00:00hi and welcome to the a 16z podcast
00:02we're here today with Benedict Evans and
00:04Steven Sinofsky in another of their
00:06hallway style of conversations they
00:08share takeaways from the recent annual
00:09Consumer Electronics Show CES 2018 more
00:12broadly they consider what's the future
00:14of the smart home and who are the key
00:16players especially given the rush to
00:17bundle smart devices together under a
00:19single system and then on the design
00:21side of things one is simpler better and
00:23which devices require more complex or
00:25simpler UI so the two of us separately
00:28went to CES in Las Vegas with another
00:33180,000 even see each other which is
00:35sort of nerdy I did bump into like 10
00:37people that I knew at random places
00:38around the show but we didn't bump into
00:40each other and we each walked for about
00:4110 miles and we saw a bunch of stuff and
00:43we saw a bunch of press releases and we
00:46saw a bunch of news coverage and now
00:47that we are week and a half on and we're
00:49sitting and thinking okay have you
00:51digested this how do you think about
00:53what was actually important as opposed
00:54to what were the 300 press releases and
00:56the exciting coverage and so I think you
00:59know there's just there's there's so
01:01much there and I think the first thing
01:02that's kind of interesting is a lot of
01:04people go and they want to see the grand
01:06master plan for every big company and so
01:09the question is can you really discern
01:10the grand master plan for every every
01:13big company yeah I mean so you see
01:15clearly there's a whole Alexis story and
01:16then you've got people going around
01:18saying well excellent voices and your
01:20platform and this is going to replace
01:20smart phones and Amazon's got off to a
01:22great start and Google's coming up
01:24behind and it starts sounding like kind
01:25of NASCAR commentary as to who's in
01:27and it was Apple are completely screwed
01:29because the home column is Christmas and
01:31all of these and Cortana is like doing
01:32press releases the whole time like
01:34they're still gonna try and and they're
01:35gonna you know take a slow pace you know
01:38and say the thing that I think you see
01:39here is particularly when you look at
01:41kind of a super big company that's got
01:42lots of stuff going on is there were
01:44some things that are actually this is
01:46the board level fundamental strategy of
01:49what they think is crucial to the
01:50existence of the company in the next
01:51five years and then the stuff that are
01:54kind of applications that flow out of
01:56that some of which may or may not work
01:58and then the things that happened
01:59because you've got 10 or 15 thousand
02:01people and you've got an SVP for widgets
02:03and taps and the SVP for widgets and
02:05taps is going to try and get into
02:06bathtubs and that doesn't mean that
02:08anyone in the board knows that you were
02:09trying to get into bathtubs it's just
02:11because you've got somebody whose job it
02:12is to do that so they
02:13it and I think particularly when you see
02:15these kind of Blizzard of announcements
02:16in these huge stands you have to kind of
02:18take a step back and think well which of
02:20these things are tactical and which of
02:21them are strategic and which of them are
02:23existential in which of them well
02:24they're just doing that because they've
02:25got people to do it I think that that's
02:27really important especially in the
02:28startup world because there's this
02:29tendency you know if you're a startup
02:31and you work in a particular space and
02:33then you just hear an announcement from
02:34a big company that they're working on
02:36something you apply like oh my god it
02:39must be important because that although
02:40I'm gonna get squashed right right it's
02:42important to me so it must be super
02:44important and it forgets that dynamic in
02:46a big company where you know you have n
02:48groups working on their vertical little
02:50things and then the new thing appears in
02:53it's n plus one the likelihood is that
02:55three four or five of those groups are
02:57each gonna go try to do it and to your
02:59analogy it's like if bathtubs became
03:01like a thing it's likely that all the
03:03groups are gonna go find a way to go do
03:05bathtubs and we joke about this all the
03:06time about like adding photos or adding
03:09messaging and so a lot of what you see
03:11at CES is it's the internal bureaucratic
03:14it's the org chart it's just the org
03:16chart of the company so like one of the
03:18ones I saw was like a whiteboard and you
03:20know Samsung has a TV group say they
03:22help the white vote they and they have a
03:24foam group for them and they have an
03:26industrial group and Microsoft has a cut
03:27his collaboration say they have to have
03:29a whiteboard but you know Epson has a
03:31whiteboard because they might displays
03:32right right and printers and their
03:34whiteboard had a printer built in and so
03:35the first thing when I saw the
03:36whiteboard which was announced at a very
03:38low price and it was a very nice product
03:39was I wonder what operating system it
03:41has and like which part of Samsung made
03:44it because that would dictate like
03:46doesn't have an OLED screen or not an
03:48OLED screen is it touch or not so and it
03:50turns out it came from like the
03:51industrial equipment side of the house
03:53which meant it didn't even really have
03:55like a whole LS in it it had only just
03:57enough software to like make a
03:59bulletproof whiteboard which turns out
04:01to be a great product but you also know
