The hidden costs of building an Android phone

Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
David (00:02):
Hello, and welcome to Android bites.
I'm David Ruddick and each week I'mjoined by my co-host Michelle ramen.
As we dive deep into the world of Android.
And on this week's episode, we havesome repeat guests who, if you're it all
into the Android hardware space, you'veheard of these people, their company,

(00:22):
and the product that they're building.
And so Michelle take it away.

Mishaal (00:27):
Thanks, David.
So we've, re-invited two very awesomeguests onto the show and that,
by the way, it was intentional.
We've invited Jason Keats andGary Anderson from awesome.
That's OS O M not to be confusedwith E S O M E although they
are two very awesome people.
Thank you for joining us.

Gary (00:45):
Sure.
Thanks for having us.
Yeah.
Thanks for having us again.
Have to
be.

Mishaal (00:49):
Yeah.
So last time we had such a longdiscussion that we had to split it
up in the two separate episodes.
We talked a lot about the process ofgetting AOS P and GMs onto a device.
The whole step by step process, yougotta go through takes many months.
You gotta get, build,ready to integrate GMs.
You gotta certify it.
You gotta go through third party labs.
You gotta make sure allyour eyes are dotted.

(01:11):
All your Ts are crossed.
There's so much you gotta do justto get GMs working on a device.
And that's fundamental isshipping a device to consumers.
If you want to have any sort ofsuccess with the regular user and
you wanna ship an Android device.
You pretty much need GMs onto it.
It's a different story withenterprise devices, but that's
something we'll get into later.
But on today's episode, I wanted to talkabout the many, many, many aspects of

(01:35):
getting a device onto market that mostusers probably aren't familiar with.
So I'm sure you've wondered.
You've seen a great new phone.
That's only available inX or Y market and you're.
Why isn't this device availablewhere I'm living, why can't
I just buy it in a store?
Why is it only available online?
Or why is there no warranty orwhy is it missing this feature?
Or like, why can't I watchNetflix and HD on it?

(01:57):
There's so many different littlethings that go into actually making a
phone, do everything you expect to do.
And it's kind of really hard to actuallyexplain what everything needs to be done
to actually get all these things working.
And so I wanted to bring Jasonand Gary back onto the show to
explain some of these things.

(02:17):
We're kind of going off onwild tangents this time.
We don't have a strict outline.
We're following.
We don't have pre-planned questions,but, uh, I think this will still be a
very interesting and fun discussion.

Jason (02:28):
Yeah, I think it's super cool.
I think one of the things that alot of people don't realize is how
much certification, how much workgoes in on the background in just
getting a device made to be legal,to distribute in any one particular
country or any set of countries and the.
It's quite significant.
A lot of it takes a hellof a lot of time to do.
And actually just in preparation forthis, I brought up our certification

(02:51):
list and it's kind of interesting.
Right?
And we, you and I have chattedabout it in private that a lot
of companies wanna hide this.
They wanna be very protective orsecretive of how to build a device.
I'm not a hundred percent surewhy and awesome really stands
for transparency in this.
We want to talk about it and shareonce the product is kind of done,
what has to be finished beforewe can put it in user hands.

(03:12):
And that's where we stand today.
Certainly interesting.
When we have a lot of customers whoare like, Hey, you know, you said
you were shipping this date, or,you know, we made a change and how
come it's taking so much longer.
And you know, a greatexample is to go over.
The lead times and certificationsthat can only be started once a
particular level of readiness is
finished.

Mishaal (03:30):
So this is kind of going off UNAT tangent.
Like I just promised I would, but I thinkearlier today, mark Erman from Bloomberg
tweeted that, um, I don't know if those,you have seen the iPhone 14 launch, but
you know, the whole dynamic island thing.
In order to hide the fact that theywere gonna call it dynamic island,
apple actually filed, I think,a trademark for it in Jamaica.

(03:51):
And in order to actually, becausethe Jamaican trademark database
isn't searchable, you have toactually physically go in person to
their offices in order to determinewhether something's trademarked.
People had no idea this name wascoming until I think earlier this
week or earlier today or something,the trademark was filed, I think
with New Zealand or Australia.
And then that referenced the earlierJamaican trademark registration.

(04:13):
And then we're like, thatis really, really, really
clever legal work right there.
But like, it kind of goes to show there'sjust so many little things you have to do.
Like even just, if you want to callsomething a name and have it be
protected, you gotta file a trademark.
But then which countriesdo you file that trademark?
And like, how do you make sure youaren't revealing too much information
in any of these certificationfilings or trademark filings?

(04:35):
You know, you don't wanna leak your ownproducts just because you have to get
something legally certified to be ableto be distributed in a certain country.
So I kinda wanna start on that.

Jason (04:43):
Why don't I go off on a separate tangent to connect that feel free.
It entertain the hell outta me this year?
Honestly.
So when we were getting our, I E Inumbers, which when you're building
a phone early on in the development,you have to apply for I E I numbers.
And when I was getting those,they're like, oh, we need a copy
of your trademark for your logo.
And I was like, seriously, they'relike, oh, it's a new thing.

(05:03):
You know, you have to have a logo on thephone and you have to have it registered.
And I was like, well, at essential,we didn't have anything on the phone.
And they're like, yeah, it's a new thing.
You wouldn't have been able toship a phone without a logo.
Anymore.
And I was so confused that that wasthe slowdown in getting I E I numbers.
I was like, okay, lemme gofinish, registering our trademark
logo so that I could get I E Inumbers so we could start testing.

(05:25):
And, uh, at one point the guy waslike, oh, it needs to be completed.
I'm like, if you long, it takes to get atrademark approved in the United States.
It's like a two, three year process.
And guy was like, oh, okay.
Yeah.
Yeah.
We just need to see youapplied for the trademark.
But that, that was the gatekeeperto getting to I E I numbers so
that we could start testing.
Cracked me up.

Mishaal (05:42):
Speaking of I value numbers, something that I found pretty funny was
that when you're submitting an FC filing,you know, you have to submit details about
the E U T or the equipment under test.
And oftentimes that includesthe I E I number of that device.
So a lot of times within theFFC filing, you won't find the
marketing name of the device.
In question, you'll just find themodel number like SM dash something.

(06:04):
But often what leakers would do isthey would look at the I E I numbers
submitted in the FC filing and theywould go to, I E I databases that
are maybe provided by some carriers.
I think ting was a popular example.
They would input that IM E I, andthen within T's database, it would
correlate that I E I to the product name.
So people would figure out, Hey,this filing is for this product.
And it's like, . Just the game oftelephone with like, how do you piece

(06:29):
together information from these bitsand pieces thrown to certifications?
And it's just like, I I'm, I'm curious,like what goes into this act, the
FC certification filing like, um,how much of that is done in house
and how much of that is relegated tothird party companies because of lack
of equipment or time or resources.

Jason (06:46):
It really depends.
So I'll tell you right now we have22 certification items that we go
through that take anywhere from,I'm literally looking at this right
now from one week to 12 weeks to24 weeks to finish certifications.
EV that includes CE FCC, UL,wifi, Bluetooth Alliance,
uh, Canadian electrical code.
I'm just pulling some random ones outIP ratings, UK C, which is a newish

(07:09):
one since the UK decided they don'twanna be part of the EU anymore.
Uh, GCF PT, C R B.
So it depends on which it is.
Some of them we do in house.
Some of them we do in parallel withour contract manufacturer, our cm,
and some, we farm out to third party.
Several of 'em are just self-reportedresults, but some of them have
to be verified by third parties.

Mishaal (07:32):
So apart from like the telecommunications regulatory agency in,
you know, whatever country you wanna shipa device in, for example, for every radio
hardware in the device, NFC, Bluetooth,ultra wide band, each of these have
their own industry standards group thatrequires you to go through a certification
process in order to license the branding.
I think Bluetooth special interest groupis one example where you have to actually

(07:55):
get their approval in order to stampthat this device supports Bluetooth.
Can you explain a bit about thoseinterest groups and like what
the process is like apart from.

Jason (08:04):
I would say, I know Bluetooth and wifi are fairly simple, to be honest,
it's in terms of doing some basictestings that we do in house and we
provide our results and they validateour results and they just double check
it, make sure we're not BSing them onnoise or volume or any other issues
that might occur that make sure that ourmeasured spec matches our promoted spec.

(08:25):
And they're quite easy.
The hardest one certainly is thecellular antennas, particularly with the
advent of 5g, when we started awesome.
Two years ago, two and a half yearsago, we were talking to certification
labs and they were giving us leadtimes of like nine months to do
certification on 5g antennas.
Thankfully, that's come down to threemonths now, but those were scary days.

