00:00:00
[Music] hey what's up everybody welcome back to another episode of the waveform podcast we're your hosts i'm marquez and i'm andrew and today we've got a
00:00:14
couple pretty new pieces of tech to talk about but we're gonna start off with some uh
00:00:19
random tesla autopilot subscription chat because we wanted to go over like the logistics of it what we think of it but also we got some new pixel six info as much info as we could possibly get
00:00:30
about a phone that's technically not been announced yet it's kind of a weird announcement yeah they did a strategy
00:00:35
strange very strange but we got tensor and we're going to talk with david about that new silicon strategy from
00:00:41
google but first andrew break down this tesla autopilot subscription price thing okay we we wanted to talk about this last week but i forgot to put it doc so we'll talk about
00:00:53
it this week we're about a week late on this but um i'm gonna break it down and you're gonna since you're the tesla owner if i'm kind
00:00:59
of miss missing anything just sure just correct me when it's possible but uh i think it's safe to say
00:01:06
that for a very long time people have been uh asking for subscription service for full self driving
00:01:12
um also before we get into this i know there's a lot of debate about what full self driving is if it's worth
00:01:18
the money we're not doing that at all today we're just talking about how it's a subscription base they finally added it um just as the
00:01:25
future currently exists people wanted the ability to subscribe to it for some time and then
00:01:30
unsubscribe from it when they don't need it for a while or and then add it back later with no penalty and also not paying ten thousand
00:01:37
dollars up front like that's a lot of money up front so monthly i mean we see this happening with everything financing it's just
00:01:43
always been a thing in our uh our economy where you try and pay per month rather than upfront so
00:01:49
people have been asking for for a really long time it finally came out and it is currently going for 199 a
00:01:55
month or 99 a month if you already paid for enhanced autopilot so i guess that's my first question
00:02:00
really quick is okay you can have a basic base model with nothing you can have enhanced autopilot
00:02:06
and then there's full self driving yeah correct what is the price on uh enhanced autopilot
00:02:12
that's a good question i'm gonna guess five i'm gonna go to tesla's website and check that's just the link i'll spec a model
00:02:19
three that's probably what i did people expect a base one yeah so this is perfect if you compare it like that
00:02:25
i can compare it to what i have already so i don't think it lets you buy enhanced autopilot by itself now
00:02:30
it's either nothing or full self driving for 10 000. okay that's it all right so still a little confused on
00:02:37
that apparently if you haven't advanced enhanced autopilot already it's only 99 right but we're going to mostly focus
00:02:42
here on just the 200 a month because i think that's what most people will be in the scenario of okay um like you said it's cancelable
00:02:49
any time um they don't prorate it per month but if you cancel it you just at the end of the month it goes away and then you stop
00:02:55
paying for it and you can do it up again whenever you'd like so at 200 a month that comes out to about 2
00:03:01
400 a year so the first thing i think you kind of do the math on is at what point does this equal ten thousand dollars because yeah that's what you're essentially trying that's what you're avoiding is paying the
00:03:11
upfront yeah everybody's doing the math about okay if i spend ten thousand dollars up front will i eventually use autopilot
00:03:18
for enough months that it will be worth it where if i had paid a subscription it would have been more than ten
00:03:23
thousand um and i guess the question that really affects all everybody's math is how much
00:03:28
do you really use autopilot yeah a lot of people think i use autopilot all the time i don't actually use it that much but i
00:03:36
do use it in specific situations that are sort of random throughout the year meaning when there's traffic so a lot of
00:03:43
people like there's a full self-driving beta that's out now that's really really impressive and a lot of people
00:03:48
are getting this slow robot where it'll take you from point a to point b stop at stop signs navigate through city streets take
00:03:55
turns exit highways all these intersections and everything and it's super cool and i'll try it for the beta but i
00:04:02
generally actually enjoy driving i don't know i'm some weird sicko that actually likes holding the steering wheel
00:04:08
so i'm actually steering what the steering yoke steering wheel or whatever but i actually enjoy driving so i don't
00:04:14
for the most part use autopilot but the one part of driving that is annoying no matter what kind of car you're in is like stop and go
00:04:21
traffic or like slowly rolling through the lincoln tunnel at eight miles an hour where it's like not
00:04:26
really driving anymore you're just kind of sitting and waiting to get through it at that point i'll turn on autopilot
00:04:31
which really isn't all that different from any other car's cruise control just holding you in the lane moving you forward when traffic moves
00:04:38
fine is that to me worth 200 a month or 10 grand up front uh to me part of the the tesla experience is having the tech to be able to do that
00:04:50
so i'm gonna say yes for me because there is enough traffic around this part of the country that it happens all the time but i'm not
00:04:56
somebody who's like literally always using autopilot all the time with tons of miles so here's the thing though
00:05:02
uh you say you're not using it all the time but it's easy to say you use it once a month right yes so like in terms of being like
00:05:09
cancelling it when you don't really need it and grabbing it again when you do need it that's not like really an option you're probably i would
00:05:16
say for this you're either doing it all the time or maybe you're doing like a trial run of it
00:05:21
for a month or two yeah um so i feel that's that's it's cool that you can cancel it one other i don't see a lot of people canceling it bringing it back cancelling it bringing it back um yeah
00:05:34
there might be some super niche opportunities there when that happens but i think it's smart for tesla because there's going to be a lot of people who
00:05:41
don't pay the 10 000 up front thinking they won't use it that much and then one day they'll try it and a certain percent of
00:05:49
those people will like it so much that they convert into paying every month for it and so that'll it'll end up making tesla
00:05:56
money from people who weren't going to try autopilot at all ever which is pretty good for them but there's also a group of people who
00:06:02
were going to pay for autopilot but have done the math on how many months they think they're going to need it and use it
00:06:07
and they're going to give tesla less money i'm sure tesla's done the math and landed on this price and figured this is
00:06:12
good but yeah i think the the overall like mission of getting more people used to autopilot and air
00:06:19
quotes i'm doing air quotes for audio listeners uh i think that sort of helps it i think um
00:06:25
the biggest issue i see with it is and i know tesla has to think this as well but it over
00:06:31
four years you are now paying that's when you equal about ten thousand dollars four years okay four years i