04:03that they have people in three other
04:04groups in Samsung he looked at that
04:05press release in this war and they want
04:07to have a meeting and they're like we
04:08got to get in the room with mr. Kim and
04:09figure out like why I'm not responsible
04:11for whiteboards anymore but then the
04:13interesting thing is what do we say
04:15about it for the the pure you know the
04:18the companies we think of as tech
04:19companies you know cuz it's really like
04:21you could divine the master strategy for
04:23like a Samsung or an LG yeah say well
04:25this is sort of the thing we
04:26talking about earlier that everyone goes
04:28so and says what's the Google strategy
04:29what's the Facebook strategy or see
04:30Amazon strategy when you go to the
04:32Google's the LG stand or the Samsung's
04:35down it's actually kind of interesting
04:36you look at the tech press coverage and
04:37they kind of know they're like why is it
04:39a washing machine here why is our
04:40dishwasher for like five years now
04:42they've been doing this yeah and the
04:43answer is these companies make
04:45everything and they want to make
04:46everything electrical in your home they
04:47want to make the air they make the air
04:48conditioner and the washing machine and
04:50the TV and the PC and the phone and they
04:53have an all chart around that and so
04:55they want you to buy that your washing
04:56machine from them and clearly in some
04:59way the washing machine is going to be
05:00digital or connected at some point in
05:02some way over the next 20 years and so
05:04they want to do that because they've got
05:06connected people and they want it to be
05:08have some software in it or some
05:10interface it means you have to buy the
05:11washing machine from LG and not from
05:13Samsung and not from Electrolux and so
05:16they've got their kind of leverage
05:17strategy of connecting all of these
05:19things together which is also
05:20incidentally what Sony were doing 20
05:22years ago so what Sony did and and
05:24interesting like Sony was very
05:26impressive but like massively impressive
05:28but didn't make nearly the breadth of
05:30things that are currently made by the
05:32Korean companies I mean it's
05:34mind-boggling just how much stuff unique
05:36products they make and get distribution
05:39like you could go to Best Buy and see a
05:41lot of stuff from Korea yeah I mean
05:43there is it's the sort of stories if you
05:44live in Korea you could spend your
05:45entire life only consuming Samsung
05:47products from the Samsung house to the
05:49Samsung car to the Samsung bank and
05:51everything else I mean one of the things
05:52that kind of comes out of that there
05:54when you look at like back to the
05:55washing machine point okay washing
05:57machine has a screen and there's a bunch
05:59of interesting things you can do like
06:01you can pour a whole bottle of washing
06:02liquid into it and then you can tell you
06:03it what loaders got and it will
06:04automatically put the right watching
06:05liquid in so you don't have to load it
06:06every single time so there's there's
06:08kind of interesting useful stuff you can
06:09do right but then you think okay just
06:11isn't you to have Android on it does it
06:12need have a little tablet on the front
06:13does it need to have a screen at all
06:14does it need to have just have a kind of
06:16it have a go button and this is a
06:18question that applies to all of these
06:20sort of non these compute devices that
06:23are sort of smart and sort of a computer
06:25but aren't going to be your main device
06:26in order PC they're not a phone they're
06:28like a fridge yeah so how much you eye
06:30does that have how much computing does
06:32that have how much value do you have on
06:34that device versus how much is it like a
06:36single purpose thing and I think that's
06:37a really important point for folks to
06:39understand because that's one of the
06:40you can sort of discern how much is like
06:43a corporate existential strategy from
06:46like a local like I make the the
06:49plumbing based appliances which might be
06:51how the org chart is organized because
06:52like if you looked at Sony way back in
06:54the day like all of a sudden everything
06:56had a memory stick in it kinda choked up
06:58and a dial and and like a memory stick
07:00in a radio seemed like a very dubious
07:03kind of thing but believe me the radio
07:05person went to a board meeting and was
07:07told memory sticks will go in all of
07:09your products and it was absolutely a
07:11board level kind of thing where is it
07:14Samsung just like the whiteboard showed
07:15you know it's not clear yet that like
07:18Android has to be in everything but it
07:21is clear if you went to both of them
07:22that Wow like Bixby was a board meeting
07:24yes plan and I picked Samsung's Bixby
07:27and Chloe at LG those were board
07:30meetings or at least like an exec staff
07:32meeting where they all were told you've
07:34got to go participate and then you can
07:36sort of see the seams because way out on
07:39the edge a home security camera they
07:41just do like the checkbox minimal API
07:43because he's like look I got to compete
07:45with ADT and rang and a whole bunch of
07:47companies for distribution in the US I
07:48don't have time to be all corporate plus
07:51if I want to actually get into Best Buy
07:53they're gonna tell me will you got a
07:54supporter like so in Google home they
07:56think it exactly you have these devices
07:57that have its you appointed it says so
08:00you've got a bunch of Sony devices had
08:02the toaster in it that had a memory
08:03stick and an SD card and a compact flash
08:05card weight and so you have the security
