0:01 We reached out to Mr. Who's the Boss to
0:03 find out how he would design his own
0:05 dream OS. In this video, we're going to
0:06 cover all the features, how much it will
0:08 cost to develop, and we've even mocked
0:09 up the design to show Aaron for the end.
0:11 So, let's get into it. Mr. Who's the
0:13 Boss or Aaron has reviewed thousands of
0:14 tech products over his career. And in
0:16 our last video, we designed his dream
0:17 smartphone, and here's what he had to
0:19 say about it.
0:22 It's a very interesting mix of curvature
0:24 with geometry. Well, you have my money.
0:25 I'm very curious to see what you've done
0:26 with the software on this thing. We're
0:28 setting out to design the perfect OS for
0:30 that hardware. But it's going to be a
0:31 little bit trickier than you might
0:33 think. Luckily for you, split the whole
0:35 process into five easy to digest stages.
0:37 Starting with number one, planning. For
0:38 this specific OS, we're going to project
0:40 an ambitious development timeline of 12
0:42 months. This is actually unheard of in
0:43 the industry, but given that this is
0:44 Aaron's dream smartphone, we ain't going
0:45 to keep him waiting. This stage
0:47 basically consists of a small team of
0:49 senior members getting together and
0:50 outlining what they want from this new
0:52 system. For example, Aaron has requested
0:53 very toned down visuals kind of like
0:55 nothing OS but with the addition of
0:57 notifications that glow so the system
0:58 can highlight what's most important to
1:00 him. When Nothing OS goes for a very
1:01 subtle approach where no app in
1:03 particular draws your attention, Aaron
1:04 wants the relevant app to glow for you
1:06 to instantly direct your eye to it. Now,
1:08 this could be as simple as adding some
1:10 glowy lights around the notifications
1:11 that have most recently come in, but we
1:12 probably want to add some AI logic to
1:14 this to feel like it really knows you.
1:15 So, for example, if your house is
1:17 burning down, you probably want to know
1:18 about that first. Things like this would
1:20 be outlined by that small group of
1:21 senior team members. They're going to be
1:22 looking at things like project
1:24 timelines, budgets, and resource
1:26 allocation. And this whole process lasts
1:28 for about 2 months. Now, you may notice
1:29 that the time given for each of these
1:31 five stages actually totals more than 12
1:33 months. Before you get in the comments
1:34 telling me my math is bad, some of these
1:36 stages are happening simultaneously to
1:39 make things more efficient. So,
1:41 shut up.
1:43 Now, on to stage two, platform
1:45 development. There are two approaches to
1:47 making an operating system these days.
1:49 The first is to build from the ground
1:50 up, which is very expensive, labor
1:51 intensive, and challenging. Or the
1:53 second one, which is way more accessible
1:55 to young startups, is behind door number two.
1:56 two.
1:58 >> What's that?
2:00 >> They'll add like a hinge sound effect.
2:02 It's the door, door number two. >> Right.
2:04 >> Right.
2:06 >> The second easier option is to fork
2:08 AOSP, otherwise known as Android Open
2:10 source project, the basics of all great
2:11 smartphones. Now, if you'd ever been on
2:13 GitHub before, you've probably heard of
2:14 this phrase, but forking is where you
2:16 take someone's existing chunk of code
2:18 and add some of your own on top to
2:19 create something new. This gives you a
2:21 head start in development by giving you
2:23 a baseline OS to make your own. And if
2:24 you pair it with GMS, you'll benefit
2:26 from improved security having a
2:28 functioning app store. And also, there's
2:29 a whole bunch of Android developers out
2:30 there already. So, it's just going to
2:32 make things a bit easier for us. Now,
2:33 you might think this sounds a bit lazy.
2:35 Why not make your own OS? But actually,
2:37 it took Huawei 4 to 5 years to
2:38 transition away from Android with
2:40 Harmony OS. And even then, that still
2:41 kind of looks a little bit like Android,
2:43 doesn't it? So, no complaining in the
2:44 comments. Also, we've been told that
2:46 adapting AOSP isn't quite as
2:47 plug-and-play as it used to be. Now that
2:49 Google has their own lineup of devices,
2:50 stock Android's looking kind of
2:52 barebones, and you have to work even
2:54 harder to meet basic expectations. For
2:56 example, you might remember nothing OS
2:58 1.0 on phone 1. It was a kind of basic
3:00 experience. And now, compare that to
3:02 nothing worse 4.0 nearly 3 and 1/2 years
3:04 later, looking a lot more elevated, a
3:06 lot more elegant. Quick sidebar, the
3:07 Android nerds out there are probably
3:08 wondering why we haven't just done our
3:10 own custom launcher. But custom
3:11 launchers are great, but quite
3:13 restrictive. You can't change anything
3:14 in the system settings and you can't
3:16 optimize for your own hardware. So,
3:17 let's move on. Now, the whole
3:19 development process is projected to take
3:21 about 6 months with a team of about 300
3:22 software engineers. And I have to stress
3:24 that this is way quicker than industry
3:26 norms. And for context, when phone one
3:27 launched, our whole company was less
3:29 than 300 people. So, it's a miracle that
3:30 that phone even launched in the first
3:33 place. Now, on to stage three, testing,
3:35 quality assurance, and compliance. Now
3:36 that we've done all the UI designing and
3:38 engineering, it's time to get into
3:39 testing. And you might have seen some of
3:41 our other product testing videos, but
3:42 sadly, these tests don't involve
3:44 throwing things downstairs. So, how
3:45 exactly do you test software? Well,
3:46 we'll break this down into two
3:48 categories. The first of which is
3:49 professional testing. These are
3:51 automated tests carried out by running
3:53 hundreds of scripts to test how the
3:54 device performs while running multiple
3:56 apps simultaneously and also measuring
3:58 how it holds up over time. And the
4:00 second category is user trials. Similar
4:01 to finding bugs in a video game, the
4:03 team go through every inch of the
4:04 software to see if there's any funny
4:06 glitches that crop up or if there's any
4:07 more serious stuff that could end up
4:09 hurting your device. Now, this actually
4:10 has a fun industry term known as dog
4:12 fooding. And this is originates from
4:14 where pet food manufacturers would eat
4:15 their own products to prove that they
4:17 were safe for consumption. Alongside
4:18 internal teams, companies will also make
4:20 use of external beta testers and
4:22 sometimes include their own community.