00:06:36
would argue most people keep their cars for longer than four years especially if we go into like i financed
00:06:41
a model three here just to take a get a guess of it so i did five thousand down for 60 months i think that's that's what i did for my
00:06:48
last car i think that's pretty pretty typical if you're financing a car um so at the model 3 starts at 40k
00:06:55
and you put 5k down for 60 months you're paying 642 a month so the nut 200 is like a twenty percent
00:07:02
increase on that yeah uh you gotta really want autopilot yeah you gotta really want it and you're also doing it for
00:07:08
sixty months so if that's twelve thousand dollars if you do full self driving the entire time so you're now paying
00:07:14
two grand more than just paying for 10 up front yeah i'm just trying to think maybe there's some customers that are
00:07:19
like i'm gonna do a road trip in june of this one year i don't really use autopilot that much but you know what for my cross-country road trip in my car i'm going to buy autopilot for that
00:07:31
month and so they'll just pay the 200 bucks to have it for that month and then stop and then it'll go away and they don't use it that much and maybe they love it so much that they'll keep paying for it after that trial but that's a
00:07:43
that's a possible use case yeah i've been so what have been my my initial thoughts on it where i
00:07:49
love that they did full self drive or the subscription model i think a lot of people asked for it and they wanted it did they price this too high to where
00:07:56
they're essentially almost there they're making you want to do the ten thousand dollars
00:08:01
because it is generally the better deal assuming you have your car for more than four
00:08:06
years which we can agree is the majority of people yeah um also an interesting point uh full self driving doesn't transfer with you if you move to a new
00:08:18
car i had that question so if you buy a tesla let's or or lease or whatever refinance you get a
00:08:23
car you pay the ten thousand dollars and then you sell that car you have to pay the ten thousand dollars
00:08:30
again on your next car to get some full self-driving with it so it's like a feature of that single device it's
00:08:36
funny because most subscription services aren't local to a single device they'll apply across everything if i pay for spotify
00:08:43
i didn't just pay for spotify on my one laptop i pay for it to access anywhere so it's kind of funny that this one
00:08:49
subscription service does not transfer um but yeah if you do pay for the subscription
00:08:55
on one car and you want it on your next car you unsubscribe on one car and then re-subscribe on the next car
00:09:00
and so when you sell it you're selling a car that does not have autopilot yeah which is bringing the value down and that's something to to notice do the
00:09:06
majority of these cars have the hardware already built in like you don't get the extra hardware when
00:09:12
you pay the 10 000. they all have the hardware okay so like that to me feels kind of strange paying
00:09:18
the 10 000 and then having to pay again where i guess that's where this could come in a little better
00:09:24
um let's say you are somebody who might only have the car for three years then it definitely makes sense to just
00:09:29
do the 200 my other question would be can you pay the 200 bucks to try it for a month and then pay the ten thousand dollars because you loved it so much and you know you're gonna have the car
00:09:41
for more than four years and you want to have the better deal that's a good question i would guess that because tesla should take your
00:09:47
money but actually don't know i don't know you could in the past if you before the subscription started you could
00:09:52
add it later right uh weren't they charging it would cost more yeah it would cost more so if you
00:09:58
did it up front from the factory it would be for example 8k if you didn't it would be 10k after the fact
00:10:04
but the hardware was always there it was always like okay now we have to go through enable software and get this new feature i think my final my
00:10:11
final thought on this is i think it should be 100 a month and i think that averages out to the typical
00:10:17
lifetime of a car better um and then while still has the potential of making more money than the 10 000 up front because there are people who keep their cars for 10 or 12 years
00:10:27
and and we're in this uh age where we are assuming electrics are going to last longer than gas cars in a lifespan because they don't have all the wear and tear on
00:10:38
a regular transmission and an engine i think a hundred dollars is great i kind of think the 200
00:10:44
is like they didn't really want to go full blown into the subscription model but so many people were asking it
00:10:49
so this to them is like a well if anyone wants it that bad we're gonna make quite a bit of money off of it yes one of the cases where
00:10:55
there is no competition like there is no alternate version of this a subscription service to
00:11:01
a adaptive cruise control version where you can go oh well this one's only 80 a month so i'm gonna get this one
00:11:06
instead also if you own a tesla there aren't even if you had other versions on other cars it's not like you could buy
00:11:12
mercedes cruise control so there's definitely no subs no there's no competition internally yeah to sort of push that price one way or the other which is interesting
00:11:23
but yeah i'm trying to think of if there's some other uh feature in in the tech world where you
00:11:29
already have the hardware capable and you're just paying to unlock some software to take advantage of that hardware like all the tesla cars have the cameras and the
00:11:40
the yolk and all the sensors around the car but only some of them actually use all
00:11:46
of that for this certain self-driving future you could think of like peloton or something like that where you have this bike that has
00:11:53
a multitude of options of how you use it and the best way to do it is by buying the subscription to the application and getting the classes and getting the leaderboards and stuff like that i mean it doesn't
00:12:04
necessarily make the bike work any different but it's it's the only way to use the peloton
00:12:09
you can't even just like free ride without encryption all you can do is free ride okay yeah i mean you still can use it to
00:12:16
the exact same potential sort of in like in terms of the exercising goal of it but you lose the classes and the leaderboards and the thing that kind of motivates you to want to do it
00:12:27
yeah it's weird i feel like my final thought on this would be if you're purchasing purchasing a tesla
00:12:33
and you really really want full self-driving if you can't afford to pay the 10 000
00:12:38
right up front i would probably not go for the subscription model because i'm guessing you're going to keep it for
00:12:45
long enough to where you're going to be paying more eventually and i would try as hard
00:12:50
as you can to pay it up front even though clearly ten thousand dollars is a lot of money up front yeah i feel like the subscription model
00:12:57
isn't quite there for the the majority of the people who can't afford to pay ten thousand up
00:13:02
front it's not benefiting them yeah we as far as we know because we don't know how many how many people are going to love
00:13:07
autopilot versus are going to pay it and then regret paying it who knows watch as many youtube videos as you can of autopilot