08:06camera that has Alexa and Bixby and
08:09Google home and home kit right which is
08:11for startups a super interesting point
08:13because it means like if you take the
08:15lead from the big companies in what to
08:17do you'll see that all of them they have
08:19the resources to like do everything five
08:22times and in fact like even some of them
08:24will they'll literally just have two
08:25products they'll have the for the Korean
08:27market which is like all up Samsung with
08:30no seams and then the best buy one where
08:33they put in like API access for other
08:35stuff and it's a different model number
08:36might be a different color and the thing
08:38is this is particularly for voice this
08:40is just another across platform kind of
08:43thing well we'll get to whether voice is
08:44a platform itself but like you know the
08:47idea that you're just going to go do all
08:48of these presumes that we've reached a
08:50steady state and I think that's not
08:53weird because I mean there's the sort of
08:54two or three connected home questions
08:56there is see how many of these things
08:58are actually going to end up being smart
09:00or connected at all I will like to work
09:03will the door lock work will the fridge
09:05work will the cooker work will the cook
09:07a hood work will you say hey Alexa turn
09:09on the extractor fan you how many of
09:11those things will have an exhaust fan
09:12just let me know how many of those
09:15things will happen and how many of them
09:16won't you can kind of look around your
09:17kitchen and you think okay well toasters
09:19worked microwaves mostly worked electric
09:21carving knives mostly didn't work you
09:23probably wouldn't have been able to tell
09:24that in advance and that's sort of where
09:26we are with all the smart stuff yes
09:28there'll be a bunch of stuff some of the
09:29stuff that's being proposed will work in
09:31some of it Wayne and that's one of the
09:32reasons you see so many of these
09:33different things because they're all
09:34riding on the smartphone supply chain
09:36there was the error as you talked about
09:37like putting a DC motor in every
09:39appliance and there was a time and in
09:42fact in the u.s. electrical codes are
09:44actually built around this notion that
09:45you're gonna have like a lot of plug-in
09:47appliances in your kitchen so there
09:49needs to be an outlet like every four
09:50feet at counter height and you know
09:52there were like carving knives like
09:54nobody really has those anymore
09:55electric can openers how many blenders
09:58and mixers and other thing and so there
10:00was a time you know mostly in the 60s
10:02and 70s when like everything got a DC
10:04motor in the kitchen and and then we
10:07worked out which ones might end and then
10:08we realize that it didn't happen and
10:10it's also different by country so you're
10:11like everyone in Japan has a rice cooker
10:12everyone in break it has a kettle and
10:14America maybe neither of those but
10:16everyone has a coffee machine
10:17exactly so you kind of can't predict
10:19which of these toys or these gadgets
10:20will turn into something that everyone
10:22expects to have so there's that like
10:23well what will be smart and what won't
10:26then there's will they all be one system
10:28or not and will you be able to say to a
10:32lexer or to your phone or to a panel on
10:34the wall or something do this thing and
10:37the lights come on in this room they go
10:39off in that room the stove turns on the
10:41water turns on the sprinklers do there
10:43so like it's all one thing or what
10:45anyway you can just say do this and it
10:46come happiness through one interface or
10:48do you have like point solutions right
10:50the network is a door lock and unlock
10:51automatically when you walk up to the
10:53front door and this is sort of nothing
10:54else he doesn't talk to anything else
10:55right and this is where you can get
10:56really sort of confused as a as both a
10:59consumer and like looking at strategy
11:02and what to build because like it is
11:03this sort of tech geek nirvana
11:06you know it's also a big company
11:07strategy and a big companies try they
11:09want it all to be one system because
11:10then they can get you to buy tool from
11:11them right right the difference I think
11:14is is that the big companies especially
11:16those huge ones are actually they can
11:18have the strategy but they also need to
11:20sell things individually like they they
11:22can't like you can't go to Best Buy and
11:24say I need a new washer because mine
11:25broke and then they'll say well not
11:27unless you buy an air conditioner
11:28yeah and if you go to Samson fridge
11:30right right and and so they you can see
11:33that in their in their boots where you
11:35know like the fridge could be the hub or
11:37a speaker could be the hub you know and
11:39then you have to ask yourself do I want
11:40the fridge and we'll get to that in a
11:41second but the big thing is that the
11:43thing that you have to realize is that
11:44the geek scenario of like is no
11:47different than the old one of I want to
11:49just relax and listen to music and push
11:51one button on a remote that then like
11:53turns the music on turns the TV off
11:55switch is the a/v in this like the Barry
11:57White made yeah yeah I'm home and the
11:59door recognizes your with a date and so
12:01it turns the music turned puts the right
12:03music on and sets the lights to the
12:04right level and does I like the whole
12:05thing right right y'all seen power
12:07scenario exactly and whilst repairers
12:09and Barry White in one but it I mean and
12:12that like the thing is that's been a
