00:00hi and welcome to the a 16z podcast
00:02we're here today with Benedict Evans and
00:04Steven Sinofsky in another one of our
00:06hallway style conversations in this
00:09episode they discuss what happens when
00:10the S curve levels out and especially as
00:12more and more businesses are relying on
00:14devices that need only browsers and
00:16Internet connectivity well will the $200
00:18bucks sitting at an employee's
00:19workstation look like or more broadly
00:21how do tech devices evolved for the
00:24enterprise from the mainframe to the PC
00:26to the tablet and the smartphone so I'm
00:28Benedict Evans I'm here with Steven
00:30Sinofsky and we've been talking a fair
00:33bit up around the kind of fundamental
00:35platform shift that's going on in tech
00:37from PC Wintel Mouse windows-based
00:40computers to mobile operating systems
00:42and we thought it was kind of
00:43interesting to talk about what happens
00:46to the PC in that environment and maybe
00:48the kind of a good way to set up that
00:50conversation is to go back and think
00:52about what happened to the mainframe
00:53because in a previous generation
00:55mainframes and mini-computers and
00:57workstations lost their scent position
00:59at the center of the tech industry to
01:00the PC into the client-server and so on
01:02but that didn't mean the mainframe
01:04business just kind of evaporated the day
01:06after Windows 95 shift instead I see
01:08IBM's mainframe business grew really
01:10healthily throughout the 2000s and so
01:12it's interesting now as we think well
01:14it's one thing to say that the center of
01:15gravity in tech moves to mobile and
01:17moves to mobile operating systems and
01:18all these use cases will slowly move but
01:21what happens to the other half of that
01:22story and what is say 5 10 20 year
01:24half-life of the PC look like yeah it's
01:26super C I'd like to put some color on
01:28that because I think that a lot of
01:29people don't really realize this but you
01:31know back in like 1993 94 IBM was on the
01:34brink of bankruptcy and the theory was
01:37that that pcs were just ending IBM's
01:40reign and disruption theory was still
01:43you know years away from being published
01:45in a book and an HBO article but the
01:47idea was like oh they're just gonna
01:49evaporate like and we think about
01:50disruption and we think about the
01:51blackberry and the company just going
01:53away overnight because of the iPhone and
01:55so but it turns out if you actually look
01:57at IBM the next 20 years were incredibly
02:01profitable for them and if you just use
02:02the stock price measure which obviously
02:04has all the other variables aside with
02:05it but but it actually IBM beat the S&P
02:08500 by 7x over that 20-year time frame
02:13on the back of a dead mainframe and I
02:16remember when the IBM Z series mainframe
02:19was announced and this was like a big
02:20this is gonna reinvent it but remember
02:21the only people working on mainframes or
02:24IBM and they came with this new
02:25mainframe and like the big thing was oh
02:27it's gonna run Apache and like every was
02:30like who cares like everybody's running
02:32them on intel-based servers and all of a
02:35sudden like the mainframe will get very
02:38exciting but it was the single source
02:40kind of dying innovation and last breath
02:43but for 20 years they profited from it
02:45and beat sp500 so it's super like in a
02:49sense what we tend to underestimate is
02:51this notion of how long the tale is on
02:53on something in technology particularly
02:56in the enterprise space yeah I'm the
02:58best example I've heard of this is of
03:00the UK that system the UK value-added
03:02tax system runs on backs
03:03created by Dec which I think
03:06discontinued it before many of the
03:08people listening to the podcast were
03:10born and yet that system is still
03:11sitting there running because well
03:13no-one actually decided to spend the
03:14money to replace it yeah and so so we're
03:17we're in this space now where it could
03:20very well be that the PC sees its most
03:22profitable 20 years but yet the least
03:25number of people thinking about it or
03:27worried about it or investing in it and
03:29the the least amount of change so if
03:32this is part of the story I mean you
03:33know one of the things you see around
03:34sort of the chatter around Apple there's
03:37people complaining that they don't do
03:38that much content to make that up
03:40to the max anymore and max don't improve
03:42very much frankly I think you can see
03:44the same thing across the whole PC
03:45industry that the s-curve is flattened
03:47out and if you buy a new PC or a new Mac
03:49it's not really very different from one
03:51that you might have bought four or five
03:52six years ago well you know Intel has
03:54very you know as announced like they're
03:56not even focused on PC chips like that
03:59it's about the data center and it's
04:00about IOT and VR and a bunch of stuff
04:02but not not like core i7 the next
04:05generation yeah and it seems to this
04:07sort of two to three parts to that one
04:09of them is you know the center of
04:11gravity the center of the excitement
04:12center of investment and and volume and
04:15everything else in the tech industry has