00:13:14
in action that's the best way to know because like unless you can borrow somebody's car and use autopilot with
00:13:20
you know a couple of trips and figuring out if you want to pay for it or not but watch a lot of videos why doesn't tesla
00:13:25
do a like a 30-day free tr when you buy your car you get a month of full self-driving for free and then at
00:13:30
the end of the month you can purchase it up front yeah the car arrives that would make people so mad but if the
00:13:36
car arrives with all the full self-driving features you get really used to them and then a month later all of them go away i mean i'm sure
00:13:42
people would get annoyed at that at first just because it's the option just like how people were annoyed when like remember in florida there was that
00:13:48
hurricane and they unlocked like a little more of the battery or whatever and you could drive faster but ultimately that's not
00:13:54
really that big of a difference those specs were what you paid for already um but getting a free trial of
00:13:59
it might make people wonder if they and it'll it'll stop a lot more regret of paying 10 grand and then
00:14:06
realizing they don't like it or don't think it was worth it yeah give them give them a free trial that they can
00:14:12
activate at any time so you're being way more lenient they don't start with it but at a certain point you'll go you
00:14:18
know what i want to try it then as soon as you try it that's your free month and then after that you'll have to pay
00:14:24
that's how i would structure it if i was running the program that's the that's a nice way of doing it
00:14:30
yeah all right well let's take a quick break we're going to come back we're going to have david on and we're going to talk all about the new pixel 6
00:14:36
what we know and what we think we know and all about tensor be right back this episode of waveform
00:14:42
is brought to you by mini cooper se so you may or may not know since i try to be subtle about it
00:14:47
uh we really enjoy electric cars here at waveform and the mini cooper sc is charged for the city and ready to
00:14:52
spark up your drive so we've spoken about the usefulness of compact evs on the podcast before
00:14:58
and i think this one probably takes the cake it's a zippy little car designed to be as stylish as the city you drive in
00:15:03
and unlike some other evs the mini cooper sc doesn't look like it's straight out of a sci-fi movie it still maintains a
00:15:08
classic design that won't look out of place driving the streets in new york city yeah i like i appreciate all the uh the
00:15:15
different crazy looking evs that we're seeing these days they're always fun to look at but whenever i think about buying one i i definitely appreciate
00:15:21
seeing just yeah well you being more low-key as like a daily for me for sure uh driving the cooper essie also
00:15:28
was surprisingly comfortable just because it handles nicely like a mini cooper uh like a go-kart kinda actually and
00:15:34
maneuvering through traffic on the way to lunch was a breeze so get up to 80 charge in just 35 minutes with fast charging at a level
00:15:41
three dc charging station or just charge overnight at your home with ac charging as well plus the electric digital instrument
00:15:47
cluster puts all the important things right at your fingertips on the inside there so it's a unique looking circular
00:15:53
touch screen display that makes listening to your favorite music or podcasts on your morning commute as simple as can be
00:15:59
plus the lighting around the screen moves with the dial as you turn it so as you turn up the temperature of the ac
00:16:06
or turn up the fan speed or music volume the lights light up too which is pretty sweet and it's also pretty refreshing having real
00:16:12
knobs and buttons and switches in the car so if you're in the market for a compact stylish quick little car to get you
00:16:18
around the city then it's worth checking out the mini cooper sc to see if it's right for you all right we're back let's talk pixel
00:16:25
6. yeah there's a lot we aren't going to do it but you are going to talk about pixel six i mean i'm excited for it but i've said kind of what i wanted to i know david's very excited
00:16:36
to talk about not just pixel 6 but tensor and a whole lot of other stuff so i'm going to actually relinquish my seat
00:16:41
okay for the rest of the episode um david will be your host for the rest of this and i think he's very very excited
00:16:47
so before you leave though sure since you're a current pixel user what are your like reactions to seeing
00:16:53
pixel pixel six uh i want the smaller one okay i think there's if i get it if if interesting we don't know a lot yet we don't like price okay let's bring in david
00:17:06
oh hey welcome this keeps happening david david you keep taking over the podcast
00:17:12
man welcome back thank you um okay we've got a bunch this is maybe the strangest
00:17:18
smartphone pre-announcement i've seen in a while so so what google did was they they hopped on well first of all they emailed a bunch of people behind the scenes including me and they were like hey wanna come to new
00:17:30
york city and just like check out the pixel but also we're not announcing it yet but yeah you want to check it out and then you can talk about it
00:17:36
and i was like yeah okay i'll take a look and then yeah we went out i saw the pixel 6 and pixel 6 pro
00:17:41
got lots of hands-on pro not ultra but exactly not an ultra i didn't get to shoot any video or photos of it
00:17:48
um but i got my demos and then came back here and then made a video talking about it and then
00:17:53
yeah google spills all these details but not a whole bunch of like official specs just a bunch of details about pixel six
00:18:00
but it's not the announcement yet it's there's no price you can't order it they're gonna announce it later yep even
00:18:06
though they just revealed it this fall it's weird yep um so here's what they revealed so we've got the two designs
00:18:11
you've probably if you haven't seen the leaks by now they look pretty accurate it's got the camera bar across the back
00:18:16
instead of the camera circle uh there's a six and a six pro that are similar in a lot of ways but
00:18:22
there's a few differences the six has a 90 hertz display instead of the six pro which has 120 120hz
00:18:27
display the 6 has a flat display with slightly bigger bezels and the 6 pro has a little bit of the the waterfall display going on over the edges and will give you a slightly bigger 6.7 inches
00:18:38
instead of 6.4 slightly bigger battery of course in the pro as well but i don't have the sizes
00:18:44
in front of me here and then you will get a 4x telephoto camera in the pro in addition to the cameras they share which is the primary and the ultra wide and the colors are all like pretty
00:18:56
pastel-y there's a there's a gray one there's an orange peach-ish looking one there's a blue-green
00:19:02
um but that's basically it we've got the the camera bar is sort of the most obvious new thing but the one thing that i
00:19:09
wanted to focus on that i talked about in our in my video about it uh was the new chip inside inside of
00:19:15
both which is called google tensor so yeah uh we have a lot of thoughts on tensor and maybe its implications and i made the video of course you can check that out already um but how would you how
00:19:26
would you describe what tensor is because it was rumored for a while yeah okay um i i'm a little mixed on the naming i
00:19:33