12:13thing since like infrared programmable
12:15modes all through x10 and wiring in the
12:18lights with specialist and like that
12:20stuff is just so much more complexity
12:22than actually where I mean you you know
12:24you try to do those scenarios and you
12:26spend all your time realizing wow I
12:27can't tell if the three-way light switch
12:30which way the lights are so it just
12:33won't ever correct it so I have to go
12:34through the house and like reset
12:35everything that there's almost like
12:37there's two levels to the problems one
12:38is getting the technology to work and
12:39the other is whether anyone actually
12:40wants to do go through and make 10
12:42different choices about what the
12:43settings yeah and so you could argue you
12:45know the bearish scenario you would be
12:46all of the connected home stuff means
12:48that what was a hobbyist scenario
12:50becomes much easier to build but it
12:51might still be a hobbyists know it well
12:53and because there's there are real
12:54they're real things that that matter you
12:56know we've both talked about like you
12:58know people visiting your house and like
13:00how do they operate anything is a very
13:03odd thing like they didn't even know if
13:04you have Alexa Siri or Google so they
13:07don't even know how to start and so you
13:09can imagine this scenario where somebody
13:10is just like the babysitter is literally
13:12Google Alexa Siri turn the lights on I
13:15can't see anything I you know and it's
13:17it's sort of absurd but the idea that
13:20all of these you know that's the thing
13:21about this this market of consumer
13:23electronics is it's generally been
13:25highly fragmented because of the
13:27different ways people live globally and
13:29even you know urban and rural
13:31you know suburban and you have like at
13:33annual 15-year placements I'm crazy I
13:34mean like 20 you have a placement cycle
13:36on a fridge so yeah I don't gonna buy a
13:37new fridge just because there's a new
13:39one let's go to us color screen on it
13:40and then even then you go you know you
13:42go to you know you're gonna do a fancy
13:44kitchen remodel or something there's a
13:47good chance you you have your heart set
13:48on like one element of it like I really
13:50want a gas range that looks cool and
13:52professional the god that means that all
13:55of these companies gonna have to make
13:56all of those parts and there are only
13:57two companies that can do that which is
13:59Samsung and LG and even when you've done
14:01your kitchen that doesn't mean that you
14:03replace all of the locks in the house
14:04and so what you talk about the whole and
14:05rebuild every light switch in the house
14:07exactly and so I think that it's a great
14:09strategy and it drives a lot of
14:12interesting things but the odds of this
14:14coalescing are very low and that's why
14:17this notion and the lesson from Sony or
14:19RCA before it or whirlpool they're all
14:21they're like nobody ever came to
14:23globally dominate the home like this is
14:26the interoperability standard do they
14:27see power right and I think that that
14:29sort of means that it's interesting when
14:31you get to the now our tech company
14:33world like how much of this like what
14:35the is it really a platform or or what
14:39are we talking about here this kind of
14:40guesses to voice I think and I write a
14:42blog post about voice sort of eighteen
14:45months ago and I said this sort of
14:46there's a whole bunch of reasons why
14:47people want voice to be a thing whether
14:49that's in a box in your home or on your
14:50phone or in your whatever it is people
14:52want voice to be a thing and one reason
14:54is because machine learning means that
14:56some parts of voice now work so you can
14:59turn audio into text pretty reliably you
15:01can turn text into a structured query
15:03pretty reliably so you can work out this
15:04person is asking you to play music and
15:06you can do it the second thing is that
15:08anyone who isn't Apple or Google has
15:10made your platform envy and they want to
15:11find a way of going around that not
15:13being worrying about Apple or Google
15:14cutting them off at the platform level
15:16on the smartphone and they want to not
15:17be a commodity in whatever is the earth
15:18and so that's yeah that's II am there's
15:20a massive Amazon Facebook motivation
15:21here which drove the Kindle Fire and
15:23then now is grooving alike so and drive
15:24a bunch of stuff a Facebook and so
15:26there's a bunch of reasons why people
15:27want it to be the thing and a bunch of
15:28reasons why people think it should be
15:30the thing I think this is sort of a
15:31fundamental misconception I see over now
15:33again about voice which if people think
15:35well it's almost there but it'll be
15:36better and it'll get better and it'll be
15:38fixed and in a year or two it'll just be
15:39all existing problems will be ironed out
15:41and I think the misunderstanding there
15:43is there's actually two challenges here
15:45one is that you say Alexa hey how much
15:48longer is there on the timer and it says
15:50there's three minutes left on the timer
15:52instead of saying The Times newspaper
15:53has been printed for 175 years which is
15:56what Google home did to me the other day
15:57and that's like the voice recognition
15:59problem and that does does get fixed
16:01that will absolutely get fixed because
16:02of machine learning the other problem
16:04however is what happens if you ask it a
16:07question that it doesn't know how to
16:08answer because everything that you say
16:10to a voice assistant is basically
16:12getting plugged into an IVR it is now