04:16moved to the mobile platforms from the
04:18PC platform and so if you were going to
04:20make a new widget or a new component or
04:22something of course you would target
04:23mobile instead of targeting pcs the
04:26though is that sort of pcs have sort of
04:28reached a point that there wasn't much
04:30more that you could do degree that you
04:31could really do with it I mean I wrote a
04:32blog post your two ago called like the
04:35best is the last talking about piston
04:37and I started out talking about piston
04:38powder and their forties and fifties and
04:40it's not that they stopped getting
04:41better because people stopped caring
04:42because Jets were there if they stopped
04:44getting better because they just wasn't
04:45anything else you could do you couldn't
04:47make them any faster woman any more
04:48efficient well I think that the the key
04:50to that and to add you know like a
04:53different a new kind of view on this too
04:55is that that what really also it matters
04:58is the API ecosystem that's being
05:00written to to provide capabilities for
05:03people and and in the business world
05:05what really took much longer than
05:07anybody thought was the transition to
05:10HTML and and the browser and so you have
05:13on the one hand a the mobile ecosystem
05:15where every business is busy building
05:17apps for direct consumer you know no
05:20matter what retailer you are no matter
05:21what airline you are no matter what
05:23insurance company medical records
05:24company you're building an app to
05:26interact with your consumers and the
05:28question is what are all of the people
05:30in your company using to do work with
05:33each other and for many years and for a
05:36very long tail those were client-server
05:38apps built on on the Windows platform
05:40and this is an area the Mac missed like
05:42nobody ever built these on the Mac and
05:45so part of what resurrected the Mac and
05:47brought it to the internet was the
05:48internet but but it took like 20 years
05:51and said this is this was the famous
05:53kind of don't get out of the Trojan
05:55horse before you inside the city gates
05:57where Marc Andreessen said that Netscape
06:00was going to turn Windows into a set of
06:01badly beat debug device drivers because
06:03all of the applications would be been
06:04done inside the web browser that was
06:06pretty matter of you to build like two
06:07sayings I do my best um but this was the
06:11point that all the applications would
06:12happen inside the web browser and web
06:14browsers will be on the PC and suddenly
06:15no one would care about winners ApS
06:16anymore and that was a so game displays
06:19are gonna happen with Java on the client
06:21which tells you how long ago the sports
06:22again people listening to this you've
06:24never used Java one client or Netscape
06:26or Netscape yeah many people never use
06:28netiquette but that was like the vision
06:30in cheese hell in 1995 yeah and it took
06:32until you know arguably at least 2005
06:35for things like Ajax to start making
06:37that at least vaguely plausible Ajax
06:39pioneered by Microsoft
06:40ironically and see like you get things
06:42like salesforce founded in 1999 i think
06:45or 2000 so just before the crash right
06:46but it really took at least a decade
06:48after that before that became something
06:49that even vaguely look sensible and of
06:51course now it seems entirely normal that
06:53you can basically use powerpoint in
06:54written in HTML inside a web browser and
06:56you can create all of these tools and
06:58applications in a browser but it took 20
07:00years for it to really become a thing
07:02for all of those enterprise and it's
07:03super I mean it's really surprising how
07:05how long that that took and part of it
07:07is because it takes of like most
07:08enterprise applications you know you you
07:10talk to a CIO and they're managing a
07:12portfolio of a thousand two thousand
07:15three thousand what they call
07:16applications and four it takes a very
07:19long time those aren't all resource
07:20evenly like in an enterprise you don't
07:22have three thousand teams so just it
07:24just to be clear when you say is that
07:25when you say they might have a thousand
07:27if that might be like five hundred
07:29different things built on the same local
07:30thing or something oh right right so
07:32they'll be they'll be like an expense
07:33reporting app but then they'll be the
07:35manager version of the expense reporting
07:36app or they'll be the reporting tool for
07:38this expense reporting or there'll be
07:40the travel people who input receipts or
07:43anything like that and and then you know
07:45forget something like budgeting which
07:47involves dozens of departmental level
07:49things vendor things up invoice approval
07:52and this so there'll be so many of these
07:55apps and many of them have their also
07:57baked into the workflows of the company
07:59and so it takes a very long time because
08:00unlike a company you know an enterprise
08:03is usually run with a pool of resources
08:05to build apps and only the biggest apps
08:07have dedicated teams year-round and
08:09other apps are starved for resources
08:11until they need changes at the fiscal
08:13year boundary or tax law changes or they