don't think the naming is going to matter that much for a lot of people like i don't think regular people are like oh apple
00:19:38
a14 i want to buy it because of the a14 yeah i do actually think the name is is probably good um for people that
00:19:45
don't know google also makes a machine learning or deep learning library called tensorflow
00:19:52
yup right and a tensor is basically it's basically just like a higher order vector so you've got scalars as a point a vector is like a to a like one dimensional point with
00:20:05
direction and then you've got a matrix and then you've got a tensor yeah so a tensor is basically just like
00:20:12
we can in n directions flow this information with an input and come out with an output it's basically deep learning so that's how they named it yes that's why they named it tensor so
00:20:22
it is we were hearing about the white chapel project before and obviously there are other smartphone manufacturers that design their own
00:20:29
silicon in their smartphones i mentioned the exynos chip that's designed by samsung um
00:20:34
uh kieran or kyren i always okay kieran chips from huawei people always correct me when i say it wrong yeah uh and obviously apple with
00:20:41
their a14 so yeah the tensor chip will now be in pixels yeah makes it competitive in its
00:20:46
own unique way but also allows them to focus on dedicating certain parts of the chip to certain things
00:20:51
and this is a lot of what they were telling me they gave me some demos on photography computational photography is a big deal
00:20:57
they're doing something new now where if they take a photo of a human and the face is blurry they're also now
00:21:04
taking some really really fast shutter speed shots on the ultra wide that they can merge in to get a clear face on an otherwise
00:21:11
blurry subject yeah pretty cool stuff also computational photography inside video yeah is something they showed me and they gave me a quick demo of basically like an hdr shot right into the sun yeah against
00:21:23
pixel 5 and iphone 12 and of course it's doing all its pixel magic so um you know that was tailored by them
00:21:29
and of course we'll have to test that for ourselves but that's pretty cool and then on device speech to text
00:21:34
recognition which covers a lot of things it covers like obviously just texting somebody with your voice but it also covers
00:21:41
google assistant and it also covers uh like turning on transcriptions yeah or translate and also live transcriptions that are also live translated into a different language all
00:21:51
happening on device a lot of really cool stuff yeah um we don't know the rest of the details of
00:21:56
like how it'll benchmark and if it's equivalent to a snapdragon triple eight yeah or us an eight seven sixty five we
00:22:03
don't really know yeah but we got all these details now and we've got thoughts on them but yeah i'm curious
00:22:10
how do you feel i have so many thoughts okay so okay here's the thing uh big reason why apple is transitioning from intel to arm for example on the max on the max is because they're
00:22:22
realizing that arm was originally developed for like smartphones right it's reduced
00:22:28
it's like a reduced architecture that was originally like let's save as much power as we can and
00:22:34
then over the years because smartphones are like one of the most replaced products of all time they were able to iterate it on it more
00:22:40
than like any other thing right so arm just became better and better and better apple realized like
00:22:45
oh these are now basically faster than the traditional processors yeah this is a similar transition it's also
00:22:53
arm of course you know because qualcomm uses arm architectures and then it basically like
00:22:59
tweaks it to its own specifications but google here is effectively realizing that machine learning is becoming more and more and more important in our phones right that's the
00:23:13
reason that apple created the a14 bionic the whole like bionic right because there's certain parts of that
00:23:19
chip that are dedicated there's like a neural core and there's the part that's this face id
00:23:24
and all these different biometrics and and often yeah ml gets looped into that right so
00:23:29
the neural processing unit so if we pull this back to the original google pixel right pixel one that was a classic pixel one a classic phone it was the first made by google phone
00:23:40
that's right that's when they announced the made by google brand they released the pixel one they
00:23:45
released the google home and then they also released i think google wifi at that time um google wifi
00:23:51
at the time didn't have google assistant but more important than the hardware that they released
00:23:56
then google made phones was the google assistant google assistant was their first step in
00:24:03
doing basically like deep learning stuff on your device and more importantly ambient computing google's future is ambient computing which is to say you don't have to
00:24:15
interact with your device that much right everything that you should be doing should be able to be done
00:24:21
just with your voice or it should predict things for you it's kind of like that elon musk like
00:24:26
yeah any input is failure of the system basically google wants the google assistant to be all around you at all times and that's why now if you go in the google store they offer so many products
00:24:39
that are all made to be all around you all the time yeah yeah there's a there's a bunch of parts
00:24:44
of this that i'm curious about obviously we have other assistants that work in their own ways siri and alexa and what's the one that's uh
00:24:53
cortana there's another one but google assistant yeah google assistant to me tends to be the most consistent and the most helpful
00:24:59
because it knows the most about me but also it's the most proactive a lot of times it's it's already about to tell me what the traffic is to work before i
00:25:05
even leave right stuff like that is really cool um so i i imagine tensor you know putting your own
00:25:11
design silicon in the phone you make is a step to sort of centralize and maximize this advantage that they have right like siri on the iphone while it's decent it
00:25:23
as currently constructed i feel like it can never match what google assistant does because apple's so focused on privacy that
00:25:29
there's so much that they don't know about you and they try to map it out and do a good job about that but google
00:25:35
sort of shamelessly knows a lot about you and is constantly ambiently computing with that information right
00:25:41
so tensor is just another way of optimizing and speeding up and and centralizing that right on the pixel and making it
00:25:48
does it make it a buying point for the pixel is that the reason people will buy pixels okay here here i think is google's core
00:25:54
problem okay uh is it will make the experience way better because all of all of the reasons to buy pixel are all of the things that google
00:26:06
assistant does and the ways that pixel are smart right like or not even just google assistant
00:26:12
but the machine learning chops that that's what it has yeah so i was just going to say that not to interrupt the camera on the pixel
00:26:18
is a feature that people care about a lot yes the reason the camera is good is because of google's software prowess
00:26:25
yeah with machine learning yeah okay yeah because like computational photography
00:26:30
everything is kind of using machine learning right like it is it is using machine learning to know if that face is
