Latest / In The Loop / Has Google Just Beaten Apple With The New AI Phone? (Google Pixel 10)

Transcript

Has Google Just Beaten Apple With The New AI Phone? (Google Pixel 10) · In The Loop

0:00
After a number of weeks discussing big AI trends
0:03
throughout the summer of 2025, whether we're in AI bubble
0:07
that's about to pop, it's made me want to do something a little
0:10
bit different now. This week's episode is going to
0:13
discuss the new Google 10 Pixel phone.
0:16
Not because I'm in love with phones, but as you can imagine,
0:19
it's just been given a bunch of new AI features, which funny
0:23
enough, I've taken aim right at Apple.
0:26
If you've been listening for a while, you'll know that this
0:29
particular topic around Apple just keeps drawing me back in.
0:33
So this week we're going to discuss all the new AI features
0:36
that are coming to mobile phones, what that means for the
0:39
Apple Google battle, and a particular area of innovation
0:43
that excites me most. This is In the Loop with Jack
0:46
Horton. I hope you enjoy the show.
1:00
So all started with an ad that's about 30 seconds long, and I'm
1:04
actually going to play it for you just a moment, but it's
1:06
obviously taking aim at a particular company.
1:09
And for anyone that's been following this series, you'll
1:13
know that I've been watching the dumpster fire that is Apple
1:16
Intelligence quite closely. So let me play the video of the
1:20
new Google 10 Pixel advertisement and I think you'll
1:23
see exactly what I'm talking about.
1:26
If you buy a new phone because of a feature that's coming soon,
1:32
but it's been coming soon for a full.
1:34
Year. You could change your definition
1:37
of soon, or you could just change your phone.
1:55
So yeah, that added, it's very short, it's sweet.
1:57
It's quite clever, to be honest. And they're obviously trying to
2:01
get people to swap from the iPhone to Google because
2:04
obviously Apple Intelligence has been a complete failure so far.
2:07
You know, for those that don't know, Apple announced Apple
2:09
Intelligence, this huge big project that promised smarter AI
2:14
features in Siri and a bunch of other features that they
2:17
actually made adverts for that then got pulled and they
2:20
actually got sued because they were doing basically false
2:24
advertising. So in September 2024, they were
2:27
talking about Apple Intelligence coming to iPhone, iPad and Mac
2:31
with loads of different tools and features coming throughout
2:33
their different applications. But obviously you Fast forward
2:36
to today and in 2025, their release mentions none of that.
2:40
And most of those features never came out.
2:42
And meanwhile, Google actually have been rolling out all the
2:45
features that Apple said they were going to roll out.
2:48
And to be honest, much of the big launch event they just did
2:51
and the adverts and the campaigns and the content
2:53
they've been releasing has been focusing on all the features
2:57
that Apple said they were going to release.
2:59
And the reason I'm particularly interested in this is because
3:02
for those that are listening right now, you're in the top 1%
3:06
of, I guess, engaged listeners and people in the world of
3:10
technology. So for the majority of the
3:12
population around the world, the proper ways they're going to
3:16
engage with AI in what I guess is an everyday use case is going
3:21
to be through a mobile phone, because pretty much every single
3:23
person has a mobile phone. So I want to really dig into all
3:26
the different features that I think are worth discussing.
3:29
And so this Wall Street Journal piece on whether Google is
3:32
beating Apple on smartphone AI, it's a really important
3:36
discussion to have. And we'll return back to this
3:37
piece later on. So let's first get into some of
3:50
the really cool features because some of them are really
3:52
interesting and and I'll come on to why I in particular find them
3:55
very interesting from AI. Guess a nerdy technologist
3:58
perspective. So Google's roll out of a eye on
4:01
to all Android devices have actually been ongoing.
4:04
So the Pixel 9 brought out a bunch of really cool things, but
4:08
as you can imagine, this new Pixel 10 update hasn't really
4:11
been focused on new hardware. And I'm not going to delve into
4:14
that because I'm not a technologist reviewer that loves
4:17
all the new fancy camera upgrades.
4:19
What I'm really interested about is how AI has been integrated
4:23
into everyday products. So obviously this release for
4:27
their new Google phone has just been all about new AI software
4:30
updates. So some of the things we're
4:32
talking about is the fact that Google Now comes with Gemini
4:35
built in. So you have something called
4:38
Gemini Live, which actually a friend had a chat with me
4:41
recently and said the reason they're using Android at the
4:43
moment is because of this. So I'll press play to the video
4:46
again. Any phone looks impressive in a
4:50
shiny commercial, but does it still look impressive under your
4:57
bathroom sink? Gemini, my wedding ring just
5:04
went down the drain. Show me where to start.
5:07
Unscrew the slip nuts so you can remove the drain trap.
5:10
Just make sure to put a bucket down.
5:17
So what you just heard there was somebody holding a phone up to a
5:20
pipe. Again, as you heard someone
5:22
saying that their ring just went down the pipe and how can I now
5:26
get to it? And Gemini Live just from
5:28
pressing a button that's already on the phone's home screen
5:32
immediately takes you to the best of the best of the new AI
5:35
capabilities built into Gemini. So not only have they built this
5:38
really cool feature into the product, they've also managed to
5:42
put the model on the mobile phone.
