of second but that was a
30:11decision that was made on the basis on
30:14the basis of the reason we have to do
30:16that we were going to outsource this
30:17because we're trying to optimize for
30:19shareholders and in fact that
30:21shareholder optimization actually was
30:23bad for a lot of other constituents yeah
30:27so there's a reason why it will went
30:30China in the first place of course the
30:31fact that it went to a contract
30:33manufacturing model powder of on the
30:34other hand enabled vastly more
30:36innovation and more see more people to
30:37make stuff if there's like a more
30:39general macroeconomic pushback to this
30:40which is that uses comparative advantage
30:43that China gets rich and we get rich I
30:46think that's absolutely true
30:49although I I will say that the notion
30:52that the only constituency that we have
30:54to take care of as shareholders I'm not
30:57arguing for protectionism per se I am I
31:00though for strategic thinking about what
31:05are all the things we're trying to
31:06optimize for the situation is way more
31:09complex than this sort of free-market
31:12narrative that we normally give and it
31:16seems to me it smacks of justification
31:19Cory Doctorow had a wonderful tweet he
31:22said we you know economists use
31:25equations to justify the divine right of
31:28capital in the way that court
31:29astrologers used the Stars to justify
31:32the Divine Right of Kings the key thing
31:34for me I think is to think of the no to
31:36the free market as a macro level and
31:38ecosystems as a micro level as this is a
31:41system and the system as it functions
31:43generally is a very efficient way of
31:45producing things that we want but you
31:46have to point it and steer it and you
31:48know you make adjustments and you decide
31:49well that as it operates is not actually
31:51generating what we want so the market
31:54does not provide a cost of pollution
31:55okay so we're gonna have to add a cost
31:56of pollution and we're gonna decide what
31:58that cost is we're not going to set at a
31:59level that economic activity halts but
32:01also going to stop people pumping sewage
32:03into the rivers and I think the same
32:04thing applies if you're running Google
32:06or Facebook or indeed Apple you you know
32:09you set your balances and you set your
32:10trade-offs and I think that's kind of
32:12and to me that's kind of a litmus test
32:14for rational discussion the one
32:16understands that all these policies
32:17decisions are trade-offs you can't have
32:19something for free and as long as you
32:21make a decision okay well we're going to
32:23accept that cost you know they'll get
32:24this benefit then you're having you know
32:26then you're having a rational
32:27conversation there are lessons from
32:29technology that can apply to those
32:31policy discussions because when you see
32:33that the way that say Google or Facebook
32:36or Amazon are making those kinds of
32:38adjustments to their platforms it's very
32:41hard to kind of go back to the idea that
32:43somehow this is just the best system is
32:46one in which nobody is applying those
32:49kind that kind of intelligent tests
32:51learn measure respond loop that we take
32:56so for granted in Silicon Valley and it
32:58is so often absent from our policymaking
33:00the best quote ever about monopolies is
33:02from the man who invented the idea of
33:04the invisible hand you know that no
33:05sooner do people have a trade get
33:07together then they start coming up with
33:09schemes to Deford the public in
33:11the person who invented the idea of the
33:12invisible hand and as he invents it he's
33:14saying yes but attention that's right
33:18we must believe that the world we live
33:21in is not the only possible world we are
33:23facing a set of great changes driven by
33:27technology and the first step towards
33:31making the world new is to believe that
33:34the world we're living in and the rules
33:36were playing by are subject to change
33:38that they are up to us that they are not
33:41inevitable and efficiency is not the
33:44only goal the secret of technology is
33:47that it allows us to do more to do
33:49things that were previously impossible
33:51and taking that belief to heart is the
33:55biggest thing that we can all do well
33:57that sounds like a good way of
33:58describing the book how can we make the
34:00world a better place
34:01WTF is out now thank you very much for