08:16add a new geo or a mergers a net and
08:19acquisition integration and so the tail
08:21is very long to get these done but the
08:23demand was there immediately and so now
08:26what's happened is we've reached a point
08:28where most enterprises have moved all of
08:30their applications except for a very
08:32very few to the browser to a web
08:34interface to a web interface and you
08:36could see this everyday you you look at
08:38you go to a big metropolitan hospital
08:40and there's an 85% chance they're using
08:43one of the two EMR systems from either
08:45epic or Cerner and so you you talk the
08:48doctor or the nurse is looking at your
08:49lab results in a browser in that
08:52application now what's interesting is as
08:54probably even doing that through a
08:56Citrix client because they don't even
08:58want any state at all on the PC but it's
09:00still a browser certainly when you're 20
09:03to 5 years ago that would have been
09:04Windows app and absolutely would have
09:06been a Windows app and and you see the
09:08same thing within the if you unfold a
09:10hospital like looking at CTS or MRI is
09:12like five years ago that was a very
09:14complicated java application that didn't
09:17really work that needed a firewall
09:19specially configured and stuff like that
09:21the same goes for American Express
09:23travel agents or 401k managers or
09:25customer service reps and all these
09:27customer this is just an aside here
09:29people talk about get that we have this
09:31conversation over never began a getting
09:33work done on PCs and you need these apps
09:34you need to do your stuff on a PC the
09:36vast majority of work is getting done on
09:38a PC is not actually happening
09:39constructing complex financial models in
09:41Excel or doing stuff in InDesign or in
09:43Xcode or in AutoCAD it's you know you've
09:46got a Windows box that cost your CIO
09:48$300 and you're sitting there putting
09:51stuff into a SAP client nsab client
09:52possible management program or something
09:54you've got or you're just doing email
09:56yeah but you've got a bunch about a
09:58bunch of those sort of specific vertical
09:59apps that need a CPU power right right
10:02or anything like that they just
10:03front-ends the data in something they're
10:05not they don't really require multiple
10:06monitors they have no device
10:08connectivity and now those apps like
10:11they don't even run locally they require
10:13doing so it's actually really
10:14interesting like to go through the US
10:16workforce and to just sort of take a
10:18quick look at like 156 million or so
10:21people with jobs in the US so this is
10:22just the full employment of the US and
10:24you could get these on the Bureau of
10:26Labor Statistics it's super interesting
10:27data used to use it all the time to
10:29figure out target addressable market
10:31yeah I use that to count up to do a
10:33chunk of how many elevate to attendance
10:34error in the USA since 1900 okay which
10:38is actually a category yeah but there's
10:40like 20 or so top-level categories and
10:42of course the top one is office
10:44administrative support at 23 million but
10:47if you go all the way to the bottom
10:49which is farming fishing and forestry at
10:51about a million but in between its it's
10:53a pretty straight linear curve and very
10:56quickly if you just put them in buckets
10:58about a hundred million basically don't
11:00need pcs at all and just you could use
11:02their phone for everything like a really
11:04good example of this would be
11:07you know five years ago like if you were
11:09to do a construction project the first
11:11thing that they would do is come in and
11:12say we need a phone line and you need to
11:14lease a PC in order to remodel your
11:16kitchen because we're gonna have the PC
11:18for looking up the project and looking
11:20up parts let's supply lists and stuff
11:21the whole time and we need a phone line
11:23to connect and now they all just show up
11:26with their phones and they're not even
11:28like issued by the construction company
11:30because most construction construction
11:32workers are vendors and so that you you
11:34can't even have a job without your phone
11:36and so all of a sudden like the whole
11:38use case sort of goes away and just to
11:41put a number on it there's about 600
11:42million business computers in the whole
11:44world and so if you look at the US
11:46there's probably about a hundred million
11:49and so then you go wow and a 50 million
11:51of the people need pcs there's about 50
11:53million that are controlling cash
11:55register is controlling elevators
11:56controlling robot arms that do machines
11:59that are in ATMs and a whole bunch of
12:02stuff and and then the question is of
12:05the 50 million that have jobs that are
12:06sitting in front of a computer what
12:08happens to them hmm and this is that the
12:11thing we kind of we kind of push away at
12:13that said some portion of those are and
12:16people always tend to react to this by
12:17saying well I'm going to I need to run a
12:19code I need an IDE I need to build a
12:21normal model in Excel okay fine
12:22that is running your payroll on a
12:25mainframe that you can't switch to a PC
12:26because it needs to do stuff said and