(08:47):
It's like, oh, we'd get to thepoint where we'd be done with
manufacturing to an extent.
And now we gotta sit on our handsfor seven to nine months while we
wait for certification to come in.
Right.

David (08:57):
I I'm curious about the antenna certification.
I mean, Aside from like the length ofdelay, obviously, because there's so many
5g devices being certified right now.
Is there a big differencebetween certifying for 4g as

Jason (09:09):
far as I could tell?
No.
They, to be honest, the labs werefairly non-transparent about what the
process was and why it took so long.
I suspect those earlier days, theywere more just trying to figure out
what they wanted to study and they justassumed it would take a lot longer.

David (09:25):
I'm curious about Bluetooth.
So we see Bluetooth performance,very so wildly across devices.
And part of that is chipset.
Of course, , they're probablyworking with different firmware,
like literally, depending on whenthat device was certified or whatever
Qualcomm chip it used was last updated.
And it sounds like obviously, based onwhat you said, Bluetooth SIG is kind

(09:46):
of like not rubber stamping it, but

Jason (09:49):
not, it's, it's more just validating the result.
Like if we say if inour own testing, we say.
Okay, this is the range it's good for.
And this is the signalstrength at that range.
All they're gonna go do is validatewhether or not what we said is accurate.
Got it.

David (10:04):
Do you feel like they're doing a good job protecting
consumers in that sense as the S I

Jason (10:09):
at our level?
Absolutely.
I think the biggest problem comes inwith the random pair of headphones
you bought for four ninety nine onAmazon might not connect as well.
And a lot of people tend to blamethat on the phone or the device
you're connecting to, not on thedevice you're connecting with.
And a lot of that comes backto power numbers, right?
If they want a really, reallyinexpensive Bluetooth device, they're

(10:30):
not gonna put a very fancy battery.
So the signal strength and receptionis gonna be very weak and you're
gonna get a poor connection.
So it's a good segue into conversationsabout when we're partnering with
other companies, we're going to becurating accessories to go with saga.
Is finding partners who are usingthe right ships that they've made the
right investments in terms of powerand performance on their Bluetooth,

(10:50):
connectivity and other features and makingsure that devices work well with saga.

David (10:55):
See, that's, that's really interesting to me because I feel like
the amount of Bluetooth validationthat happens out there in terms of
accessory side is basically zero.
Google is tried a little bit, but Idon't think they've had much success.
I'm wearing a pair oflike Google certified.
These are the mot blog.
NBO one, the Buies pair of headphonesI've ever owned far and away.
So , it's just a personalcuriosity for me, I guess, given

Jason (11:18):
Bluetooth.
Yeah.
Cert, that's funny.
I I've gone.
I love audio equipment.
I, I like, I have so many random fancyheadphones and amps and things like that.
And, uh, I went full wireless and nowI've sort of made my way back to wired
because at the end of the day, Yeah,there's a benefit to being wired.
You don't have to worryabout battery at all.
You don't have to sacrifice that spacein the device for battery, and you can

(11:41):
just punch so much more power into audio.
If you're hardwire.

Mishaal (11:45):
You know, I actually think David, your issue might just be
more with Bluetooth in general.
Bluetooth is notoriously very buggymess, no matter what platform it is.
And there are many efforts on theAndroid side to make things more stable
with like a rewrite of the Bluetoothstack on the certification side,
something that's even more of a mess.
And I think we can allagree on this is us.

(12:06):
How in the world, are you supposed totell what a USB 3.2 gen one X five,
whatever the heck that's supposed to mean?
Right?
The average user probably has no ideabecause there's so many different
labels and so many different numbers.
Something that I found that's quiteinteresting is like in the world of us,
BBC enthusiast, CS, there is a communityon Reddit that is prided themselves

(12:26):
on being enthusiasts around USBC.
There's so many differentcables and chargers and stuff
you can buy off of like Amazon.
But like if it's not USB, I Fcertified people generally say to
avoid those products because only acouple of these products actually go.
As far as receiving USB I F certification,that kind of wanted me to ask you

(12:46):
Jason, how many of these certificationsor which certifications are actually
necessary or rather mandatory toactually bring a device to market.
Do you have to get U S P I certification?
Do you have to get Bluetooth sixcertification, wifi lines, et cetera.

Jason (13:01):
Um, I'm just looking at the list right now, honestly,
to give you an exact answer.
They're basically out of these 22 thatwe're doing 18 of them are required.
The ones that aren't necessarily required.
So like one of them is for Amazon only,only Amazon actually requires us to
have a UL certification, but we do itanyways, just in case we sell through
Amazon at this point, we're not doingit, but we wanna make sure it's there

(13:22):
and we're doing 'em all at the same time.
I'm literally reading throughall of these at the same end,
actually, all of these, we do.
We, we mandate forourselves that we do these.
It makes sure that the productthat reaches the customer meets
our high standard that we expectfrom our products, IP rating.
It's self-reported to an extent,frankly, IP rating's always one of

(13:45):
those certification that cracks me upat IP 68 because IP 68 literally just
means more than IP 67, like IP 67 is onehour, uh, sorry, one meter 30 minutes.
Anything beyond that's considered IP 68.
So you could do five hours of 10 meters.
You'd still be IP 68,

David (14:02):
1 of the ironies about that.
And I'm, I'm curiousto your thoughts here.
So to me, IP 66 would be the mostintense way you could ruggedize
device because that's hightemperature, water jets, right?

Jason (14:15):
Uh, not really IP 67 is definitely harder.

David (14:17):
Okay.
Interesting.
Because I see few devices,cause it's a prolonged

Jason (14:21):
pressure right.
Over the entire device.
Cause what can happen is as you getpressure over the entire device at
one meter, you could cause a liftin one spot that creates a gap
on the other side of the device.
Just because you're under pressure,the entire device under pressure.
Whereas if you're just targetingit with IP 66, that one's
not that difficult to pass.
Once you get to 67 and68 gets more difficult.

(14:41):
I'm very, very proud to saythat saga is IP 68 rated now.
So if I have a lot of timeto get there, to be honest

Mishaal (14:48):
yeah.
I mean, for consumer peace of mind, thatwas always nice to know, you know, whether
or not a product is IP rating or not.
But as you mentioned, IPratings, like the, the data is
collected on it's self-reported.
But if correct me, if I'm wrongto actually label your device
with an IP rating, do youhave to pay a licensing fee?
I remember hearing something like that.
You do?
I

Jason (15:07):
don't remember what it's not much to be honest.

Mishaal (15:09):
Okay.
I was always wondering at whatpoint is there a, a decent trade
off whether or not, you know, okay.
We can say that we do meet thesestandards, but we're not going to call
ourselves IP rated because we don'tthink we should pay this fee in order
to just slap this label on our device.
Even though we are functionally equivalentto IP 68, but we don't wanna have this

(15:30):
branding or we don't think it's necessary.

Jason (15:32):
I think customers really wanna see that number.
I think for marketingpurposes, very important.
What's quite interesting.
I had this conversation with,uh, one of the char, the CEO of
one of the major us carriers.
I was like, you know, what's yourreturn rate for water damage?
And the guy goes practically,N it's incredibly rare.
The device actually gets returnedfor water damage, any phone, even

(15:52):
ones that aren't IP rated that couldalso come from just people going,
oh, I dumped in the toilet and thenflushed and then threw it overboard.
That might be on me.
But people like to know that their devicecan go outside and get accidentally a
little wet and they can be safe withouthaving to worry about it being damaged.

David (16:07):
And I think that it speaks to it mostly being a peace of mind kind of
marketing thing, because in all honesty,Motorola tried to make the unbreakable
phone and nobody cared if that's whatpeople actually unbreakable bought that

Jason (16:19):
thing.
And drove.
I had a conversation when we werebuilding or middle of the road when
we were building the essential phone.
And one of our marketing guys was that,oh, can we call it an unbreakable?
And I looked at 'em.
I said, You hand me a block ofbillet titanium, just a solid
block of titanium as an engineer,I'll figure out how to break it.