16:14accurately able to recognize what
16:15question you've asked but there are only
16:1720 questions they can answer and each of
16:19those questions is somebody sat down and
16:22wrote a thing to be able to answer that
16:23question and at a certain point you
16:25start being able to scale that you can't
16:28answer 55 questions or hundred questions
16:30or 5,000 questions and Amazon has tried
16:33to solve this with skills by having an
16:34API and so now there's 5,000 skills or
16:3710,000 skills on Alexa but the problem
16:39with that is you still don't know what
16:40you can ask great this is per gets you
16:42to a common thing people always say this
16:44is a command line yes the command line
16:45has a million commands but you don't
16:47know you have to remember which ones you
16:48can run right and I think that a lot of
16:50this comes from the spectacular work
16:52that Google has done over you know the
16:54past ten or fifteen years of just you
16:57know like being able to type into a one
16:58line edit control you know vast numbers
17:01of things and what they have done with
17:02what I think is an interesting lesson is
17:04is that it started off by being this
17:07pure like we will match your question to
17:10kind of thing and then what they
17:12realized is that could only take you so
17:13far and then it turned into like they
17:16just have different teams they have
17:17specialized teams that say well if we
17:19detect this as a travel related thing if
17:21we detect this is you know a film or
17:23movie related thing if this is a
17:24knowledge related thing and then even
17:26just the appearance of Wikipedia change
17:28their ability to answer directly home
17:31pull the structured right because now
17:32you can ask how old anyone is and if
17:34they have a Wikipedia thing or where
17:35were they born and those are easy things
17:37to do but there's a basic fallacy here I
17:39think I forget who does it pointed out
17:41basic fallacy in sort of non tenant for
17:43people looking at AI in any sense is you
17:46do something and you think well for a
17:48human being to be able to do that they
17:50would have to be able to do all these
17:52therefore the Machine must have that
17:53capability right and you forget actually
17:55you're looking at a washing machine a
17:56washing machine can wash clothes but it
17:58can't wash dishes and it can't cook you
17:59dinner and what we have here you know
18:01back to our of talking about talking
18:03about connected home here what you have
18:05is somebody built a washing machine and
18:06somebody built a dishwasher and somebody
18:08built the toaster and somebody built a
18:09fridge but each one of the things that
18:11Alexa can do somebody had to build and
18:13that doesn't scale and so the idea this
18:17becomes like a universal platform or
18:19it's like it's the next smartphone is a
18:20complete blind alley
18:21what does get interesting is back to our
18:23connected home you actually probably can
18:26remember how many lights you have or how
18:28many rooms you have and like the door
18:30they turn on the lights
18:32turn up the heating turn preheat the
18:33oven you can remember those kinds of
18:35queries and so voice might be like your
18:38universal UI for those yeah but that
18:40presumes that that actually is something
18:42well that and that that's why I was
18:43actually looking for it last year at the
18:45hotel I stayed at for CES they announced
18:48that they're working on having Alexa for
18:50the rooms now this is a relatively new
18:52hotel in which all of them in Vegas are
18:54but this one they already had like what
18:56I think is probably the best central
18:59automated switch that I've ever seen in
19:01a hotel like it has six buttons on it
19:03and it's like lights drapes and front
19:07door like privacy and and clean and and
19:10it's it's wired and it's very reliable
19:13because if you tablet it it's it's not
19:16even a tablet it's literally a resistive
19:18screen on this little tabletop thing
19:20with a USB charger in it and oh and the
19:22other is the thermostat and so you push
19:25a button like drapes and it says you
19:26want the shear on or off and you want
19:29the drapes on or off and notice that's
19:30one of the first things it has is their
19:31on or off and you can push on or off not
19:34a toggle because one of the notorious
19:36home automation things is someone gets
19:38up and manually closes the drapes and
19:40then the automated system doesn't even
19:42know which direction they are which is
19:44really common with lights and the same
19:45thing on lights it has all on all off
19:47and then like dim that's it yeah and I
19:51thought oh when that that is gonna work
19:53really well for Alexa because there's a
19:55narrow domain and you know exactly what
19:56you cannot and and you don't have to
19:58notice like very much and so then I
20:00got there this last week and then the
20:02lights are and I spent like ten minutes
20:03trying to get the thermostat down
20:05because the room of course was Vegas
20:06cold and I could not get it like you
20:09know it Alexa would come back I'm sorry
20:11there's not a thermostat named
20:12thermostat and you know I would say I'd
20:15like to lower I said Alexa lower the
20:16temperature to 70 degrees and or raise
20:19the temperature to 70 degrees and it
20:20says temperature set to auto 61 degrees
20:23and I was like did it tell me what it's
20:25set to or late and I just went in this
20:27loop and then I gave up and then I just
20:29started in Alexa turn to open the drapes
20:31Alexa closed the drapes and it was
20:33flawless well this is kind of this is
20:34and I'm reading a two-page index card