12:28your mainframe can do find that's some
12:29portion that's 510 percent of the base
12:31then there's all the other stuff which
12:33is stuff we've been talking about what
12:34is that PC at the nurse's station in the
12:36hospital what are they doing with that
12:38and at what point does that become a
12:40browser and what what does that platform
12:43mean what is that device and this is
12:46sort of you know you get to the kind of
12:48the $200 device the $200 PC here and
12:51this sort of there's one narrative that
12:52says well of course that should be an
12:53iPad or it should be a cheap and would
12:55have Burton you borrowing arguments
12:57about Android and so on yes there's
12:59another argument that says well really
13:00this should be and we can we come back
13:03to you know the browser was a platform
13:05from from mark from 20 years ago there
13:06was a parallel there and recent fantasy
13:08that you were going to have a Java note
13:10was it called the net PC
13:11yeah well that was Larry Ellison and and
13:12and the network computer but it was the
13:15same thing you'd have a list interesting
13:17all the things that maybe cringe and the
13:1890s well you know there's a differ
13:20between there's no difference between
13:21being early and being wrong yeah well he
13:23was wrong at the time but now if you to
13:26me a Chromebook is exactly what Larry
13:27Ellison was trying to do 20 years ago I
13:29mean are technically the architecture
13:31might be different it's exactly the same
13:32thing right it's a web browser that runs
13:33client that runs applications over the
13:35network and and and you know we you know
13:37the this notion of having a device it's
13:40a terminal it's a terminal
13:41it literally is an IBM 3270 terminal
13:44accepting connected to a single computer
13:46by a wire it's connected to all of the
13:47internet and that's actually a huge a
13:50huge huge difference and and and so it's
13:52it's very interesting to think about
13:54like what are the the alternatives
13:56because you you could have a Windows PC
13:58that has a browser in it you could have
14:01a tablet like from an iPad or an Android
14:05tablet that has a browser and apps in it
14:07or you could have a Chromebook which has
14:09a browser and now with the most what
14:11pixel on the most current one also runs
14:13Android apps and so I think it's super
14:15interesting from an IT perspective
14:16because you know really what's on our
14:18mind is just like this is a cost
14:21containment problem for IT because now
14:23they what IT is very clever like when
14:26when pcs came out was what happened was
14:29he would buy a PC for like $1500 but it
14:32turns out that all the profit went to
14:34Intel and Microsoft for that and then
14:36all the cost went to IT
14:38so they've seen this before so now
14:40they're very smart they like tell
14:42everybody you have to have a phone to
14:43join our company and you can't you have
14:46to have fun and they publish your mobile
14:47phone and the listing the end but then
14:50they put their own monitoring software
14:52on it so they can wipe your phone and
14:53remote manage it but they're not paying
14:55for connectivity they're not paying for
14:56the device and it's sort of a compromise
14:58you you make it's sort of like using
15:00your car to commute to work and if you
15:01break it you have to take you to the
15:03Apple Store yourself right right and and
15:04so they've really pushed pushed in in
15:07that direction and so the question is
15:08they they don't want a device that
15:11brings them cost you put a number on it
15:14Gartner coined this phrase the total
15:16cost of ownership of a PC and you would
15:17buy a computer if Apple left that charge
15:19yeah yeah well and we at Microsoft at
15:21the time really hated it because you
15:23would buy like a $1500 computer and
15:25Gartner would tell you it cost like
15:26$40,000 a year to give that computer to
15:29an employee and and it turns out like
15:31wow it was the greatest shell
15:33in the world to push all the costs of
15:35the customer it's not unlike on-prem
15:37computing does the same thing like
15:38people complain about AWS pricing but
15:41the the flipside of it is like running a
15:43data center yourself as ten times as
15:45expensive and the same goes for the for
15:48having a client device so the question
15:50is given the choices of life you're
15:52sitting and looking at that nurse's
15:53station you've got a Windows PC there
15:55that has management cost do you replace
15:57it with another device that has
15:58management cost or do you replace it
15:59something with something that give you
16:01some fundamental advantage right and I
16:03think and and I think that that is just
16:04super interesting because like you could
16:06take the doctor station and the nurse
16:07station you you could even take like
16:09like a like the bankers is my favorite
16:12because I think if I if I look at the
16:13numbers there's eight million people in
16:16the US that work in business and
16:18financial operations and you know okay
16:20so you it's not a hundred percent of
16:22them that are doing Excel modeling in
16:23fact the vast majority of them aren't
16:25doing like anything like what we think