5:43
And I'll come on to that in a minute.
5:45
Another really cool thing they managed to do is release
5:47
something called Magic Q, which is basically this connector
5:51
between all of your applications through a conversation.
5:55
So the examples they had, I'm not going to play the video, but
5:58
some of the examples they have is your friend might say, have
6:01
you got the address for the location you're going tonight?
6:03
And instead of you needing to then go into another app and
6:06
then check your phone actually understands the question and
6:09
goes to the relevant application and pulls that information in.
6:13
So that could be used cases such as when's your flight, you know,
6:16
all all the information that you might have on your phone,
6:18
because obviously your phone has so much context about you, your
6:22
world, your life built in. And what they're trying to do
6:27
here with Magic Q is a really quick and easy way for that AI
6:31
to integrate into everything. So another really cool feature
6:33
they're just bringing out is conversational editing
6:37
capabilities for your camera roll.
6:39
So you can now take a picture and you could literally just
6:42
speak to it as if you're having a conversation with GPT image
6:44
generator and ask it to do various different things with
6:47
your photos right in your application.
6:49
And from the reviews online, it's actually supposed to be
6:52
fantastic. Again, we have to remember that
6:54
Google are actually bringing up the very best models in image
6:58
and video generation with sound pre built in.
7:01
So there's a lot of great ability and opportunity for
7:04
Google to integrate this into their mobile phone.
7:06
You can ask it to say remove the cars or restore, restore this
7:10
old photo, which is this particular one that I'm actually
7:13
playing in the video. If you're watching it on any
7:15
device or app that lets you play videos, again, many of these
7:18
features that we're talking about is what Apple advertised.
7:21
You know, that ability to have your AI integrate into all of
7:24
your favorite applications and give you the answer soon as a
7:28
friend asks the question. So where, where are you heading
7:30
tonight? It will just pull the answer
7:32
straight from your calendar or for a previous text message or
7:36
whatever it might be and give it to the person.
7:39
So yeah, a bunch of really cool things coming out for Google.
7:41
Another really cool thing is they've actually managed to put
7:45
the language model on the mobile phone.
7:47
So the optimizer model for the phone, which is a really big
7:49
step for the ability to have fantastically fast responses and
7:54
everything just work on a mobile phone.
7:55
And now this on device model, which is about 1.8 to 3.2
7:59
billion parameters, which is very big for such a small
8:02
device. Basically, Willpower automatic
8:05
transcription, caption generation, a bunch of different
8:09
features that I talked through. And it will keep all of the data
8:12
private because it's actually held locally, so held on your
8:15
device. So this is all well and good.
8:17
So I could list all these nuclear features all day.
8:19
But I guess the real question then is how is this being
8:22
received? So I think Reddit is always a
8:24
really good place to sense check how people are feeling.
8:27
And really the description from many people is calling it
8:31
incremental. I think generally what the
8:33
mobile phone technology space people perceive anything that's
8:37
just a software update is incremental and they expect
8:40
massive hardware leaps that Apple used to bring about every,
8:43
you know, 3 or 4 years. So therefore nothing has ever
8:45
received too well in the mobile phone industry.
8:48
And as we all know, it's becoming harder and harder to
8:51
make people upgrade or switch devices.
8:53
So this person here, anyone else not buying the Pixel 10 after
8:56
being underwhelmed? But for many reasons, I think
8:58
this is quite unfair. I think this person hits a nail
9:00
on the head. You know, the Pixel 10 seems to
9:02
be the right direction. And I just don't understand the
9:04
criticisms because predominantly this, the argument hinges on the
9:06
fact that Google are bringing out a bunch of new things.
9:10
And the people that are criticizing it are saying that,
9:13
well, there's a bunch of things that are the exact same, but
9:15
you're just rebranding it, which is what Apple do every single
9:18
year as well. So yeah, I'm not defending the
9:20
stagnation, but I think people are often overly critical when
9:25
Google or any mobile phone manufacturer does anything these
9:27
days, especially when it's got AI in it, because ultimately
9:30
these are big leaps in software. Yes, there are no massive
9:33
hardware changes yet, and maybe that'll change in the next 12 or
9:36
18 months when Jonny I've and Sam Altman bring out their
9:39
latest cool device, but maybe not.
9:42
And I think Marquis did a fantastic video which really
9:44
broke down the psychology of this, I guess, barrier for a
9:47
Google or an Apple these days. And he said in that video that
9:52
65% of people are unwilling to switch even if it saved them
9:56
money. I'm one of them right now.
9:58
Now this has got my head a little bit turned.
10:00
I want to experience what it's like, but am I going to switch
10:03
from Apple to Google? I honestly don't know.