00:26:35
blurry it'll always take another photo with the wide camera and then it uses machine learning to clear up you know it'll crop
00:26:41
in with the wide camera clear up that area with machine learning and then make the face not blurry it's
00:26:47
crazy um but they have this like whole ambient computing push where they don't want you
00:26:52
to have to interact with your devices that often and when i reviewed the like pixel 4a and the pixel 5
00:26:58
i basically was like well it's okay that these phones are like not pro phones in
00:27:05
a way because they make you not need to touch them as often you've got things like now
00:27:11
playing on your phone so you glance at your phone when you're in a coffee shop and you already know the song that's playing one of my favorite features
00:27:18
yeah you've got um hold for me so if you're making a call it can like hold for you you you've got call screening so you don't
00:27:25
have to really touch your phone that much like google's whole thing and if you go on their website right now
00:27:30
they have so many products that you can buy that all have google assistant built in now the nest wi-fi which used to be google wifi has google
00:27:37
assistant built-in you've got the nest cams there's just like all of these products with google assistant so
00:27:44
i think their end goal is google assistant should be able to sort of do everything for you and google has just
00:27:49
been so good at deep learning compared to every other company ever because they have the most data
00:27:55
because they're google they have everything they have the internet's they have the internet google knows a
00:28:00
lot it yeah yeah a lot not even just about you but just about like trends and everything
00:28:05
and like about what i'm likely to do and likely to become yeah it's kind of scary it's like funny we were checking the
00:28:11
other day about like oh how many cities have we been to how many states have we been to and we're like how do i even
00:28:16
figure out for myself how many places i've been to i could just sort of count back in my memory or you could open the google maps like
00:28:23
timeline and it'll give you a bunch of points on a map where you've ever been yeah because you have your phone on
00:28:28
youtube it's like they know so much about you yeah um okay so do we just so to go back to temperatures in the pixel yeah um pixel 6 and pixel 6 pro i'm gonna guess
00:28:41
what i think the prices will be okay i just the way they were talking about them and the way i saw them built and after
00:28:48
holding them i feel like we can say or at least i'll say i think pixel 6 pro is going to be right
00:28:54
around a thousand dollars and i think pixel 6 is going to be right around that 750 mark where we saw like
00:29:00
uh you know oneplus 8 and galaxy s20 fv and a lot of those phones which is you know it's it's expensive
00:29:08
but it's competitive-ish in that way yeah uh do you think this brings pixel to like the finally competitive flagship landscape that we've been
00:29:18
hoping for or is it going to take more than just tensor so yeah this is the problem is like google is so focused on making the
00:29:24
experience for you amazing um and seamless their whole thing is ambient computing which means that
00:29:31
you shouldn't have to worry about things things should just happen for you and the experience that they offer on
00:29:37
pixel with things like their deep learning algorithms is so amazing unfortunately from a marketing perspective it's really hard to market that like you can say like pixel is
00:29:48
delightful a lot of people will call pixels delightful but i'm like helpful like google uses that term all the time but unless you
00:29:54
have like a headline feature like the camera right everybody uses the camera everyone knows
00:30:00
pixel cameras are great so i really hope that they're really pushing how much better these cameras
00:30:06
are now especially since they're finally putting new um they're putting new uh new hardware hardware in it oh yeah
00:30:11
thank goodness i'm excited for that very yeah yeah yeah that's a good point like if
00:30:17
they have like a a really catchy maybe not just slogan but like an entire campaign cause they've they've flirted
00:30:22
with this they've talked about like how i don't know if exactly the way they phrase it but how like google works for
00:30:28
you yeah or like let google let google do it do it for you yeah and it's like it makes sense to me
00:30:34
because i know what they mean but i just i just don't see it yeah a lot of people you know what's funny though is
00:30:39
they they actually they had their earnings call recently and they told the investors by the way we're going to be spending a
00:30:45
lot more money than usual near the second half of the year for the holiday season which means they're going
00:30:51
to market the crap out of pixel 6 and pixel 6 pro i'm very interested i think they really really want to get past that like single
00:30:58
digit like penetration number in the u.s market share um but yeah when i say that it's a
00:31:06
similar that tensor is a similar transition as from like x86 to arm it's because like
00:31:14
companies like qualcomm have upped the amount of machine learning cores that they're dedicating in their
00:31:20
chips like qualcomm is doing a good job of like scaling that up more and more but i think google is realizing
00:31:27
sort of like single core cpu performance and multi-core cpu performance are fine
00:31:34
but we've kind of gotten to this point where the chips are fast enough that you don't really notice slow downs
00:31:39
so let's go over here and just like put a ton of resources into the thing that isn't quite there yet for everybody right like just allowing your phone to seamlessly
00:31:51
do all the deep learning stuff as seamless as possible and then tensor is like like you said in
00:31:58
your video tens are so important because you can scale it up to bigger hardware you can scale it down for
00:32:03
things like watches and then google also cuts out the middle man they can make it cheaper and so i'm when you say when you did the
00:32:10
pricing guess i'm a little mixed i'm like they they've gone through this right they've done the
00:32:16
premium pixel where they charge a thousand bucks they've done the cheap pixel where they charged uh 700 349 for the a series i mean for
00:32:24
the a series that's the baseline for the a series okay but then if we're talking about like pixel five yeah that was what
00:32:30
79.99 that lost the 6.99 or 799. i think it launched at 7.99 and then obviously dropped i have a feeling they're going to try to make it
00:32:41
8.99 for the pro um okay and 6.99 for the non-pro cause i think a 250 delta is kind of a lot for people especially since the only thing you're losing you're only losing a couple things on
00:32:54
the smaller one yeah that is true like when you think about putting those phones next to each other in the store one of them is going to be slightly smaller one of them is going to
00:33:00
have an extra camera that a lot of people don't really know what to make of like you know the iphone pro has an extra
00:33:05
camera but as far as i can tell people aren't buying it for that yeah um it'll have a little bit of a bigger screen and
00:33:12
that's kind of it same exact chip same exact software and a lot of people haven't noticed the
00:33:17
difference between 90 and 120 hertz so that's not a differentiator either yeah for the pro name you know
00:33:24