(16:39):
So you can't call anything unbreakableeverything's breakable within reason.
So anybody says they're gonna build anunbreakable phone is outta their mind.
I mean, I, I have built some phonesin the past, some prototypes that
could withstand a hell of a beating.
We did one that I would still love to do.
One day it'd be kind of crazy expensive,but when we were building project jam,
when we were looking at the materialsfor the housing, I built some outta

(17:00):
Silicon nitr, which is what, uh, similarmaterial they make the profess out of.
I could take that housing outinto the parking lot and spike it
like a football and it'd be fine.
It was kind of unreal.
sure each housing costs likea thousand bucks, so maybe
not great in mass production.
But it was super, super cool, uh,to break it, I had to freeze it

(17:22):
and then hit it with a hammer.
So speaking

David (17:24):
of breaking and freezing and hitting with a hammer that gets to the
other group of organization standard,you brought it up before I cut you off.
I'm sorry, which is the mill eight, 10 G,which to me actually like they're way more
fun things you can claim with mill eight,10 G because it covers so many things.
But I guess like, youknow, LG tried to use it.
A couple of brands tried to use it.
It doesn't seem to stick with consumers.

Jason (17:44):
Yeah.
I don't.
I am not that familiar.
Well, I am vaguely familiar, but it'sbeen a while since I've looked at the
mill spec standard, because to your point.
Nobody really cared.
And in some cases it's weakerthan the standard that we put
on ourselves that has become thedefacto for, uh, mobile devices.

Mishaal (18:00):
So speaking of standards you put on yourselves, I wanted to ask
you what are some examples of optionalcertifications beyond like IP ratings
that a lot of companies seem to go forbecause they think it'd be beneficial
to the baseline of the product or justthe branding in general, you know,

Jason (18:14):
the craziest one is always drop testing.
And as a person who's been building phonesand laptops and tablets for the better
part of 20 years now, it's both the mostfun and the most heartbreaking to watch
the abuse that we continuously put ondevices while we're doing development,
random drop tests, tumble tests, which areliterally when you go to the factories,

(18:35):
oftentimes it's like a handmade dryer.
like literally just atumble, like a wheel.
That's just dropping phonesconstantly against things.
So we do a one meterdrop test onto granite.
The traditional test is either one orone and a half meter onto polished Gran.
And that's what most phone manufacturerswork to at least in the Western world.
And we sit there and we drop and wedrop and we smash and we smash and

(18:58):
we make minor little tweaks to thegeometry of the housing or the geometry
of the cover glass or the thickness orthe position of foams on the inside.
It's a constant little twist andturn and just trying to get closer
and closer to being better andbetter under those drop tests.
But at the end of the day, thereis no amount of testing anybody can
do or any self certification thing.

(19:18):
We can do the guarantees that aphone won't break when you drop it,

David (19:21):
I guess, in terms of breakability, I'm curious, what do you think about
the evolution of glass materialsand where we're at right now?
Because it seems tohave slowed down a lot.

Jason (19:29):
Uh, I don't think it's slowed in a lot, uh, partially because
I have very close friendshipswith the team at Corning.
They've been wonderful supporters of usat essential and then us now at awesome.
We're using the latest andgreatest gorilla glass on the
front of device and the big push.
That's maybe cause of the outwardlyvisible reduction in what's going

(19:51):
on in glass development is thisbig push towards 3d glasses and 3d
glass is much more complicated tomanufacture in mass production, both
in terms of tolerances and fragility.
But like we saw inproject gem, it's doable.
It just takes a buttload ofdedicated work to get there.
But in terms of flat panel, glass andglass ceramic, There's still quite

(20:12):
a bit of development going on there.
The team at Corning has a, just abuilding full of mad scientists,
constantly trying to improve bothstrength and transparency of the glass.
That's what a lot of people don'trealize because we're talking about
minuscule percentages improvementsin transparency, while also greatly
increasing strength and scratchresistance and frequently strength.
And scratch resistance isinversely proportional to clarity

(20:36):
of the glass transparency.
And so they can make it stronger, butthen they lose some transparency and
your front camera might not look as good.
So there's a lot of back and forth onthose spaces and it's constantly ongoing.
And I was a little surprisedCorning didn't make as big a stink
about their latest cover glass.
And honestly, at the moment,I'm forgetting what it's called.
It's not gorilla glassseven it's Invictus.
Thanks you.
Thank.

(20:57):
Invictus class.
I dunno why that marketing changed,but, uh, yeah, we didn't hear as
much of it when they, as, as we didwhen they launched grill glass five
and six, but Invictus is stronger.
It's slightly more transparent.
I mean, we're talking fraction ofpercent and it is more scratch resistant.

David (21:13):
I mean, and don't get me wrong here.
I think that there arestill improvements, I guess.
You're right.
They've not communicated it asclear as they could have, because
I have my iPhone 13 pro max here.
And the number of scratches onthis thing is almost embarrassing.
I have watched it fly 20 feetacross an airport floor and
just skid the entire way.
Not a single crack on it.

(21:33):
So they're doing something right.
But my pixel six pro is utterly destroyed.
I've dropped it several times.
It's got like five different crackson the pack now at this point.
And, uh, I'm just waitingfor the front crack.
So yeah, the big reason it's

Jason (21:47):
all anecdotal.
Yeah.
But you're absolutely right.
And it, I don't, I don'tremember what's on the pixel six.
But there is also a trade off betweenstrength and scratch resistance too.
So it's a constant battle between thosethree variables, scratch resistance,
crack resistance and clarity.
And then, and then that gets, wecould spend a whole hour talking about
why foldables are still a bad idea.
And it really comes down to the display.

(22:08):
I don't even wanna call it glass becauseit's the minute thinnest piece of
glass and then a bunch of plastic, astrata, perhaps something like that.
I've had conversations with a fewglass manufacturers about foldables
and each one of them are like, yeah,glass is not meant to vent despite
everything they're doing, glass doesnot bend nicely, which means having to

(22:29):
use plastic and plastic is kind of a.
Material for covering your display.

Mishaal (22:35):
Yeah.
Foldables, that's kind of thereason why I'm holding it out.
Personally.
I've seen like great results.
A lot of people are really enjoyingthere's Z four threes and Z four S
and there's a lot of competition, butlike, unless you're already established
brand with a deep relationship witha display manufacturer, and you're
probably not getting into the foldablebusinesses, a startup, at least

(22:55):
not without a lot of connections.
Oh, to be

Jason (22:57):
honest, they're they, they are desperately trying to sell
foldable displays to other people.
Cuz the display manufacturer hasput a ton of money, invested a ton
of money into foldable tech anddevelopment for all their processes
into foldables and other than Samsungand a couple Chinese brands nobody's
pursuing it that hard for that reason.
So probably once a quarter geta call from our display guy.

(23:18):
Hey, we made minor improvementsin the cover glass with the fold.
Well you sure you don't wannause one on your next phone?
So it, it, it's not a matter of that.
It's a matter of.
It's not that great ofa user experience yet.
And I haven't seen one that really blows

Mishaal (23:31):
me away.
Well, their margins are probably alot higher on those foldable displays,
so I'm not surprised they're pitchingthem to . I kind of wanted to ask
you though, the two of you, you haveextensive experience in building a
smartphone startup, and actually youhave a lot of partnerships already.
You have a lot of contacts at allthese companies, and as we've talked
about, there's so many differentcertifications, so many different things

(23:54):
you mentioned there's 22 certificationsthat you're pursuing for the saga.
Yeah.
And I'm.
If you are someone who wants tobuild your own smartphone and
you're not already well connected.
And in this industry,where would you even begin?
How would you even beginto navigate this mess?
are there companies that actuallyspecialize in helping you get these
certifications around the world?

(24:14):
What do people do?
Most people just go forAudi 30 M designs, even

Jason (24:18):
then you still have to get certified because if you wanna
market it under your name, all thecertifications have to be in your name.
So even the saga, which is soldthrough, so salon, it still has a strong
connection to awesome because all thecertifications are done by awesome.
If I had no connections and no knowledgeof all the certifications that are
required, besides starting at the bottomof bottle of a whiskey um, I, yeah,

(24:41):
I don't even know where I would start.
I'd probably go through my cm and be like,Hey, who have you used in the past for us?
Or it depends whereyou're trying to launch.
If you're trying to launchin Asia, it's a lot easier.
You're trying to launch in the us.
It's way more difficult.
As we saw there was, you know, I'mnot gonna call 'em out by name, but
there was another recent companythat's moving into the mobile space.
They're not launching in the us,cuz it is much more difficult here.

(25:04):
So it really depends whereyou're launching what you want
to achieve with what you have.
There are certainly third party companiesthat can facilitate some, but not all.
You'd probably need two or threethird party companies to get
you through this entire list.
And you need to see 'em that reallywanted be on your working on your behalf.
Yeah.
And that's

Gary (25:21):
how we got introduced to salon in the first place because
they wanted to do that exact thing.
And they went out and search bar and widewith what it was available out there.
Just to kind of learn what it would taketo build a phone, a mobile phone and
launch it in, in the states and, and in.
In the initial effortsthat we're doing here.
So they went far and wide andcame to the point or conclusion.