10:07
I don't think so. Anyway, the nerd in me finds
10:10
this all very interesting for a different type of reason now,
10:13
and that's context engineering. If you haven't listened to the
10:16
Context engineering episode, I think it was episode 23, do go
10:19
check that out. The reason it's so interesting
10:22
is because of the challenge of making AI really work at scale
10:26
and is something that's useful. Context becomes key.
10:29
So for those that don't remember, context is what the
10:33
language model already knows at the point of you asking a
10:35
question. So the job of a technology
10:38
provider is to feed the language model with the perfect context
10:42
at the right time. And if you think about how hard
10:45
that can be when you've got lots of context, well, then I'm
10:49
always really interested about how people have done it and done
10:52
it well. And because ultimately mobile
10:54
phones have so much context on us and our lives.
10:57
It's got access to my emails, my calendar, my previous text
11:01
messages, my application usage. There is so much data there that
11:05
adds context to who Jack is. That if you can find the right
11:09
way to feed the right thing to a model, suddenly you can create
11:12
this assistant of AI that knows me so well and can actually take
11:18
over things in my life that I would love to hand off.
11:21
Google built something called the Context Hub, which is
11:25
essentially a way of an application on a very low power
11:29
chip. So a chip that's on a mobile
11:31
phone to basically be constantly listening to data everywhere and
11:35
then feed into all their key assistant features.
11:39
So all of these can therefore trigger an AI response.
11:41
So if your camera sees text, it can trigger the feature called
11:44
Circle to search feature to say highlight what that is.
11:48
If it sees a landmark, it can use a visual lookup feature that
11:53
then identifies it. There's also going to be a new
11:55
feature called Tell Me About This, which essentially lets the
11:58
Gemini, so the language model on the phone analyse any part of
12:02
your screen that you wish. So yeah, that context
12:04
engineering part is what I find very interesting.
12:07
And ultimately really we're we're running out of of many
12:11
people would argue we're running out of data on the Internet to
12:14
feed into models. You know, we've downloaded and
12:17
used all The Pirate Bay, if you believe the rumours, all of the
12:20
Internet. And that's something that we've
12:24
spent many, many years building up a data reserve, but now we're
12:28
entering a world in which we're going to have to find new data
12:31
sources to make these things even more powerful.
12:33
See, I think this is a really important part of, I guess,
12:36
making AI work on a mobile phone or a small device that people
12:39
often don't realise. So what does all of this mean
12:53
for Apple? Yeah, they're clearly behind.
12:55
It's becoming harder and harder to ignore how far behind they
12:58
are. And I mean, if you've been
13:00
following this series for a while, then you'll have listened
13:02
to the episodes I've covered Apple and what I think they
13:05
should be doing right now if they're going to leapfrog
13:07
everybody else, which is by one of the largest language model
13:10
labs integrated into their technology very, very quickly.
13:15
Because really, although it's a high switching cost and and when
13:18
I say high switching cost, people don't like switching from
13:20
an Apple to a Google, for example, you never know,
13:23
eventually it could end. They've still got this huge
13:26
ecosystem of products that makes it even harder to change.
13:30
But again, Apple also have much of the equivalent there.
13:33
But Apple is starting to be more pragmatic now.
13:37
Again, as I say, I I said this about 5 months ago, they should
13:40
go buy one of the large language model labs and take all the
13:43
talent there and bring them internally.
13:45
But now they're considering whether they should be using
13:47
Gemini in their Siri features. If Apple does end up sourcing
13:52
Gemini like models from say Google, their whole strategic
13:56
narrative and entire strategy for the last 1520 years of
14:01
owning the entire stock will come to an end.
14:03
Again, remember Apple's always had a strategy of trying to own
14:06
and integrate the entire stock perfectly for the best possible
14:10
phone. But it's arguably now less
14:12
important, maybe because actually every phone application
14:16
and manufacturer, because of the likes of Android, is able to
14:19
produce really fantastic devices, whereas 1520 years ago
14:23
that was far from the case. So yeah, it raises very
14:25
interesting questions about, I guess the value of owning the
14:28
entire stock. And now Google's marketing line
14:31
is that they have all of the technology stock they need to
14:34
produce incredible technology. They own the model, they own the
14:38
hardware, they own every part of the stock.
14:40
Now they even build their own chips.
14:53
To conclude, what's striking about this entire situation, I
14:57
guess, is how it reflects the broader changes in the
15:00
smartphone industry we're moving from this year, where hardware
15:02
improvements used to drive much of the innovation.
15:05
And that's what the market expects.
15:06
Big hardware changes like Apple used to bring all of the time to
15:11
a world where software becomes increasingly important, and
15:14
specifically AI. Google seems to understand this
15:17
shift and is acting on it fast. Apple are far, far behind them.
15:21
The real test is going to be whether consumers actually care
15:25
about these features and actually swap to the Google
15:28
Pixel 10. I, for one, am not there yet.
15:31
But again, I've never been so tempted.
15:33
But really, what this is all about is not who controls the
15:35
mobile phone industry, but who controls the entire user
15:39
experience for most consumers. And right now, Google seems to
15:43
be taking the most meaningful step to owning that experience.
15:47
Anyway, thank you for listening. I'll see you next week.