20:36cheat card well this was kind of my
20:38point that there's two problems says the
20:39one problem is the thing has to
20:41recognize you and that will get fixed
20:43you know it will get to the point that
20:45it knows will always know that you're
20:46talking about the thermostat whatever
20:47you say the problem is if you then say
20:49to it hey what's a good restaurant that
20:51I could have dinner at night somebody's
20:53going to have to over in there well and
20:54that for a hotel is actually super
20:56interesting because it might use Yelp
20:59and then send me to Caesars and so like
21:02and it turns out the queries that aren't
21:05built in the hotel skills that they made
21:07like the thermostat and lights and the
21:08drapes are just going to Alexa and it
21:11turns like anything like that like you
21:13know what's the weather it's not going
21:15to the weather that the hotel provides
21:17which might come from any particular
21:18source and so the idea that that hotels
21:22would turn us all into their own like
21:25huge system is pretty complex sort of
21:28time kind of thing we kind of keep
21:29circling back to is there's the vision
21:31of the unified system I mean it's like
21:33when you watch a sci-fi movie about the
21:35far future somehow there's no
21:36applications is just one interface on
21:38the screen right right that aside the
21:39fact that they wouldn't even be a screen
21:40and you wouldn't have pilots and you
21:41wouldn't be finally they wouldn't even
21:42be any people on the ship but yeah well
21:44that was a comment I did I did actually
21:45see like the automated motorcycle driven
21:47by like a robot which I just thought was
21:49the weird it was like a Cylon ship that
21:51made you put him why have you put a
21:53droid in the back of the x-wing why
21:54isn't he - I don't know in Star Wars for
21:56everybody we've all lived through that
21:57the summer that but there's a sort of
21:59there's a sort of a separate thing
22:00there's always a fantasy that it'll be
22:02unified system the reality is you'll buy
22:04one thing now we need five years time
22:06you'll buy another thing and 15 years
22:08after that so you'll have all the and so
22:09this kind of brings us back to the stuff
22:10we were talking about the beginning
22:11which is you have the big company
22:14strategies and you have the desire to
22:15create unified systems because that will
22:17get you to buy all the staff and because
22:19that but but the reality is it will
22:20always be kind of more messy and more
22:22complicated right and that's what that's
22:23right I do think like like the real sort
22:25of advice like if you're making things
22:27it was added to things which is one
22:30which we need to go into a little bit is
22:32just how much intelligence to put in
22:33your device versus the rest of the
22:36network and then the second is which one
22:38of these sort of systems do you want to
22:41plug into so let's just talk about for a
22:43second because one of the things is is
22:44that that's driving me crazy is just
22:46like yeah there you know Samsung makes
22:48screens and they make charm chips so
22:50they seem to decide that every little
22:53device has to have a screen with an ARM
22:54chip in it and that just seems like
22:56overkill yeah well this is the UI
22:57question you know should the I mean if
22:59there's a the Razer I always use here is
23:01how many questions should the computer
23:02ask and actually so half of the stuff if
23:04this stuff means now there's more
23:06questions that you have to answer and
23:07the other half is now there's less
23:09questions that you have to answer so you
23:10could argue like the problem hold
23:11problem with quantified self and fitness
23:13trackers and so on with it was actually
23:14more stuff you had to do rather than
23:16less stuff so to the washing machine
23:17should the washing machine have a screen
23:19that asks you what you've put in and
23:21what load you want on what settings you
23:23want and all this should they take the
23:25dials and put them into the screen or
23:27should they just be a go button and
23:28nothing else and actually the ideal for
23:31the washing machine is this just a gay
23:32button the ideal for a dishwasher where
23:34you don't have to worry about like
23:35different fabrics it's absolutely you
23:37should just shut the thing and press GO
23:38and that should be in there shouldn't be
23:39any settings at all and the rest of it
23:41should be the chips problem right which
23:43the the example for me was just all the
23:45connected microwaves because for 25
23:48years all the microwaves you could buy
23:49have had like this ridiculous grid of
23:51buttons you know chicken stay frozen
23:54reheat you know and popcorn they always
23:57have a popcorn button and like it just
23:59was ridiculous because like the food you
24:01want to do is never quite there and so
24:03the question was like did this really
24:05help because the directions on the food
24:07always just say heat on high for 2
24:09minutes yeah Lee which turns out on most
24:11microwaves to be tricky to do now
24:13because they want you to use all of
24:15these things that make their microwave
24:17look unique and so I I do think that the
24:19conclusion is like lesson less UI less
24:23intelligence as the or more intelligence
24:25is the right way to approach building
24:28reduces the cogs and the maintenance and
24:30extends the longevity of the device the
24:32example I've used which is I'm sort of
24:34writing about something about about this
24:35now is there's two parts to this one is
24:39you shut your door when you leave home