(25:42):
And some of 'em grantedhave come from Qualcomm.
You know, that the, uh, CEO of ourco-founder of Solana had worked
at Qualcomm for 10 plus years.
Uh, a lot of their foundingmembers came from there too.
So they had touched parts of it, butnot launching a full on device with all
the certifications, everything else.
That's why they came to us.
And in the speed they wanted to move.

(26:03):
I think that's where it made themost sense to go with a company
like us was able to make changesand adapt to some of the features
that they wanted for the saga.
And it being a key directingfactor for some of the other OEM.
In the space.

Mishaal (26:17):
So while we're talking about thes, I actually wanted to
ask, I believe you confirmed thatthis device is launching in Q1
of 2023, or was it the first half

Jason (26:26):
Q1 Q1.
And we've also announced that there willbe a limited number of developer units
available before that, because the bigpush behind all of this is to bring more
and more developers who are generalists inweb three, who have probably been building
for desktop into the mobile space.
So we're getting them accessa bit earlier, and these
are pre-production devices.

(26:46):
So when those devices are made of it,they are pre-production, but we are
making sure to bring in and again,be more transparent with what we're
doing and open up the developmentof a device to a broader community,
which nobody's ever done before.

David (27:00):
I have a question about pre-production we get told all the
time when we get a phone to review,especially ahead of launch, like,
oh, this is a pre-production device.
And usually that's a coverall for.
We're gonna deploy azero day firmware update.
Like that's all they usually mean.
What about what to you doespre-production really mean?
Like, are there hardware differences?

(27:20):
What happens in thosefinal stages of refinement?

Jason (27:23):
So that's a great question.
So I'll give you the whole answer.
So bear with me.
So typically what it is is we go intoa D VT bill, which is design validation
test build with the knowledge that ispencils down from a hardware standpoint
and largely firmware pencils down as well.
And to be very transparent,we're in that build right now.
We're just exiting ourDVT build right now.

(27:43):
Then those devices are givento these 22 agency or 22, uh,
uh, certification agencies.
And you cannot change thehardware at that point.
If you have to, then you have toredo all these certifications.
So when you're given some pre-production.
If it's a D V T level device, it meansthat the hardware is gonna stay the same,

(28:07):
but there will be significant softwareand firmware improvements over time.
Firmware limited updates.
If we can't change radio andstuff like that at that point,
but we can improve camera tuning.
We can improve audio.
We can improve display batterylife over time, but we cannot
tweak the, uh, radio perform.
Once it's gone through certification,

Mishaal (28:28):
right.
And from time to time, you'll seea lot of these prototypes or like
actual pre-production units pop upon like eBay, because people who
have these who are held onto thesedevices and who honestly shouldn't be
selling them, decide to go sell them.
And that's how you figure out that, Hey,that divot on the back of the next to six
was actually supposed to be a fingerprintscanner because there are actually,
pre-production NEX to six units on eBaywith fingerprint scanners on the back.

(28:52):
So like,

Jason (28:53):
uh, you could, yeah, that would, that would've been a, a pre-certification

Mishaal (28:56):
device.
Yeah.
And I've actually, I mean, I've useda pre-certification device before,
like I'm not gonna mention what it wasor like who it was from, but like it
was wrapped in this giant thick case.
And like, you could actuallysee the phone case itself.
It was like clearly meant tokind of hide the identity.
But I think it was like superpreproduction and, you know, they
didn't have the backing fully ready yet.

(29:17):
All I could tell basically was a software.
Like they were doing software bring up.
Yeah.
And you know, I was offering somefeedback on software, but it's called

David (29:23):
a lunch boxing.
Right.

Jason (29:25):
Uh, it depends.
We call dog fooding.
Uh, yeah.
All depends on the vendor.
You know, it was fun thingwhen we did a project jam.
I know I harp on a lot, but it was funbecause we made weird cases that made it
look like a regular phone, just to tiethe fact that there was so weirdly shaped.
We didn't do that with saga orour, with our, we had a different
internal code name, but wejust let 'em out in the world.

(29:45):
It's been quite entertaining when I'vebeen traveling with it and I've traveled
all over the world with it already.
The most common comment I'vegotten because it doesn't have
the logo, the pre-production unitsdon't have the salon logo on them.
And final color schemeis, is that the iPhone 14?
It's obviously the iPhone has beenlaunched now, but, uh, up until recently,
people would automatically assumebecause they feel the build quality

(30:06):
look at how unique it looked and justassume it had to have come from apple.
And I, depending on whoI'm talking to I'll joke.
Well, it did come fromCupertino, but not from them.
So yeah, the, you can do it.
Some people wanna hide it.
I, you know, I don't think the averageperson on the street cares that much.

Gary (30:23):
Right.
So the, so the leak are pretty, I think,conscious, or if you were to see a device
out in the wild people, aren't tryingto hide it as much as they used to.
I think for pre-production units, youknow, when Keith talks about D V T
and then there's a phase before that,which are engineering validation, test
units, which may have varying hardwarein them, like you might have a smaller

(30:44):
memory cuz it's cheaper or more, uh,portable and accessible for the time
during engineering validation teststhat just allows us to develop still
internally from the software end.
But from a hardware perspective,it may not be that imperative with
what we would be for productionunits and then for pre-production
from the software perspective.
Yeah.
You're not gonna have a lot of features.

(31:04):
There is a factory acceptance testbuild that you do provide that the
factory will just end up flashingthat old, old, old software on there.
And it's usually like three monthsupdated at the point to production.
So it means that, Hey, there's really oldsoftware on that, unless you take an OTA.
Or it's a camera thing where the, um,Keith mentioned tuning, that's actually a
really, really long tail thing where youhave essentially three phases of tuning.

(31:28):
One is objective tuning, which makes yourgreens green blues, blue and red red.
But then when you do subjective tuningphases, it's like your greens are greener.
Your blues are blue wearand your reds are redder.
Uh, and that's for different scenes too.
And subjective.
The longer you tune, thebetter your camera is.
And that's really what stops you frombeing able to have it production ready.

(31:50):
A lot of the times, it's not able tostart testing until D B T comes out.
Like you won't start to beable to tune stuff until your
actual hardware comes out.
At that point, the longer yourtune, the better the device is.
Uh, so yeah, I think that kindof sums up what pre-production
means is just software related.

Mishaal (32:08):
Speaking of software, I wanted to kind of get an idea from you to how
important the actual factory provisioningpart is because I'm sure people are
familiar with the term factory images.
Like you can download them for a pixeland, you know, you can flash 'em out your
device and Google says they restore yourdevice to a factory state, but that's not
actually true because there are certainthings that the user just cannot flash or

(32:30):
modify or things that have to be set onlyat the factory, such as like what Google's
doing with the remote provisioning, keysand other such like cryptographic things.
Can you talk a bit a bit about that?

Gary (32:40):
Yeah.
So there's stuff on the device thatis considered persisted, uh, and
that's stuff that necessarily won'tbring you back to an official factory.
And factory is really dependent onwhere you're taking that context from.
Like, if we concert factory reset, thereare things that go into the factory build
initially, where you actually have certainapps that check the quality control

(33:03):
outwardly before it leaves the facility.
So we'll have like a test on there thatruns a gamut of 40 plus tests, essentially
checking that all the things on thedevice are functioning the way they're
supposed to before it leaves the factory.
And then we have other things that goonto the device that test camera modules
on quality control, meaning that you havedifferent settings for camera images, to

(33:24):
make sure that the different glass that'sattached to and layered on top of that
module, it's taking clear enough pictures.
There's a lot of stuff that goes into thefactory that the users will never see.
And you've seen that in Shammyphones where you can restore that
test in there, and then it writesessentially a file that says.
Before it left the facility, it wasin the state and it was good to go.

(33:44):
So it was more at the factory level.
That's what you'll see atthe factory factory level.
Um, when you do a factory reset,there'll be some things that from the
system side, we're not doing anythingthat people won't be able to reset.
Like the stuff that we're doing withsaga, everything will be able to be
cleared essentially, but there's alsofactor reset protection, which is
inherently made to allow people whoscavenger and steal devices from people

(34:08):
to try and sell 'em on the secondarymarket for that to be protected.
So what you'll end up seeing is ifyou ever registered a fingerprint and
you try to factor reset the devicewithout unlocking the phone, for
example, first, uh, it'll go intovector reset protection mode, which
is something that Google implementeda while ago for, to kind of work, you
know, those, those type of effortsfor people on the secondary markets.