4:23 This helps find things that algorithms
4:25 might not pick up, like inconsistent
4:27 typing or maybe some things that just
4:28 aren't very user friendly. This is all
4:30 super important. Now, it might only be
4:32 software, but code can be unpredictable
4:34 sometimes. So, it's important to do
4:35 these tests to make sure the OS is
4:37 working as we expect. As part of this,
4:38 the team also does a bunch of security
4:40 testing and threat modeling to make sure
4:42 the phone is safe against any potential
4:44 vulnerabilities. Now, this process isn't
4:45 perfect, and sometimes things do slip
4:47 through the cracks. If you remember back
4:49 to February, Apple had an issue where
4:50 people could bypass their lock screen
4:52 and holy, that was embarrassing. And
4:54 with this, it takes about 6 months in
4:56 total to pull off. Now onto the panomous
4:58 stage, hardware implementation. Now this
5:00 stage is pretty short and sweet.
5:01 Hardware implementation is where you
5:02 take your final device and you install
5:04 the software onto it. This is so you can
5:05 make sure everything's working as
5:07 intended and gives you a chance to
5:09 implement any hardware optimizations and
5:11 tuning. At this stage, the software team
5:12 will work directly with the factories to
5:14 make sure the OS comes pre-installed on
5:16 all of the devices. In total, you can
5:18 expect this stage to span over 3 months.
5:20 And with that, we move on to our final
5:23 stage, number five, deployment and
5:25 launch. So, it's been 12 months of hard
5:26 work. You've banded together as a team
5:28 to make something beautiful. And now
5:30 it's time to prepare to ship this out to
5:31 all those lovely people who can't wait
5:32 to get their grummy little mitts on your
5:34 new phone. This stage is basically the
5:36 final polish. You're getting that final
5:38 OS build ready, installing any
5:39 pre-installed apps. Shout out you
5:41 Samsung. And preparing for that
5:43 allimportant day one patch. This is the
5:44 amalgamation of a year of hard work and
5:46 all that initial planning has paid off.
5:48 But the work definitely doesn't stop
5:49 there for the software teams. Software
5:51 is a living and breathing thing with
5:53 updates continuing to be pushed after
5:54 the devices launched to bring new
5:55 features and improvements to the
5:57 community. So bear in mind whilst this
5:58 might be the last stage of this video,
6:00 it's definitely not the last stage of
6:01 work for our software teams. We got to
6:03 keep cooking. With that said, at the
6:04 start of the video I said we were going
6:05 to give you some figures. So how much is
6:07 this all going to cost? Obviously every
6:09 company pays people based on different
6:10 benchmarks, but we've done some napkin
6:12 maths, so this is what you could roughly
6:13 expect to pay. There's a lot that goes
6:15 into this final cost. There's the price
6:16 of devices and the cost of the testing
6:19 lab, user research and design testing,
6:21 cloud GPUs, and the tools that the
6:22 team's going to need like Figma for
6:24 designing and all the engineering and
6:25 design team salaries. But overall,
6:30 you're looking at spending $40 million.
6:33 Isn't that more than the hardware?
6:36 >> Yeah, that was about 26 mil. >> What?
6:38 >> What?
6:39 Anyway, it's time for us to show Aaron
6:41 the final design of his new OS. And
7:16 [Music]
7:17 I mean, there's like a thousand things
7:19 to take in there. Let me just let me do
7:21 this one more time. This right here, I
7:23 actually love this. If I could open up
7:25 my home screen and the one icon that
7:28 I've decided has a notification from
7:30 someone I've decided is important is the
7:32 only icon that's glowing, that is
7:34 amazing at just like directing my
7:35 attention to that thing and that thing only.
7:36 only.
7:38 >> Yeah. And it's the trick that app icon
7:40 designers use is like drawing your
7:41 attention in with color. So if we take
7:43 all the color away and use that same
7:45 kind of philosophy more intentionally.
7:47 >> It's really it's beautiful. I'm not sure
7:49 about the um the screen kind of melting
7:50 together. That might just be a strange
7:53 transition. This I like though. This I
7:54 really like. We've seen companies do
7:57 priority notifications before, but it's
7:59 never been smart enough to be useful.
8:00 This is almost contingent on a
8:02 hypothetical world where it's better
8:04 than it currently is. I love how the
8:06 number one priority notification that
8:07 nothing has subscribed to you on
8:09 YouTube. I quite like the idea that you
8:11 see right now the brightness slider is
8:12 glowing. The brightness slider is only
8:14 glowing if there's something to either
8:16 look at on it or you're currently
8:19 interacting with it. Then they revert so
8:20 that they're not the primary point of
8:22 attention. I'm not sure I could deal
8:24 with these transitions every single time
8:26 I open up a menu. Back of the phone
8:28 looks so spicy here. It
8:29 >> is very Tron, isn't it?
8:32 >> Yeah. Amazing. Guys, you killed it. Uh-huh.