24:41you shut the door the door locks by
24:42itself and that's like a 2000 year old
24:44technology when you come home in ten
24:46years time the door will also open by
24:48itself and that's based on smartphone
24:50chips and computer views like cars or
24:51cars work today a cloud but they from
24:54the users point of view it's a door that
24:56locks by itself and unlocks by itself so
24:58one of those is super advanced
24:59technology and the other isn't but the
25:00user experience you'd actually be
25:01completed the same way of those and it
25:03probably shouldn't have a screen to
25:04unlock anymore than it should have a
25:05screen to lock itself right and so this
25:07sort of that thing the other thing I was
25:09thinking about walls our kettle so
25:12certainly everyone in the UK has an
25:14automatic electric kettle you plug it in
25:15you press the button it heats it up
25:17until it's boiled and then the heating
25:18turns off and you look at this and I
25:20feel like channeling these people who
25:21look at CES and say look at all these
25:23stupid devices look at a kettle you an
25:25idiot you could buy a saucepan put the
25:26saucepan on the stove that will boil
25:27water why have you got a kettle right
25:29right which is true but it's actually
25:31better to just press a button go away
25:33and come back and the water's boiled
25:34right turn the heat off rich it's even
25:36further the the vegetable peeler no let
25:38your vegetable three you just like a
25:39kind of a five-dollar vegetable peeler
25:41yes you could peel through it with a
25:42knife the 15th time that half of the
25:44Apple has gone into the sink and a small
25:45piece of your thumb has gone into the
25:47st. you go and you buy a vegetable
25:48peeler right about it and I think that's
25:50a way to think about a lot of these
25:51things that they remove friction and
25:53they remove a question or they remove
25:55something that's taking a tiny piece out
25:57of your life every single day and yeah
25:59you could boil water on the stove yeah
26:01why have you not just got a kettle why
26:03does the door lock have a screen it
26:04should just unlock right well why does
26:06it cover me we have a kettle here that
26:07has 15 different temperature settings
26:09based on what kind of tea you're making
26:10like just really yeah press the boil
26:12button and then pour the water if you're
26:15making things you like the screen you
26:17you want to have the screen and you you
26:19can actually be different and this is
26:21something of course that Apple tends to
26:22do better than anyone which is be
26:24minimal about things like you know if
26:26you're making a Windows laptop it's like
26:28cool to continue to have all the legacy
26:30plugs and all of the different variants
26:32of adapters and plugs built into the
26:34machine and Apple makes these bold
26:36choices that really do move things
26:38forward I mean it was just the
26:39anniversary of the MacBook Air
26:40and you know I was at that event and I
26:42sat there with all the famous reviewers
26:44and everybody talking about there's no
26:46Ethernet what are we gonna do because
26:47back then you know basically you and you
26:49know Apple at that demo Steve Jobs went
26:51through a lot of lengths to talk about
26:53how to use a DVD player or actually a
26:55cd-rom at the time and it was like air
26:57CD or air driver settle the next tunnel
27:00drive as well but yeah yeah a disc yeah
27:02air disc or something it was called and
27:03it was like trying to comfort everybody
27:05and I just love that it's exactly the
27:07same as I'm shipping a headphone adapter
27:09with the iPhone right right and it just
27:10doesn't these things this that life does
27:12move forward and I think that's really
27:14important for everybody to sort of
27:16recognize as as a key takeaway finally
27:19just really want to touch on this point
27:20about being cross-platform and whether
27:22you should embed the elects the runtime
27:24or the cirrie runtime or the what you
27:26can bet or the Google runtime which is
27:28was Google's hey Google strategy which
27:30was everywhere like they spent a lot of
27:34money to do hey Google everywhere but
27:36part of what they did is it's the
27:37Android thing where it's just it's free
27:39to make your commodity device be
27:41different you buy a small amount of
27:43chips and you get the free runtime like
27:47does that make sense to you that
27:49everything should also embed like be
27:51basically the alexa microphone at
27:53endpoint yeah it's sort of weird i mean
27:56it comes back to this you know you want
27:58to make it a thing google want
28:00everything to be an endpoint for google
28:02home amazon want there to be amazon home
28:04everywhere doesn't necessarily make
28:06sense to a person that everything should
28:09be that endpoint i mean this is you know
28:10the joke that you'll ask a question and
28:12you'll have Google and Alexa arguing
28:13with each other right on your kitchen
28:15counter as to which of those is the
28:17thing I mean the son we actually haven't
28:18mentioned we talked a little bit about
28:19voice earlier there's a lot of
28:20experimentation here there's
28:22experimentation around what devices will
28:25have smart and what kind of smart and
28:26what UI and what system they'll be
28:28connected to and is it voice and what
28:31kind of voice and voicing what devices I
28:33think there's also just near going back
28:35a level to the sort of the Anxi to what
28:37actually is a super high level strategy
28:38but certainly for Google probably for
28:40Amazon as well which is what do we do
28:42with a rolling learning what's the
28:44product we've got this kind of