(34:30):
But yeah, I think from a consumerstandpoint, when they say factory
reset, there'll be things that can't be.
Cleared completely.
It'll be for the benefit of the ownerof the device and not so much, you
know what OEMs do at a factory level orpre preproduction testing for quality.

Mishaal (34:47):
So some of that factory level stuff, like the tools that you
mentioned, they've sometimes leapedout into the wild and they're like
ridiculously powerful in what they can do.
Like I believe there was an oppotool that was able to restore the
wide vine provisioning, key somethingthat's only supposed to be able
to be provisioned at the factory.
And there's a lot of these tools, likeI think believe from one plus they're

(35:09):
these, um, I think they call them MSM,download tools that the community is
very fond of them, but there's actuallylike an extremely low level factory
flashing tool that allow you to wipeand restore all partitions on the
device, back to the actual factory set.
So.
And like there's a lot of these kindof tools that you users will never
see on a production device, becausethey're all stripped from bills before

(35:31):
they're actually shipped to users.
Can you talk about any ofthese tools, Jason or Gary?
Like what are some of the usefulthings that engineers will have
access to that, you know, userswill never see the light of day.

Gary (35:44):
There's a lot of things for convenience, especially on like the camera
testing side of things, but those aretypically like, Hey, these are user debug
builds or rooted builds to allow them to.
Through UBI or something like that, butyou'll never see the light of day for
that to show up on a production unit.
There are other stuff that is made forthe factory to do what you had said,
like create different memory, footprintsand stuff in the original factory mode.

(36:08):
So factory mode, meaning.
What you had mentioned earlier is that,uh, they have their own build at the
factory that allows them to do theirspecific testing, but then our build goes
into it and then they have to do testing.
That's different from theirstrictly hardware testing.
I've worked at companies where they'vehad internal tools for like Odin or the
ones that you've mentioned, like oneplus, uh, for you to, to do those, um,

(36:31):
types of provisioning, but to do anythingnefarious you would need to essentially
have what's green keyed in the rom.
So typically when you see that scarymessage, when you boot a device, uh,
that you rooted or flashed a, a customRam on, you'll see something that
says warning, like a big red letters.
That seems very, very alarming.
That's essentially why that'sbuilt into there is because you

(36:52):
don't have a build going into thedevice that has been green keyed.
And that green key is only owned bya set of individuals in the company.
So the factory won't be able tosign a factory image that goes
onto that hardware device thatwon't show that warning message.
So typically we've hadlike something, a UBI key.
That's physically attached to signinga build and that green key exists in a

(37:15):
physical world and is completely offline.
And you have a machinethat only signs that build.
And then we release that sign, build to,uh, manufacturers, uh, or, or the factory.
And then the factory never hasthe ability to do anything with
their original factory build.
But the thing that goes out toconsumers is just that green key
build that's flashed by the factory.
And yeah, there's a lot of like handshaking there, but the factory has

(37:37):
their own factory build that hasbaseline for what they're used to.
We have our build that has likea Delta between their factory
and what we did for our hardwarespecifically for like Qualcomm
libraries and then, uh, differencesin Google ASP that we've added on it.
And then all the things that wework with our hardware vendors,
and that gives a baseline for thehardware that the software that

(37:57):
actually goes out to consumers.
Uh, it might be unclear or muddythere, but yeah, that's more or less
where those two builds differentiate.

David (38:05):
I mean, just as a kind of curious anecdote it's I never would've
considered, there was essentially afirmware nuclear football that, uh,
smart home companies have to hold onto.
And I'm curious, I mean, whatdo you all know about how
companies tend to handle that?
Because I imagine Samsung might do it alittle differently than y'all are doing
it just because of scale . So I'm just

Jason (38:26):
personally curious.
That's a good question.
I want Gary, do you know how sayyou worked there in the past?
Any idea how they can?
Yeah, so

Gary (38:34):
it's like a nuclear thing where it's three people need to come
together to turn the key to sign it.
Yeah.
So it's like a split type of thingand you usually have build and release
engineers or a group, and those peoplehave to be very, very trusted and we
have something similar, um, set up.
I

Mishaal (38:51):
mean, just imagine the level of damage you could do if you
had those Seine keys, like if youwere able to sign your own system
image using like Samsung's keys, forexample, then you could literally put
whatever you wanted into that systemimage and flash it onto any device.
And the boot loader would accept it asa valid system image and boot it up.
And then you'd basically havefull control over the operating
system of some user's device.

(39:12):
Like they definitely have tokeep that key under lock and key.

Jason (39:16):
Yep.
yeah, I just, I just put it with a, a, a,uh, a cage that has a bunch of poisonous
animals in it, and nobody goes for it.
It's weird.
No,

Gary (39:26):
yeah.
Essentially snake it's in with

Jason (39:29):
yeah, exactly.
That's

Gary (39:30):
where we keep it.
Do that.
Definitely.
In the physical world.
I I've seen, uh, some of pizza's reptiles.
They, they, um, they wouldkeep a, a green key bowl in

Mishaal (39:38):
Indiana Jones trap.
yeah,

Jason (39:41):
she's been a bit of a jerk lately.
I'm not gonna lie.
She, she could definitely defend it.
The

Gary (39:46):
safety of those things is, is a big thing.
And it usually mixes the same typesof security you would expect from a
two-factor authentication, that mixes, whoyou are with what you know, with, uh, what
you have and where you are type of thing.
It'll be essentially a mix ofthose things split across, uh,

Mishaal (40:02):
multiple people.
Yeah.
And key leaks, you know, even ifthey don't happen at the OEM level,
they could happen at the factory orsomehow some malicious developer or
hacker were able to extract the keysfrom the trust zone on the device.
There are different ways to do it,all of them incredibly difficult.
But if that were to happen forcertain features, I believe Google

(40:24):
has the ability to revoke keysthat are used for at, for example.
And if that were to happen, thenall of a sudden you'd find yourself
unable to use so many applicationsthat rely on at adaptation.
Like a lot of yeah.
Banking

Gary (40:35):
and go ahead.
Right?
The at adaptation's keys and the widebind keys that you mentioned, there
are mechanisms built into the chipswhen it's wide buy ones, there's
like a back off mechanism to whereit could gets completely inaccessible
or completely wiped, essentially.
Yeah.
Those type of things, thereare built in measures.
Android devices have been around longenough or Google to be very, very

(40:55):
conscious of what that would look like.
Essentially, if, if.

Mishaal (40:58):
That ever happened.
And it's happened before.
If you look up news reports of after anupgrade, my wide vine L one certification
went down to wide vine, L three, and allof a sudden you can't watch HD videos
and Netflix, it's a very common problem.
You'll see.
And at some point it happened to a deviceand users had to actually ship their
device to the OEM so that they couldmanually re provision those devices.

(41:22):
It was a big deal.
Thankfully, Google's movingtowards remote key provisioning, at
least for the ATEST station keys.
If that happens, they're able toreissue a key without having to actually
reprovision at the factory level.
But like, you know, thesethings, if they go wrong, yeah.
They go horribly wrong.
Right.

Gary (41:39):
And a lot of these devices are fused.
So you do have these that are prod fused.
So you have something that willtry to do something ne barriers
to that device will eventuallyblow the fuse at a Harbor level.
Yeah, I, I think it's a little bitdifferent about what we were talking about
Michelle, but there are other set guardswhen people are trying to do something
in that areas to the device itself.

(42:00):
But yeah, that would be a nightmare.
If any OEM, uh, would need to goin, have things re provisioned it's
it's not even great on the hardwarelevel for us to doing anything there.

Jason (42:10):
That line of questioning calls back to our earlier conversation about
pre-production devices and certification.
It is very funny when you're testdriving pre-production devices that
have not been CTS certified, whichis the Google certification, peer
devices, certain apps just won't work.
They aren't available.
And I personally find it quiteentertaining because as the
CEO, I start dog fooding a unit,basically, as soon as it's available

(42:32):
for me to start dog fooding.
Absolute garbage, fresh off the factoryline, VT one to now D V T where the
hardware and the software's really good,but it hasn't finished CTS certification.
Netflix just doesn't work.
You can't downloadNetflix onto the device.
Precertification.
The rationale behind that is aroundsecurity and privacy, which makes a lot
of sense, but certain apps just won'tbe available on pre production devices.