i feel like you're writing that yeah anytime you see pro in a name you immediately go oh okay this better cost
00:33:30
more yeah it's the better one obviously but like what are the real features but yeah this this phone isn't going to
00:33:35
really have that many different features right versus the regular pixel six so you're right it probably
00:33:41
can't be that different from each other um the other thing i i i really like the
00:33:47
comparison with uh the mac transition but not just because it's going from x86 to to arm but i think because
00:33:54
they're going from not having control over what new the chip designs will look like with intel
00:34:00
versus having exact control over everything they want to prioritize and design for in every new generation of the chips
00:34:06
yeah and that's what i see when google's going to qualcomm and like crossing their fingers for a chip that
00:34:11
has more ml core and more stuff that focuses on the stuff they they care about yeah but they don't have
00:34:17
that level of control so they can take all the control into their hands and i think i saw some people sort of speculating
00:34:22
this will probably be in line with like a snapdragon 765g like nobody really is going to notice
00:34:28
the day-to-day difference in performance versus that and a triple eight or whatever the next one is 895 or whatever
00:34:34
and at first i was like kind of bummed by that i was like i was really hoping this would be a flagship but the fact that we probably won't
00:34:40
notice a performance difference and it will do better at those ml things yeah probably
00:34:45
makes up for that for me i slightly disagree okay because i because for the regular cpu and gpu cores
00:34:52
there's no way they're not just using the off-the-shelf arm parts so i have a feeling that they're going
00:34:58
to use probably the best of what arm offers right now which is basically what qualcomm does but just tweaks it but i think they're not going to be like optimizing
00:35:09
cpu and gpu like qualcomm will do okay i think that's that's a big reason why they didn't talk
00:35:15
about speed yeah they were never going to tell me yeah yeah yeah never yeah yeah i'm curious curious i don't
00:35:21
know enough about ship architecture to like really flush out this theory about like
00:35:27
okay if you use more space on the die for ml does that leave less space for normal
00:35:32
computing tasks but also you're trying not to use those cores as much i don't really know that's just the way my brain looks at tensor and like the
00:35:40
things they can they can focus on but yeah honestly i'm excited for it i think when we get the pixel in hand and use it it's gonna be
00:35:46
like the telling moment yeah yeah i think it's definitely gonna be one of those different experiences kind of like we
00:35:52
had with the m1 max again it's like they're not switching from intel to arm but they're switching
00:35:57
what they prioritize it's it's similar to the transition between sync like single threaded workloads were
00:36:04
like all the rage for a really long time and then only a couple years ago amd came in and they were just like hey
00:36:11
everything's fast enough like single threaded workload everything can be handled by these single core speeds
00:36:17
at 4.4 gigahertz fine what if we just throw more cores at it multithread out all the things
00:36:23
yeah exactly and suddenly that started taking off so then intel had competition intel had to start making multi-threaded
00:36:28
stuff and so now there's that battle going on there but i think that's a similar example for google it's like yeah the
00:36:35
888 has a 25 faster cpu 25 faster gpu that's not making a difference in your life from an 865 to an 888 yeah what's going to make a real difference in your life is how
00:36:46
like seamlessly it can do all these ml tasks that you don't actually notice and that's like the frustrating thing
00:36:53
for that i feel frustrated for google is that a good experience is when you don't notice the good things
00:37:00
happening that's so true that is so true yeah the the best tech fades into the background and doesn't
00:37:07
impress you as much as the flashy as tech yeah um that's sad truth yeah here's another here's a question uh should qualcomm be worried and i asked this because i you know the
00:37:19
obvious logic is okay that's one more customer gone from buying qualcomm chips and if they set a good example for
00:37:24
themselves maybe there will be future that also drop but also when you look at like the
00:37:30
like the snapdragon versus the exynos samsung phones the exynos version is often worse yeah
00:37:36
which is quite often which threw a wrench in my whole like optimization mindset because i'm like
00:37:41
samsung you make you design the chip and it's worse than the off-the-shelf one how is it so what do you think about this qualcomm
00:37:47
well qualcomm has a bazillion patents and they like pride themselves on making a ton of patents if
00:37:53
you go to their offices like they have walls and walls and walls of patents you know what patents do they don't allow other people to use the
00:38:00
technology that they've kind of figured out right right so they're kind of building this wall around themselves i think they
00:38:05
should be worried but not necessarily because of google but because of this sheer number of companies that are
00:38:11
starting to go full stack architecture oppo is rumored to be making its own chips for oppo
00:38:16
and vivo phones okay huawei is already doing it samsung is like potentially going to be
00:38:23
moving completely to x and those at some point now google is it's like this race like welcome to the future
00:38:28
where every single company is apple yeah yeah i i get confused about you know why
00:38:34
the xnose version is worse and like you bring up the patents it's like there's a lot of things they can't do that qualcomm has a lock on
00:38:40
but does that also mean that tensor will be worse because there are things that they can't do and will it end up being
00:38:46
the exynos of the pixel world like i don't i i hope not but i figure if we're as optimistic with this as we
00:38:54
are about like how good apple is with their architecture and their chips it should be better how how is xynos not
00:39:00
better at this point yeah i think then sorry go ahead and like how do they how are they going to go full exynos when
00:39:05
the snapdragon version is still better right right i think exodus main problem has always been battery life
00:39:11
and that's been the main constraint like on a cpu speed level it's generally been similar
00:39:18
okay and i think the most the thing that most people complain about with the exynos versions of the snapdragon chips
00:39:23
well of of the arm chips yeah is the battery life uh and so you know i don't know what
00:39:29
google's gonna do in terms of battery life like they could optimize it because it's a full stack thing but like you said like
00:39:36
the exynos version of these galaxy s phones galaxy note phones are often worse yeah so it's possible that we see a rocky iraqi start i think apple had a big leg
00:39:46
up because x86 is such an old architecture whereas google is competing directly with
00:39:52
qualcomm because they're both using arm parts yeah i think the reason i'm so optimistic
00:39:57
is because when you look at the iphone apple makes ios and apple codes and makes that silicon um google makes android yeah and so when
00:40:09
they are coding and making android and optimizing they have truly the full stack like all the others