28:46fundamental capability that computers
28:48can answer these new kinds of questions
28:49what's the product well the most obvious
28:52immediate products are right now you can
28:53image consumed image recognition so you
28:55have google photos which is fantastic
28:57and yeah we can do voice okay well what
29:00could we do if we recognize voice well
29:02maybe we could answer questions let's
29:03try that let's see if that works and you
29:06can see in this I think an echo of
29:08Google now which is Google now is sort
29:10of well we've got all this information
29:12what other product can we create out of
29:14that and Google now for a bunch of
29:15reasons didn't work and they kind of
29:17shifted that to be okay well now it's
29:18Google assistant so instead of it being
29:20passive it'll be active and you'll ask
29:21it a question but there's a lot of sort
29:24of almost kind of top of the s-curve of
29:26one s-curve beginning of the next
29:27s-curve of like we've got these primary
29:30technologies in this primary set of
29:32information whether that's the arm ships
29:34or the frames or the microphone or voice
29:36recognition and we're sort of
29:39experimenting well what would the actual
29:41consumer proposition for that look like
29:44what would it be right is it a voice box
29:46is it glasses is it like is it a text
29:50assistant is it a bot it's like we don't
29:52quite kind of a resolution on what
29:54that'd be and I think that the key thing
29:55for people to understand when they're
29:57making things is is that we're at the
29:58very beginning there's a lot of
29:59experiments but you can't have your
30:01product rely on like an experimental
30:04approach because especially if you're
30:06making something in business that needs
30:08to have an answer or in the consumer
30:10space that the lights do need to
30:11actually turn on you know like the thing
30:14about the Google search experience is
30:15that it was highly resilient to being
30:17wrong like if you typed a query and you
30:19didn't like the answer you typed another
30:21one and you didn't lose anything Google
30:23now the problem is is that if you start
30:25to rely on it to tell you when to leave
30:27for a meeting and what traffic it has to
30:29know a lot of other stuff in order to
30:31get that right all the time and the cost
30:34of being wrong just once is very high
30:36like it could be the job interview or
30:38something like that that you miss and
30:39this is like literally the biggest
30:41lesson from Microsoft Clippy which was
30:44like automatic features and automation
30:46are all cool except they kind of have to
30:49be right literally all of the time all
30:51you should kind of you shift your
30:53expectations so the thing about Google
30:54search is they don't give you one result
30:56they give you 10 10 so it's actually
30:57manually created when you're saying turn
31:00the lights on they can't give you a list
31:01of 10 things that they think you might
31:02have meant right and also even again
31:05with Google search like they give you a
31:07if you don't like it you can click the
31:09two at the bottom of the screen or type
31:10a different thing in and you're you
31:12don't lose out anywhere and you're
31:14instantly cuz if you typed in by a
31:16hundred shares of like G which is one of
31:18the first voice things I did on tell me
31:20I asked for tell me the stock quote for
31:22a Cisco and I'm sitting there on the 520
31:25bridge you know on my analog phone and
31:27it's giving me the stock quote for Cisco
31:29the company that makes like kitchen
31:31supplies and I could not get the stock
31:33quote for Cisco and like if I were
31:36telling us buy or sell that would be
31:37really bad so there's a lot a lot here
31:40for makers to really ingest hmmm and a
31:42lot of sort of early experimentation I
31:44mean one of the ways I talked about IOT
31:46is like the smartphone supply chain on
31:49the one hand means there are these all
31:51these cheap commodity components and
31:52suddenly you can make any connective
31:54thing you want really much much easier
31:55than you would have been able to in the
31:56past on the other hand ubiquitous
31:58wireless and cloud means you can kind of
31:59light them up in ways you would never
32:01been able to in the past what we don't
32:03have is like the really clear answers to
32:05all what would you create so it's always
32:07come in the opposite direct right
32:08previously you'd have the vision of the
32:10thing you wanted and there'd be like a
32:1110-year march to get the tehnology
32:13mailed it right now all the technology
32:15is there for almost anything you can
32:16think of and people are trying to work
32:18out what is and also you can make it
32:19very inexpensively you know there's 100
32:223d printers to choose from and so I
32:23think that for me the big lesson to walk
32:25away from is just that there's still
32:27just as rich an environment for making
32:29the right products and the great
32:30products and they are all changing our
32:33but it's not happening as fast as people
32:35can write blog posts about like this is
32:37the future be here and that's
32:38frustrating people and I think as you
32:41know as people building companies the
32:43key is to really focus on what it is
32:45that the problem you're solving and not
32:47worry about like trying to discern a
32:49galactic strategy across all of these
32:50things that there's still it's all about
32:52building great products that improve
32:54lives exactly there's like specific
32:55problems you sell specific questions
32:57that somebody gets asked every day that
32:59you can take away well thanks everybody