(42:55):
And people have to be aware of that.
That's what we go through.
As we do develop.
Yeah,

Gary (43:00):
I think high level there's something called safety net.
And when you don't have a CTS certifiedbuild on there, safety net just breaks,
which everybody like so many appsrely on and play store will note that
your build is not a CTS certifiedbuild and you won't have anything.
A lot of things show up likethe banking apps, anything that
relies on safety net won't show up

Jason (43:20):
as part of that, just terrify everybody.
All of my banking apps work perfectly finewithout safety net the table , uh, really
scares me about all my banks actually.
And

Gary (43:28):
a lot of the open source projects used to be able to SP like
they, they have something calledMicroG, which is essentially a way to
simulate what Google APIs would do.
So it's a way to like rerouteGoogle APIs to do something else.
And it'd be just like a mask and safenet used to be able to be spoofed.
But they have since changed that as a, uh,maybe like three years ago or two and a
half years ago to where it's not somethingthat can, can be spoofed, uh, any longer.

(43:52):
So people yeah.
Are on that boat or I think.
Calx and lineage and some otherand graphing all start doing
something differently to allow youto have a, a different experience.

Mishaal (44:04):
Actually, speaking of MicroG, perhaps by the time this episode goes
up, we'll already have our episodespeaking with a MicroG developer up.
So if you wanna learn more about that,go back and listen to the MicroG episode.
Go give that a listen.

Gary (44:15):
Yeah.
MicroG is one of those things where Googlecan just change something to make it
not work which it has always been therisk of anything getting around Google.
Yeah.

Mishaal (44:26):
This is, uh, kind of fearing away from software back to hardware,
but for the sake of users likeregular users who might have this
question and I knows of it, but Iwanted to hear more from you guys.
The us market is particularly difficultto support because of the bands.
That carriers deploy in certaincities are not really the same

(44:47):
as what you'll get in Europe.
So like, if you wanna support a certaindevice, a lot of the bands you might
have to support will be different.
And so I wanted to ask you, what doesit actually take if you wanted to ship
a device in the us with support forband 71, which is very commonly used by
T-Mobile, what does that actually mean?
What do you have to do to get that?
To work?

Jason (45:06):
It really, we work with our cm and our radio team, our, our, our
modem team to make sure that we supportthe bands that we need to support
in this country and check which onesoverlap with different countries.
It is a, I'm trying to think of aless convoluted way to say this.
We basically look at all the countrieswe would like to support and ship into,
look at where carrier bands overlapand break them out into separate skews,

(45:29):
depending on where the overlap lies.
Actually, I was just having aconversation about this with a
friend of mine who works at, uh, theother phone company in Cupertino.
He he's a big wig over there and.
Even, he was telling me that he isgrabbing RO w skews, which for everybody
that sounds for rest of world, whichtypically means Asia and middle east.

(45:51):
The reason is because the iPhone 14,obviously the us skew doesn't come
with a SIM card, uh, a physical SIMcard, which frankly, I was astonishing
to me in a lot of other people.
So internally even trying tograb rest the world units because
they want to keep a physical SIM.
But that, uh, I, I guess that does answeryour question is we do look at where
the bands overlap and work from there.

(46:12):
And typically we get anorth America EU skew.
We get a Africa slash middle east skew,and we would do a RO WKU, which would
typically be China and India targeted.
Now India has some very specialrules though, and we can get
that's a whole separate story too.

David (46:29):
I think in terms of the us, maybe Michelle, correct me if I'm wrong, the
discussion would be to me, like we havea situation where carriers in the us,
for lack of a better word, effectivelywhite list network features, they
operate network, feature white list.
That's how at and T does it, how T-Mobiledoes it it's how Verizon does it.
T-Mobile's probably theleast egregious about this.

Jason (46:49):
I would say.
Yeah, for they they're byfar the easiest one to work.
At and T is in the middle and Verizonis the most difficult to work with.
Right.

David (46:55):
And so it's far less about band support in my mind, right?
Because Qualcomm reallyhandles that that's their job.
It's more about like, Hey, I'm on Verizon.
I want voice over LTE to workappropriately and to have wideband voice
and for calls to not sound terriblewhen somebody calls on somebody on at
and T because at, and T and Verizonprobably have a cross Ty agreement,

(47:18):
but it probably only works for handsetsthat are certified for their networks
directly and whitelisted, I would guess.
Yeah.
There's all these little bits andpieces, especially for interop that
can destroy your phone's featuresbecause you won't pay at and T a quarter
of a million dollars or whatever it

Jason (47:33):
is.
Yeah.
The number's a lot bigger now.
Oh, cool.
Yeah.
So there are two versions of certifica.
There are many versions, butthe high level to explain it.
Each carrier basically hastwo versions of certification.
One would be B Y O D, which means wesaga Solana are selling a phone that
is open market and is safety use on usbased networks, meaning TMO, Verizon,

(47:57):
or at, and T plus all their MVNOsbuilt off of those three networks, but
basically that's who we're dealing with.
And what that means is we do go througha certification process with each of
those carriers, ensuring with them thatour device won't break their network.
And that allows to be white listedon their network, allows for
interoperability and allows forthose high level features to work.

(48:17):
You know, right now the phone willwork before the certifications
are done, making calls, sendingtexts, but you won't get Ty.
You won't get a lot of other thingslike that until it's finished that
white list that B Y O D certificationand keep in mind, each carrier
has a different name for this.
We just internally call it BY O D to not lose our minds.
On top of that, there is a full levelcarrier certification, and that's, if
you're selling through their stores,That's where the big money comes in.

(48:41):
If you wanna sell a device throughverizon.com at and t.com, tmo.com,
you're gonna go through theirfull certification process and
the full certification process.
How do I say this without megetting sued by one of them?
Um, sucks.
It is complicated.
It is inherently biased against thephone manufacturer because they want to
get you to pay more, to ask for waivers.

(49:04):
Uh, I don't remember if it was on the lasttime we did a podcast or it was somebody
else that I was talking about the mostegregious one isn't actually in the us.
The most egregious example I've everseen in my career was in Japan where
part of their certification process forfull cert was the phone had to survive
a 1.5 meter face down, drop onto asteel L bracket without any damage.

(49:26):
No scratch, no crack.
Nothing we'll pass that again.
We go back to my billetaluminum or bill name.
You will get a scratch and a ding.
If you drop it from one and a halfmeters onto a steel L bracket.
And it was just there to force thephone manufacturer to either pay a
little bit more money for a waiveror accept worse terms on the sale.

(49:47):
And so that is a big reason why we wentopen market on saga was we didn't wanna
deal with that beyond the headache.
You're talking somewhere betweenhalf a million and a million dollars
per carrier per year per skew.
So you can imagine that adds up really,really quickly for even a big company,
let alone as smaller mid tier company.

Mishaal (50:07):
You actually did bring up that Japan example on the,
on that's my favorite podcast.
Yeah.
Speaking of Japan, I think a lot of peopledon't know about the country is that
there's a very specific piece of hardware.
That's pretty much only shippedon devices in that country.
And it's called, uh, NFC dash FICA.
Right?
And like, if you wanna ship a device inthat country, you pretty much have to
make Japan only skew or you risk havingto ship that hardware on every single

(50:31):
other device around the world and, youknow, increasing the cost everywhere else.
Um, yeah, we waste a

Jason (50:35):
lot of money at essential going down the build a FICA only skew that I
don't even think we actually shipped.
Or if we did, we ship like500 units and we spent yeah.
Million

David (50:44):
about there.
We're talking about the countrythat also came up with its a
completely different version of theATM card in authorization system.
Yeah.
They have their own bankcard system in Japan.
That's completely incompatiblewith the rest of the world.
They had a completely incompatiblecellular network iden like
they've always been a little bit.
Different in terms of

Gary (51:02):
tech adoption.
Yeah.
And I think in terms of adoptionin general, they do favor
probably the manufacturers thatcome from their country as well.
So even breaking into the market on numberof units, you're probably moving in that
country considering needing in your ownSKU and having the delicate support.
It becomes an interesting conversationpretty quickly on whether or not

(51:23):
we should be launching there.
And yeah, this is a conversationthat we'll be having, you know,
I'm sure on our next gen device.
Oh,

Mishaal (51:30):
and speaking of favoritism, I'm sure you're aware well aware of India and
like how it heavily favors manufacturersor OEMs that manufacture the majority of
parts of the phone within the country.
Like if you don't manufacture the entirephone inside India using Indian plants,
then you're gonna pay a hefty import.