00:40:15
obviously are making android phones but they don't make android they'll take it and they'll optimize it for themselves yeah
00:40:20
but that's like the one extra bit where we never really look at the milliamp hour size of
00:40:26
an iphone battery because it doesn't really match up versus the expected battery performance
00:40:31
because they don't have the same level of vertical integration as apple so now google with pixel will finally
00:40:37
have that full stack that's why i hope i'm crossing my fingers for audio listeners
00:40:42
i hope that they can take advantage of that and deliver an awesome battery life despite you know
00:40:48
not having the biggest batteries in the world i think they're definitely getting closer to what apple can offer
00:40:54
they are constrained under a similar constraint that windows is constrained by when it comes to windows versus mac
00:41:01
whereas windows has to have support for all these older versions it has to support other phones and yes
00:41:07
google can optimize android for the pixel but it still has to work on other phones
00:41:13
right whereas apple on the mac and on the iphone can literally update to a new os and say hey
00:41:20
developers if you want to be on the new os you have to change your code base whereas windows can't do that android
00:41:27
can't do that android's mostly coded in java java is old now apple does everything in new and
00:41:33
swift and it's great right it's fast it's new it's it's lean yeah so that's that's the constraint that
00:41:39
google's under they said they've been working on tensor for four years i think they've been working on it they've at least been
00:41:44
conceptualizing it since they decided to launch the pixel program i believe that we want to talk about a couple other quick things with
00:41:50
pixel six because i still have uh future hope i have i'm optimistic uh the one thing i wanted to talk about was software updates so i mentioned this in the video and this came from a lot of the ideas you were giving me
00:42:02
so when an android phone manufacturer promises three years of software updates or a lot
00:42:09
of times they'll say like it's three generations or whatever let's say three years yeah um they are depending on like a couple
00:42:17
different variables to be able to deliver those three years yeah they obviously are gonna keep getting updates from
00:42:22
android and they're gonna keep coding and and sending resources to changing that
00:42:27
and making sure it works with their new device and three years later they stopped putting in the effort and it just
00:42:33
has that version forever with pixel and tensor and google making android they are not just able to deliver possibly longer but it feels like
00:42:44
they're incentivized oh absolutely to do so can you break that down yeah so when you have a phone when you
00:42:51
buy phone from say uh i was gonna say lg but r.i.p let's say lg because i think they only
00:42:57
ever promised like two years of software okay uh basically that's cons the software updates are constrained by
00:43:03
the chip maker qualcomm because they know this yeah yeah so they have to basically support
00:43:10
that android version on that phone for a certain number of time now qualcomm the max amount of time they usually
00:43:16
support this can be changed depending on their relationship with an odm but the max amount of
00:43:22
time that they usually support android updates for is about three years per device and if you'll notice the companies that
00:43:28
make a big deal about like supporting your phone for longer uh samsung for example
00:43:33
offers four years of security updates so there's three years of android versions four years of security updates sure the
00:43:40
reason they don't push more than three years of android updates on their exynos chips is because they want to
00:43:48
have feature parody with all their phones so they don't want there to be like confusion right there's been that issue before where like
00:43:54
samsung could have done something in the s20 with but only on the exynos version but
00:44:00
they don't because qualcomm didn't offer it ah right so they need that feature samsung you've avoided the confusion you're not
00:44:05
confusing at all now yeah but with google now they can just be like oh well we're gonna support
00:44:11
android for as long as we can and that is beneficial to them because they make way more fun they make way more money on
00:44:18
you after they sell you the phone based on like things like google assistant things
00:44:24
like search like they make all their money through ad services basically and this is a huge
00:44:29
reason why apple has really opened up their ecosystem of monthly payment services like apple fitness plus apple news plus apple tv plus yeah because they just saturated the
00:44:41
market of iphones like in the us this is it was so it was so like obvious when i
00:44:47
they they're basically coming in from the opposite direction like okay google has wanted to just give get a
00:44:52
phone in your hand so they can make money from you over and over because they sell ads right um apple has made a lot of money
00:44:59
from selling phones but then a while ago they didn't really make that much money on services they'd get you the phone they don't have that
00:45:05
much data they're not selling ads they're not really making as much money as they could obviously the app store is huge
00:45:10
but when they started making all these services they generated tons and tons of revenue per customer
00:45:17
per device everybody out there with an iphone who pays 10 bucks a month for apple music who pays for icloud storage who pays for
00:45:22
apple news and apple tv is just giving apple more and more money and that's that's a new development for
00:45:29
apple whereas with google they want to have the device in your in your hands as long as possible
00:45:35
now because they've already been making lots of money on the device where they've been making their money the entire time when they first sold the
00:45:41
nexus 7 tablet they lost money on that hardware for every unit sold remember that because they were trying
00:45:47
to get people to make google accounts that was literally the reason they sold that tablet in the first place yeah so they've been doing that forever
00:45:54
apple is now like oh we made tons of money on hardware but now we need to start really ramping up
00:45:59
our ecosystem lock-in through monthly subscriptions apple used to make a ton of money on
00:46:04
hardware they still obviously do when you see their iphone sales it's still a lot of their revenue but they're starting to realize like
00:46:11
that iphone sales are going down every year people are replacing them less often and it seems a little bit counterintuitive right because you would
00:46:17
say if they're replacing them less often then you should support it for less time but they're realizing
00:46:23
they're making more money off the residual services that come with your iphone like apple music plus apple music apple
00:46:32
tv plus apple fitness plus all the pluses google's been doing that forever so now it's really in google's best
00:46:38
interest for you to have that phone for as long as possible because on google's own phone you're going to
00:46:44
have gmail be the stock email app you're going to have chrome be the stock internet app they make the most money if
00:46:51
you use their services whereas other android phones you know it's becoming more common for other android phones to use google