16:28of as sophisticated Excel use in fact
16:30the other variable that's really changed
16:32is the rise of the browser and the rise
16:34of SAS applications has made it such
16:37that the amount of ad hoc analysis that
16:39you have to do is greatly reduced
16:41because the tools themselves this is the
16:44what Salesforce has done the tools
16:45themselves have added the to of doing
16:47work the tools are doing the work
16:49because what the PC was not just about
16:51like moving we're computing happened but
16:53it was the data was on a mainframe and
16:55the format that you could get the data
16:57was fixed and so what you know if you
17:00wanted to see sales by country and the
17:02report was only sales by region you had
17:04to get all the region download
17:06everything into Excel and manually
17:08edited all the features of Excel around
17:11list management and pivoting were all
17:12designed to rule unbundling a mainframe
17:14they're all in bundling the mainframe
17:16and so then this notion that Excel is a
17:18universal tool came to be because it was
17:20the only tool where you could just be in
17:22control of what you wanted to do and now
17:25what's happening is you it turns out if
17:27like you're a customer support agent you
17:28want to understand the flow of the data
17:30well whether you're using Zendesk or
17:33some other CEM solution or something
17:35like the tools or uh have bundled in the
17:38capability to to do all that in a way
17:41that doesn't tell the business oh you're
17:43just gonna be a cookie cutter business
17:44by using our software it's in fact just
17:47you can be a supply chain and be your
17:49own supply chain if you use our software
17:51and so that to me this SAS plus the
17:54browser has really changed the dynamic
17:57of what IT is willing to very SAS and
18:01then there are no it's a browser and
18:03then there is a device itself and what
18:05do you have to deal with right right and
18:06as you go to a sandboxed model of iOS or
18:10to a lesser extent Android you take a
18:12whole step change in the management cost
18:14of having how you deal with that device
18:16right and if you go to a peeler terminal
18:18quote-unquote device like a chrome box
18:20or the network PC or 20 years ago then
18:22you get arguably another step change
18:24which is incidentally is why the added
18:26putting Android apps onto Chromebook
18:28you've signed up our obstacle because
18:29you turn you back you and happy suddenly
18:31you don't have the advantage that Chrome
18:32has a round wooden yeah I have to say I
18:34like to dive in on that one I found that
18:36one you know somewhat puzzling to begin
18:39with because the whole value proposition
18:41of a Chromebook is it's stateless is
18:43less hassle than Android it's less
18:44hassle than Android and and it's just a
18:47browser and and I've been using a pixel
18:50for you know three weeks now and it's a
18:52great great device it's a little big for
18:55what I carry around but it works super
18:57well but the Android apps on top of the
19:00the security management you know there's
19:03a store now you have to think about it
19:05you know you you give the device to
19:07what's cached on it what do you have to
19:09think about then also like if you put it
19:11out in a school you have to go turn off
19:13the store you can't let all the students
19:15go and download apps it's sort of
19:17whereas before you could just block
19:18things at a firewall or use ratings or a
19:21whole bunch of other stuff that before
19:22it even gets the device and so and then
19:24on top of all that the Android apps
19:26aren't really built for a device with a
19:28clamshell keyboard in fact one of the
19:30things I've been struggling with quite a
19:32bit is that the keyboard the soft
19:33keyboard pops up all the time and that's
19:36exactly our leader on and building
19:39surface like even on arm that was one of
19:42the big things and we realized whoa if
19:43there's ever a keyboard attached you
19:44never ever want the soft keyboard and
19:46the iPad with keyboards it handles this
19:48super super well like you never get the
19:51the keyboard button but so it's I I find
19:53that somewhat puzzling yeah but this is
19:57but this is kind of the transition is on
19:59the one hand you can look at this and
20:01we moved all of our stuff from on-prem
20:03or mainframe into the cloud and we moved
20:06the front end from a Windows application
20:08that we had to think about all the time
20:09and managed an update to a browser and
20:12so the end user device now for all very
20:15very large proportion of the people who
20:17have a PC actually only needs a browser
20:19yeah and so what you want therefore is a
20:22device that only has a browser in some
20:24way because you don't you know the
20:25trade-off of being able to run
20:27third-party apps is a whole other layer
20:29of management problems and cost and so
20:30on so if you don't have the apps and you
20:32get rid of all of that and so I think
20:34maybe the kind of the genesis of this
20:35conversation originally was I sort of
20:37thought well you know imagine the office
20:39floor for a health insurance company in
20:42a city somewhere in the Midwest and
20:43there's 500 people in the building and