Jason (51:50):
Yep.
There are basically fourcountries that fall under that
category, which are Russia.
Doesn't really matterthese days, India, Brazil.
And I always forget the fourth one.
It's like some random one in Africa.
Yeah.
We're not buying a lot of, but yeah, Indiaand Brazil being the two main ones where
you have to manufacture, you have to dosome percentage of manufacturing there.
Typically you get away with finalassembly just because it's not like they

(52:13):
have chip fabs in Brazil or India thatcan meet foreign animator processes.

David (52:17):
Right.
And that's why Moto madeso many phones in Brazil.
, that's why they're also so big there.

Mishaal (52:24):
Yep.
And it's also why pixelsare so expensive in India.
. Makes a lot of sense.
Circling back to the carrier example,I kind of wanted to bring up the
software side of things beyondlike the whatever certification
process you have to go through.
From what I've heard, carriers alsooften have software asks, like they might
say, we want you to modify your statusbar icon so that you show our specific

(52:47):
branding for the 5g logo or that youshow our boot animation for your product.
Whenever it boots up.
I'm pretty sure this is actually thedriving force behind one of Android's
most used platform features, whichis the runtime resource overlay,
which is like the feature thatlets OEMs actually overlay like
the resource of an application.
And because carriers were constantlyasking companies to make these changes

(53:10):
to the Android bills themselves,that would have to necessarily create
multiple versions or like complicatethe build of the Android bill that
OEMs are shipping out the deviceslike creating this whole system.
Just so OEMs could say,okay, you can put all your
customizations in the separate app.
And then they'll only apply wheneverthe user inserts a SIM card that's
from that carrier or that deviceis provisioned on that carrier.

(53:32):
So like even these tiny little things,like just having a device sold in
the United States, there's so manycomplicating factors that go into it.
Yeah.

Jason (53:40):
And that's where we get into that.
Oh, if you don't pass their individualcertifications and you ask for waiver,
they're gonna come back and say, well,if you want a waiver, we're gonna need
you to do X, Y, and Z in software.
That might be an example.
And it's just so ING, which is why

David (53:52):
that some cartel level

Jason (53:55):
behavior, if you, if you own all the hardware to make networks work, you,
you do have a monopoly or you have acartel going between the three biggies.
So yeah, that's exactly it.

Gary (54:05):
Yeah.
I think you'll see a lot of carrierbloat on devices that have flagship
carrier, where you see them in stores,they'll have their own agenda to
kind of load in certain softwarethat might compliment their network,
really visual voicemail apps or.
Things that they use as like agateway into other products and
services that they offer as a company.
The worst part of it is they'renot unins installable at most.

(54:29):
You can hide them cuz theyare part of the system image.
A lot of the times for those.

David (54:33):
I mean, I, I do think that trend is changing though.
I believe most bloatware these daysis uninstall because Google has
provided that provisioning mechanismto carriers during onboarding.
Right.
Um, such that even when you loada new SIM card, the device free
provisions and then it says, Hey, doyou wanna load this crap on your phone?
And usually it lets you avoid some of it.

Gary (54:53):
Yeah.
But if you're going through thefull blown, not B by O D route that
we are going currently, then theyhave a lot of hidden agendas to
say, have the boot logo, like you'llalways see like an at T a boot logo.
I don't know if that goes away now.
Uh, I haven't tested that.
I haven't switched, you know, out of mycurrent carrier to see, uh, oh, I haven't

Jason (55:12):
had we've individually been open market for a very long time.

Gary (55:15):
So kinda,

David (55:16):
I don't know, even if Michelle, are you aware if at, and
T's even still branding the bootanimations for like galaxy devices?

Mishaal (55:23):
I, so actually, yeah, I mean, I just bought, I mean, this
is an older device to be fair,but my galaxy tab S five, that's
undergoing the journey of transforminginto an Android automotive tablet.
It came with a stock at and T firmwareand that had an at and T boot logo.
I had to flash the us so firmware,just so I could unlock the boot loader.
And that had, when I did that, thatcame with a cellular boot logo.

(55:45):
And so, yeah, they're still doing itand it's really interesting to see how
they're able to throw their weight aroundbecause of course they're massive in the
us and they can, but like, You're alsostarting to see actually just apps or like
services throw their weight around too.
Like Netflix is one example.
I think a lot of people don'tknow is with wide vine L one.
So even if your device is wide vineL one certified, that doesn't mean

(56:08):
that the device will be able to stream4k videos on Netflix or will be able
to stream HDR videos on Netflix.
I believe Netflix has their own separatecertification process just for that.
And I don't know if you can talkabout that, but I just wanted to
bring it up because it's somethingthat I always had questions about.

Gary (56:25):
I think we're, we haven't gotten there yet with the Netflix
streaming, like a lot of companies dothat to make sure that you're given
the right experience on the hardware.
I know Spotify does this.
If you do anything with alteringtheir application or, or provided
service that they'll have theirown type of certification.
We, we had a lot of that kind of chat whenwe were doing project gem at essential.
So a lot of people are justvery, very protective of how

(56:47):
their services is consumed.
I feel like Netflix is probably one ofthose ones where it's only going to it.
You have to be allowed list.
To get the optimalexperience on that device.
And I'm sure it's a mix of not only thehardware capabilities, but probably your
wifi chip or something that I'm not fair.
We've never gone through it, but I'mcertain that they have a lot of talented
engineers who are doing that for areason, but I don't know, you, you have

(57:10):
any insight on why they would do that.

Jason (57:12):
I honestly do not.
Um, it's something I'll look into though.
I'm quite curious.

David (57:16):
Yeah.
I mean, my guess would be forNetflix's side of things, and this is
speculation, but as a service provider,they're heavily incentivized to
maximize revenue, to like the devicesthat are HDR, 4k certified, and I'm
guaranteeing those are all sold throughcarriers because why wouldn't they be?
And so the carriers probably havesomething on the back end there you're

(57:37):
selling a high end Netflix device.
You get a better bounty when you sellthis device and you get a subscription.
Basically.
There's no way in my mind, 4kHDR, these are very standardized
things, the Kodak, and like you canadvertise your decode abilities.
This is not complex.
This is Netflix playing business.
I think.

Mishaal (57:54):
Right.
Right.
And you'll see this often withpreload agreements, like why do
certain devices have Netflix orother applications preloaded?
Well it's because their OEMs may besigned a revenue sharing agreement with
them, or why does every single TV remoteseem to have a Netflix button on it?
Well, that's probably more of arequirement to even be allowed
to use Netflix on that TVversus anything revenue sharing.

(58:15):
Like there's all these different behindthe scenes agreements or requirements
that 99.9% of consumers will never know.
But as an OEM, it's such a huge,massive headache to deal with.
There's so many differentthings to keep track of.

Gary (58:29):
Yeah, I think there's also incentive for OEMs to have rev share too.
Like we had a lot of considerations ofthat in the past, beyond even awesome
for rev share agreements for pre-loadsand they can be pretty substantial
for companies to have that revenueshare for a small company, but for
larger companies, I'm sure they're alsoincentivized from that other end to
accept an agreement, to put a dedicatedbutton on a remote, like you said, or

(58:53):
even something on the home screen andpart of the onboarding experience.
Like a lot of times we, on the businessside, we were looking at real estate
and real estate comes in the form of theonboarding experience, your home screen.
In a folder or is that dedicated hotseat button, or even as part of some
of the minus one screens that comesalong with the business talks, which is
always interesting from a manufacturer.

Mishaal (59:16):
Yeah.
And this is the benefit of Android,just becoming more and more modular
over time as more and more entitieshave a vested interest in having
more requests, more integrations, andhaving all these things, you need to
be able to manage all of this stuff.
So it's been a few years since Ilooked at like how Samsung's one UI is
distributed, but it is so, so modular.

(59:38):
They have these country specific codesthat enable certain flags within certain
carriers within certain countries.
They have all of this built into, Idon't know how many different shipping
firmware builds, but a company like that.
That's the reason they're able tooffer so many different devices across
so many different channels that arebarely modified from each other.
But like, if you're a smallerOEM, how do you deal with that?