00:46:57
services as the primary dedicated app but the eu will actually find google for making
00:47:05
their app the default app it's anti-competitive behavior because it's anti-competitive and they'll lose money so the more people that are on
00:47:10
pixels where they can say this is our product we're allowed to make all the google services the default apps yep the longer
00:47:16
you're on that the longer you stay on a pixel and don't go to a huawei phone or a samsung phone where
00:47:22
you might be using samsung internet the more money they're gonna make yeah this reminds me of a question i asked satya nadella but it kind of would
00:47:28
apply to google as well google makes android and it's open source and anybody can use android
00:47:34
and anyone can make a competitive android phone but google has the biggest advantage theoretically to
00:47:40
make the best android phone yeah and i'm so curious how that's going to pan out now that this seems like
00:47:45
they're really they're really going to spend and they're really going to try and they're going to actually try to if they want market share they can be
00:47:51
competitive yeah they can actually push this phone so i'm curious about that one more thing yeah uh
00:47:57
pixel is all about camera yeah and i noticed that these phones all have new sensors like
00:48:02
bigger sensors there i think they literally said in the blog post that these cameras are so big now that they
00:48:08
don't fit in a normal camera bumper yeah why we have a camera bar yeah so yeah i i wonder if this is
00:48:15
and this is just pure speculation but if this is going to take the leap that we were finally hoping for
00:48:21
where you know pixel 2 was like an incredible smartphone camera and maybe the photos were a little blue but they
00:48:26
eventually improved that and they had better glass in air quotes because as we learned it's plastic lenses and
00:48:32
smartphones but like they've had a really good camera for a while that didn't make huge leaps and they're making a big hardware leap
00:48:38
here do you at the risk of hyping it too much do you think that this is a big leap
00:48:43
yeah i mean you know what's really funny is that uh the pixel 2 actually had a worse sensor than the pixel one like if you if you look at that wow
00:48:50
a worse camera right it had like a higher aperture so it let in less light and it was just it was very
00:48:57
slightly worse i think it was probably just a cheaper part and because the pixel 2 was like they started using
00:49:03
those uh oled screens from lg um it was more expensive to use the overall phone anyway regardless of that it's really funny watching the curve of the last maybe like four years when the reviews come
00:49:14
out and the pixel reviews come out and everyone says camera still the best next year camera
00:49:21
still the best but barely camera uh one of the best i feel like it's one of the best again and then iphone 11 pro come out came out and everyone's like i think apple's starting to make a
00:49:32
better camera than google i gave it my camera of the year and the smartphone awards yeah yeah this is uh the pixel has
00:49:38
always been or the camera sorry has been one of the reasons why i love the pixel so much and it's one of the best cameras in any
00:49:44
android phone not just as far as quality but it's always been really fast with instant shutter and all these new ml things to like
00:49:50
remove blur and it just seems like it's going to be a great camera again uh but i wonder like you know we have
00:49:57
huge huge sensors in some other smartphone cameras and they really don't do all that much more with it right
00:50:03
like if you compared let's say a pixel 5 shot with that small sensor with like the xiaomi mi 11 ultra with
00:50:10
this gigantic like almost one inch sensor and you take the same shot on each it's like ah sometimes i just prefer the
00:50:16
pixel photo like it's yeah it's very detailed it doesn't have the big sensor look but like pixel moving to a huge sensor
00:50:23
now has me really curious you bring up that big sensor look and i think people are starting to notice that
00:50:29
more over time because a bigger sensor look will make things look more lifelike whereas like even if a pixel's tuning
00:50:35
and sharpening from the older pixels look better you start to zoom in at all and it just completely falls apart you look at things like grass and it's over sharpened if you actually look
00:50:45
at any details of pixel shots it none of the individual details are good the whole shot is good
00:50:51
right that's a good point a lot of times if you put a side-by-side shot of the pixel versus the m11 ultra or whatever
00:50:56
other big sensor phone you want to give even a samsung like yeah you're right the zoomed out shot i picked the pixel nine times out
00:51:03
of ten yeah the second you want to zoom in on pinching and some detail you're gonna get a sharper image from the other yeah
00:51:08
so i'm curious i've heard rumors i don't know if this is i don't think it's confirmed at all but of like a 50 megapixel primary sensor i'm like
00:51:15
whoa i think the reason i'm most excited is because there are like a multiplicative of reasons why
00:51:22
this could be amazing right you're letting in way more light right and google was
00:51:28
already using ml to add in like artificial pixels right they were filling in space that wasn't there using machine learning so not only do you get
00:51:39
the combination of natural physics benefiting you but you get natural physics plus tensor
00:51:46
which is like we're gonna use a lot of this chip for ml so you have like physics plus insane
00:51:52
machine learning capabilities and real physics both on your site exactly i mean the the interesting thing is like if they had
00:51:59
used the a88 they actually could have made the camera better because qualcomm to qualcomm's credit they did add like
00:52:05
multi-frame video like hdr video which is literally one frame they're actually taking three frames one of which has a short exposure one that has a medium exposure and one that has a long exposure so even
00:52:17
if it's one thirtieth of a second uh it's actually doing like a one ninetieth a one one seventieth and then doing a lot it's crazy yeah um but
00:52:29
i am really really excited for both video and photo capabilities of these phones yeah i think they're going to be
00:52:34
very awesome i think night sight from a massive sensor with tensor is going to be sick insane sick yeah
00:52:42
well at the risk of uh overhyping and making our expectations way too high we'll end it there i i do expect to see
00:52:50
more info on pixel over the next couple weeks it's probably going to come out in october again or at least get announced in october
00:52:56
again and then come out after that but yeah six and six pro you can watch the video on the main channel if you haven't already seen it for our
00:53:02
our organized condensed thoughts on tensor but definitely check that out and uh also head over to the studio channel if
00:53:09
you haven't already go ahead and subscribe over there it's brand new but it's good stuff um all right that's been it thanks for
00:53:14
joining me david shout out to everybody who's been with us since uh since the beginning of the video podcast and we'll catch you
00:53:20
guys in the next one peace waveform is produced by adam molina we are made in partnership with studio 71 and our intro outro music is by vayne sil
00:53:31
stick is that good yeah we might as well just use it right you