20:45presume for the sake of argument the
20:46whole thing hasn't been disappeared by
20:47AI and they were still doing their job
20:49but they're all they need is a browser
20:51and a keyboard and maybe a mouse and so
20:53what is the CP what is the box that's
20:55powering the browser is that a
20:57200-dollar PC from Dell which is
20:59certainly one option yeah is it a
21:01200-dollar is it or is it a chrome box
21:03or is it a tablet and it seems like in
21:06that use case where it's never going to
21:07get picked up and carried around then a
21:08tablet may be not the right solution
21:10well that you know the screens not gonna
21:12be big enough it's touch only which is
21:15not the best interface for a person
21:16sitting doing a call center mmm so that
21:18becomes so it becomes something that
21:20looks like a PC except that what's in
21:22that little box which actually is
21:23probably going to look like a hockey
21:24hockey puck it might or an HDMI stick
21:27yeah exactly that's how what I think
21:29Google should be doing with Chrome they
21:30should be making HDMI sticks but it
21:32should be a it's going to be something
21:34that basically has no native IP api's at
21:37all and no African native well and also
21:38it's not going to necessarily have to
21:40focus on battery life
21:41it certainly doesn't need to make
21:43everything more complicated by thinking
21:44about offline or Wireless LAN
21:46connectivity like those are things where
21:48the tablet and and tablet apps are gonna
21:52be way way better and that actually
21:54introduces one other dimension to this
21:55which is while IT in enterprise is very
21:58focused on building browser-based apps
22:00for the past say five or seven years and
22:02migrating the other thing is they're
22:03under a lot of pressure now to produce
22:04like mobile apps like for iOS and a leap
22:08frogging thing right right there isn't
22:09that because you took your mainframe app
22:11and you turned it to an on-prem PC app
22:14and then you have the Windows app now
22:16not talking to on Chrome but talking to
22:17the cloud but it's still a Windows app
22:18and then the cloud is now talking to a
22:20browser app but then that browser app
22:24then turns into a native app on the
22:25piece of native app on a smartphone and
22:27so you've got these kind of parallel
22:29development tracks of where is the stuff
22:30being stored and what does the client
22:32look like and what is the development
22:33process for those kind of leapfrog right
22:35other but there's also I also think that
22:37there's two other dynamics at play if
22:39you're the IT developers building the
22:41software for it to run your company one
22:42of them is that that you're your
22:46marketing team is saying we need a
22:47direct-to-consumer app and so they're
22:49saying like we're a big retailer we need
22:51our retail shopping app and so there's
22:52all this pressure on that and these
22:54companies have been making software for
22:5550 years they just never made it for
22:56consumers exactly the stuff you saw if
22:58you kind of leaned over the desk right
23:00in the school so you've got you've got
23:01that sentiment the other dynamic which
23:03is very real is the IT people want
23:05people to want to use their software and
23:07and so like they're kind of motivated to
23:10like well we'd rather have it not look
23:12like back office software and so they're
23:14actually drawn to thinking about maybe
23:17if we just use the mobile app and the
23:18sales associate and the customer both
23:21used it at the same time and in many
23:23ways that's what what's like the the
23:25convergence of electronic medical
23:27records like they're now like epic has a
23:29consumer app that you look at to look at
23:32your own chart and your test results
23:33from your physician and things like that
23:34and like the physicians are looking at
23:36it going while the browser one is kind
23:38of clunky compared to the consumer plus
23:40I I have to use the mobile one anyway
23:42when I'm not when I'm at what I'm doing
23:44my family life and like I'm at the
23:46soccer game and a patient calls and
23:48wants some information like and so
23:50there's this very interesting thing
23:52where it's not just cait's the
23:54consumerization of the experience right
23:55and which is a phrase that's been around
23:56for a decade yeah but it this
23:58consumerization is also likely to create
24:00some tension between what do we do in
24:03the browser and what do we do as a
24:04mobile app because the the good news is
24:06the cloud means that that's the
24:08definitive source of truth for data
24:10which is another huge change that's
24:12happened which is now there's only one
24:13budget that's the one in the cloud not
24:15the one that's on everybody's
24:17spreadsheet XLS on a desktop floating
24:19around but the question is how do you
24:21get to it and like what's the
24:22interesting thing and I actually think
24:25people are under estimating the the
24:28for mobile and because mobile is two big
24:30platforms you're gonna start to see
24:31companies invest much more in one of the
24:34two because that'll be the only way to
24:36really deliver on and you can see Apple
24:38of a C trying to drive that with the
24:40deals that they're doing with IBM and