(01:00:00):
It's.
It's gotten harder, but it'salso gotten a little easier with
the modularization of Android.
And of course, if you are going tobe developing your software for your
own device, or you, you wanna make adevice, you're going to be needing some
talented software engineers to do that.
And I believe you both havean announcement regarding

Jason (01:00:19):
that.
Well, we, we touched on it earlier.
I think we, we are keeping a physical SIM.
Was that the big.
Oh, no, that we closed our series a,um, yeah, that was really awesome.
It took a lot longer that I, youknow, had lawyers, uh, but we're very
excited to say we've closed our series.
A, we are not sharingthe exact dollar amounts.
However, it is more than enough tokeep awesome, healthy and running

(01:00:40):
rampant for many, many years to come.
And we're excited to say thatSolana was the, uh, investor
as well as our partner in.
So we've raised quite a bit ofmoney and we're, the main use of
those funds honestly, are not forhardware, not for equipment, not for
marketing, but rather for people.
And we are going to be making a bigpush to hire software engineers.

(01:01:02):
They are backend system engineers,but I think Gary would do a much
better job of explaining exactlywhat awesome is looking for.
Yeah.
So

Gary (01:01:10):
here's the hiring plug.
We are hiring quite a bit on thesystem software engineering side.
So people with BSP experience and haveknowledge with working with different
hardware modules from a mobile phoneperspective, there's so many different
parts that are involved beyond a CPUinteracting with all of that bring up.
So anyone who has C or C plus plusknowledge is what we'll be looking for.

(01:01:31):
Initially, we do have open spotsacross the board for software,
for QA automation, testers.
We also have people who are morefront end and product related
roles as well, and then stuff thatwill definitely interest people.
Uh, there's a lot of opportunitythat you'd be able to work.
Here at awesome.
And we are just looking to matchpeople's interests and experience

(01:01:51):
with what we have to do, not onlyon this first release, but a lot of
other things that we are working oninternally beyond our mobile device, our

Jason (01:01:58):
first mobile device.
Yeah.
Well, I'll throw an extra plugfor just coming to work for us.
Obviously we're a startup thatis a lot of fun to work with.
We're working on newand interesting things.
The first product is a traditionalphone, but certainly the future
products we're gonna get a lot crazier.
And if you're into wild, exciting newprojects where you'll have a lot of say

(01:02:18):
in how it happens and say in what theproducts look like, this is the company
to come work for a hundred percent remote.
So we're hiring all over the world.

Mishaal (01:02:26):
And David, I think you could tell our listeners here what ESER has to offer,
how we're interested in BSP developmentand you know, AOSP development in general.
We sure

David (01:02:37):
are.
So Asper is really in the business ofhelping companies, maybe not unlike
awesome, but probably not in theconsumer space necessarily do things
like provision devices at the factory.
And the way we're helping out withthat is really at the management layer.
And we're making sure that once thosedevices leave the factory, whether
they're flashed with our AOSP firmwareor the manufacturer's firmware, that

(01:03:01):
they can be managed, updated, controlledanywhere in the world at all times.
If you're building something that runsAndroid and maybe it's not something
that goes in somebody's pocket every day.
And is this extremely important lifedevice that's very personal, but
maybe it's something an employee isusing something like a mobile point
of sale device, which looks like aphone, but takes credit cards, or

(01:03:23):
it could be something stationarylike a traditional cash register.
You would just call ita point of sale system.
If you're in that world, you're makingdevices that transact that assist
customers provide customer service,enhance customer experience in any way.
Come talk to us at SPER.
We have so much experience, not justin managing and updating and servicing
these devices from the software.

(01:03:43):
And, but also in helping you identifythe right hardware for your use case,
which is such a big part of this process.
So if you wanna come talkto us, we're at esper.io.

Mishaal (01:03:54):
Thanks, David.
And thanks, Jason and Garyfor joining us on Android.
Vice again.
It's great talking to you both likeI'm sure folks listening will agree
that you're probably some of themost open execs in the tech industry.
Like it's rare to talk to people whoare actually willing to talk about
these things without going throughlayers and layers of PR speaking of PR

(01:04:14):
you haven't heard him talk yet sincethat started the recording, but we
have Ryan Hagar on the show and yeah.
If you know him from his days at Androidpolice, you know, he's a great guy.
He's, uh, very, very well spoken.
And, uh, Ryan, I don't know if youwanna say something to the folks
here, say hi, I didn't hear anything.

Jason (01:04:33):
Hello, close.

Mishaal (01:04:37):
right.
If you're, if you're a PR a journalistlistening to this, you probably
hear from Ryan in the future.
I'm certainly lookingforward to hearing from.
Him and the rest of awesomefor you, whatever you guys are
working on beyond the saga.
But, uh, yeah,

Jason (01:04:51):
we kind of, uh, you know, maybe we should aim to do is once
a quarter or something come in.
So just maybe a shorter conversation.
Just talk about what we're working on.
Random things.
Yeah.
Just a lot of fun to be verytransparent about how products are
made and how decisions are made.
I'll tell you what, as somebody who knowsall the little intricacies of how the
hardware decisions are made every day,when I'm on Twitter or wherever, reading

(01:05:12):
about people's comments about how come,blah, blah, blah, didn't have this, or had
that blah like, oh my God, people don'thave a clue how the rules of Lego inside
a phone or, or any electronic device are.
So there are some that are incrediblystrict and they're super obscure to
anybody who isn't in the industry.
And it'd be really fun to do a, like,just almost a Q and a of why is it like.

Mishaal (01:05:36):
Oh, gosh.
Yeah, I'm trying to, I about to regulateI about to regulate so much of what
I say on Twitter, because like, I'mpretty sure I don't actually know as
much as I think I know about that topic.
So I'm just gonna shut up andjust keep my focus on Android.
So that's why I onlytweet about Android now.
Okay.

David (01:05:51):
I have a question.
We can end it on.
It's a silly one.
Volume rocker.
Love it.
Or

Jason (01:05:55):
hate it, hate it.
I hate it.
It they're they're painto get the feel right.
And I'm, I'm a stickler for the feel.

David (01:06:01):
I agree.
Pixel six, press volume, rocker,terrible best volume rock of all time.
Pixel four, a Ooh.
You know what?
I agree.
Very

Jason (01:06:09):
clicky, very clicky.
Yeah.
Difficult to waterproof.
It also like really difficult to getwa rockers to waterproof nicely because
you have to consider now one side it'sgonna lift rather than just compress.
Every time you push it.
And the buttons, the holes in yourdevice are always the ingress for
water or dust or pocket li so.

(01:06:29):
I'm a big fan of individual buttons.
I'm also a fan of no buttons at some point

David (01:06:35):
I thought you were gonna go there.
That's an interesting position.
HTC tried it and did a very bad job.
I would love to see somebody do it better.
Yeah.

Jason (01:06:42):
The, the haptics are still being built out.
The pressure sensors.
It's it's, it's one of those thingsit's like through display cameras, even
through display fingerprint sensors,it's like almost there, but not quite.

Mishaal (01:06:53):
We definitely gotta get you back on the show.
Whenever the first company goesfully pointless on their device.
I'm sure you'll have alot to say about that.
You

Jason (01:07:00):
know what?
I'll tell you.
Who's gonna scream.
It's software engineers.
Literally I've had conversationseven this week about it.
It's because during developmentyou need to be able to access it.
So you need to have a USB port.
And if there's a problem in MP, it's apain in the butt to try to figure it out.

Mishaal (01:07:15):
Oh gosh.
The wear S smart watch is like,it's already such a hassle,
just the ADB into those things.
You gotta do it all wirelessly.

Jason (01:07:22):
Exactly.
And then if there's actualproblem, You could be screwed.
So not having a plug is a, youknow, it's an interesting question
or a, at least a single port,I can imagine going completely
buttonless, but pointless will be

Mishaal (01:07:36):
difficult.
Well, you know, it's gonna happen.
It's it's gonna happen.
And we can already guess whichcompany's gonna Herald that future.

Jason (01:07:44):
Yeah.
There, there is a gym, a gem that hadno holes in it because we're making
this really complicated piece of glass.
It was like, Hey, it's a lot strongerif I don't have to cut holes in it.
And so we made one and yeah, there wasa lot of yelling I was told to be quiet
and sit in the corner kind of thing.
Pogo pins are just as good.
They're fine.
But I keep saying well, cool.

(01:08:05):
Thank you very much for havingus guys really appreciate it.
Yeah.

Mishaal (01:08:08):
And thank you everyone again for listening.

All Episodes

Episode Transcript

Popular Podcasts

Stuff You Should Know

Dateline NBC

On Purpose with Jay Shetty

.css-15opob5{left:0;position:absolute;top:0.8rem;} All Episodes

.css-14f5ked{margin:0;word-break:break-word;display:-webkit-box;-webkit-box-orient:vertical;box-orient:vertical;-webkit-line-clamp:2;overflow:hidden;}The hidden costs of building an Android phone