24:41Accenture and PwC and so on yeah just
24:44coming back to the kind of the mainframe
24:45PC point and I'm sort of thinking out
24:47loud here but definite this is yeah
24:49exactly but but what what happened to
24:51for IBM was the platform became
24:53something no one really thought about
24:54but they carried on selling mainframes
24:56and mainframe software and all the stuff
24:57the business around that for 20 years um
24:59the thing for Microsoft is obviously the
25:02kind of the server bit has gone away
25:04nothing gets run on you know Windows PCs
25:07the Windows PC is client dead great out
25:10of the internet because what else were
25:12you gonna buy now we're talking about
25:14okay the only thing in the enterprise
25:16that's Windows as Microsoft also sort of
25:18there's Windows is these clients that
25:21are only there to run a browser and no
25:23one is writing anything for any Windows
25:24API anymore or any Microsoft API is
25:27there and so at that point this kind of
25:30Microsoft kind of becomes completely
25:32replaceable in a way that kind of on
25:34sort of in a way that IBM wasn't if you
25:36said I mean because you are still
25:37running I mean IBM mainframe
25:39applications well of course I mean and I
25:43understand but I'm gonna push back on
25:44that because AI don't like to think
25:45about Microsoft is just totally
25:46replaceable but I will look at it I'm
25:48thinking that the PC is it possible
25:50because it's not running a PC specific
25:52the mainframe wasn't right so the
25:55mainframe the real business of the
25:56mainframe was keeping running like 20
25:58years of software investment you know
26:00like you really you know until you've
26:02really met with like a big bank or a big
26:04insurance company telling you about
26:05their cobalt like we met with a
26:07government agency this week and they
26:08have one application for one branch for
26:11one scenario as seven million lines of
26:13COBOL code so it's not going anywhere
26:15they and they have to find people to
26:17help them to manage it and deal with it
26:19now there's a great you know web
26:22front-end with AP eyes and stuff now
26:24that's all you see it through the
26:25browser but you you know you're very
26:27constrained in what you can do because
26:28it's responsiveness is a giant COBOL
26:30application going through an API but the
26:33things that the most interesting thing
26:34in this conversation is the the the vast
26:37richness of the the Windows platform has
26:39rolled up into office and
26:41and is the pretty much the active
26:43development on win32 modulo the count
26:46that you do of like the autocad's and
26:49the SolidWorks and the disso system fan
26:52shop and the Photoshop and like this so
26:54there's a core of perhaps and I think
26:57you said you estimated 100 million I
26:58thought it was like a hundred million
26:59maybe 150 million people worldwide that
27:01use all of those what you would think of
27:03as like a professional application
27:05that's doing like hardcore sophisticated
27:08creative stuff on a PC yeah like chip
27:10design is not like a browser-based thing
27:12it's incredibly associated but there's
27:13you know only something again just to
27:15put like you know there's 2.6 million
27:18architects and engineers in the US and
27:20many of them aren't actually the kind
27:23that sit in front of any of them in no
27:24tights you want to go to camp yeah like
27:25well they're you know they're part of
27:26architecture firms but they're not
27:28necessarily the engineers and so the
27:30interesting thing is like you know how
27:33much and this is really where people
27:34really push back because what's gonna
27:37happen is people as you've pointed out
27:39many times like people's jobs are gonna
27:41change and they're not gonna be like
27:44formatting a Word document for 12 hours
27:46at a time or as I use think about it's
27:47like debugging the word document and so
27:49like on my Chromebook I've used I only
27:52use the browser apps in office 365 and
27:55so they work great and in fact they're
27:57in a race to add back all the features
27:59even though I don't need them like I'm
28:01never gonna use kerning or page layout
28:03because I don't really print anything
28:04and and things like that and so what's
28:06super interesting is that's the split
28:09and that's really what happened with
28:10mainframes like the platform became
28:12existing software said the analogy that
28:15would be there were specific kind of
28:16applications even now that one better in
28:19principle on a mainframe than they do on
28:20a PC server form because they're sort of
28:22there's character character of that kind
28:23of application running a national tax
28:25system for example is very different
28:27from doing Google search is a different
28:28kind of computation right and so there
28:30was a core bit of the mainframe market
28:31that actually needed to be a mainframe
28:33yeah and there's a whole a much larger
28:35portion of it that could carefully move
28:36over to PC and it takes much longer than
28:38anybody thought but then and then what
28:40we were all kind of getting ideas of a
28:41similar piece here of there's a sort of
28:44a core part of the PC market that
28:46actually needs to be on a PC because
28:47you're running Excel or you're running
28:49for a job or watercad and the other 90%
28:51of it actually doesn't need to be on a
28:55book or it couldn't be on something it
28:56has a browser or can be on a smartphone
28:58or tablet and and one of the things that
28:59I think really capitalized on on what
29:02IBM did was they also got smart about
29:04how to manage it and this is something
29:06that that I think is super interesting
29:08which is they said well if if it's
29:10really important to be valuable in this
29:12scenario running this way let's not
29:13radically change this way like they
29:16didn't listen to like mainframe
29:18enthusiasts they listened to the
29:21business leaders and and so what's
29:23interesting is how how those systems
29:26evolve so that they capitalize on those
29:28kind of applications without churning
29:30them because those people are also not
29:32like the big if it's about change I mean
29:34on the PC just got an update and the
29:36forums are just exploding with like the
29:39keyboard shortcuts changed or this
29:40change or that and I kept thinking why
29:42did they do that like why bother to have
29:46that big a change when you fundamentally
29:49the scenario doesn't change at all hmm
29:52yeah well there is always that there's
29:53like the love the one lost reset yeah
29:55yeah well it's super interesting so let
29:58quick I'm gonna wrap up with like my my
30:00desire or what I think if I were running
30:01Enterprise IT my view of it is that
30:04what's going to happen is the call
30:06centers of the world the vast majority
30:08of administrative assistants of office
30:09workers of mid managers and things are
30:12all going to gravitate towards basically
30:15sitting in a browser on all day but
30:17mostly using their phone so it'll be
30:19like 80% of the communication will
30:22happen in their phone mostly using
30:23communication apps and sass apps and
30:25things like that and then when they're
30:27on their PC doing sort of the creation
30:29step of their job they're connecting to
30:31SAS apps they're integrating things and
30:33and it's gonna be in a browser and then
30:35I believe that IT is gonna really push
30:37for a device that doesn't have an
30:39execution engine on it and I think it's
30:42gonna be really interesting to see how
30:43Chromebook evolves because I kind of
30:45feel like the Android step was a step
30:47backwards when they really had like the
30:49exact right device that everybody wants
30:52it just they wanted like one more thing
30:55or something so I and I and I do think
30:57that that the tablets are on the uptick
30:59because that's really what's going to
31:01replace the the core marketing executive
31:04traveling sales executive press all the
31:06executives traveling without yeah
31:08and even executive we say in a very
31:09brothel people right people traveling
31:11with laptops yeah yeah we'll travel with
31:13a tablet instead yeah yeah I think I
31:15think that's right and I said you know
31:16to come back to the right the way open
31:17this it's it's interesting we I remember
31:21like sort of 18 months ago or something
31:22I was trying to find stats on mainframes
31:24and I got this number for the install
31:26base of IBM mainframes measured in MIT's
31:28of course with millions of instructions
31:30per second it's not like how many
31:31mainframes it's how many instructions
31:33per second are installed and the number
31:35went up I think something like 10 times
31:37between 2000 and 2008 yes nobody in the
31:41tech industry was thinking about
31:42mainframes in right and yet the IBM
31:45business was actually doing really well
31:47and I think we now we're now at this
31:49period where people are still arguing
31:50about whether the PC is going away at
31:52all but actually as we think about what
31:54happens over the next you know 10 20
31:56years or so you're going to see a sort
31:57of a similar process where the business
31:59is going to be there in some form but
32:01just kind of changing yeah well it'll
32:02drop what's happened is you know it's
32:04half the run rate that people thought it
32:06was going to be and what happened with
32:07the mainframes was super interesting
32:08which is that the need for information
32:11technology exploded and they couldn't
32:13move to that more scalable platforms so
32:15you I be M had them locked in yeah the
32:17PC is different because it's an
32:19individual empowerment tool everyone
32:21wants to do that taxi's online say
32:22suddenly you need more mainframes right
32:23and you don't need more devices running
32:26America online connected that that have
32:29the processing power to download chips
32:30off to do your taxes and I think that's
32:32the difference in the evolution with the
32:34PC is that the number of people that are
32:36going to build Excel models isn't going
32:38to increase or run Photoshop or AutoCAD
32:40hmm but the number of people
32:42communicating with them and coordinating
32:44and vendors and things like that that
32:45are gonna be cloud-based is gonna go up
32:47astronomically which is why we see the
32:49mobile phones so that's that's what we
32:50have for today that was our hallway
32:52conversation thank you
