1
00:00:00,120 --> 00:00:03,440
So Open AI have just brought out
their own AI powered browser 

2
00:00:03,800 --> 00:00:05,920
just months after Perplexity 
brought theirs out. 

3
00:00:06,760 --> 00:00:09,640
Many people online have tried to
claim that this is the start of 

4
00:00:09,640 --> 00:00:13,160
a new browser war or that Google
is suddenly going to get 

5
00:00:13,160 --> 00:00:16,040
dethroned. 
And we actually saw a huge 

6
00:00:16,040 --> 00:00:19,720
amount, billions in fact, of 
their value as a company 

7
00:00:19,720 --> 00:00:22,240
collapse in the space of just 48
hours. 

8
00:00:22,440 --> 00:00:25,440
As always, there's been lots of 
hype with this new release and 

9
00:00:25,440 --> 00:00:27,680
in today's episode I'm going to 
break down what I really think 

10
00:00:27,680 --> 00:00:32,320
about it and why I do not think 
this browser is going to take on

11
00:00:32,320 --> 00:00:34,040
Google. 
This is in the loop with Jack 

12
00:00:34,040 --> 00:00:36,040
Horton. 
I hope you enjoy the show. 

13
00:00:47,900 --> 00:00:51,380
So as the title and intro I 
think of this episode explains, 

14
00:00:52,060 --> 00:00:55,980
I don't actually think Open a 
Eyes browser called Atlas or the

15
00:00:55,980 --> 00:00:59,180
Perplexity browser is going to 
dethrone Chrome. 

16
00:00:59,640 --> 00:01:03,080
You know, I don't think it's 
going to capture even 20% of the

17
00:01:03,080 --> 00:01:06,840
market. 
And I may look back in six, 1218

18
00:01:06,840 --> 00:01:10,080
months, three years and eat my 
words, but I don't even think 

19
00:01:10,080 --> 00:01:15,160
it's going to become part of 
most people's process or daily 

20
00:01:15,160 --> 00:01:17,080
workflows. 
But in many ways, I also think 

21
00:01:17,080 --> 00:01:18,480
that's the wrong way of looking 
at it. 

22
00:01:18,480 --> 00:01:21,160
And it's definitely the wrong 
question because I don't think 

23
00:01:21,200 --> 00:01:24,600
that question should be, is this
going to be Chrome? 

24
00:01:25,400 --> 00:01:29,280
I think the more useful question
is, does Atlas solve specific 

25
00:01:29,280 --> 00:01:32,200
problems we have today? 
And I guess what can you learn 

26
00:01:32,200 --> 00:01:34,480
about testing it? 
So when that was launched just 

27
00:01:34,480 --> 00:01:38,280
over a week ago, Google's stock 
price plummeted. 

28
00:01:38,280 --> 00:01:42,200
It closed about 2% down overall.
At one point, it dropped by 

29
00:01:42,200 --> 00:01:44,800
4.8%. 
You know, some people just made 

30
00:01:44,800 --> 00:01:47,000
a lot of money very quickly if 
they bought at the dip. 

31
00:01:48,000 --> 00:01:51,000
But that's roughly $18 billion 
in market value that just 

32
00:01:51,000 --> 00:01:53,640
vanished overnight. 
And, you know, headlines have 

33
00:01:53,640 --> 00:01:58,280
been talking about the return of
the browser wars, but within 48 

34
00:01:58,280 --> 00:02:01,000
hours, it's been confirmed by 
TechCrunch and many others that 

35
00:02:01,000 --> 00:02:04,040
Atlas is actually built on 
Chronium, which is Google's open

36
00:02:04,040 --> 00:02:06,640
source infrastructure for 
building web browsers. 

37
00:02:06,800 --> 00:02:09,240
I'll get into why I think this 
is really important in more 

38
00:02:09,560 --> 00:02:12,760
depth throughout the episode, 
but it just shows, I don't think

39
00:02:12,760 --> 00:02:15,080
Open AI is even competing with 
Chrome at an infrastructure 

40
00:02:15,080 --> 00:02:16,960
level. 
They're trying to build this new

41
00:02:16,960 --> 00:02:32,840
interface into the Internet. 
So before we really get into the

42
00:02:33,080 --> 00:02:36,480
the meat and potatoes of the 
discussion today, I think it's 

43
00:02:36,480 --> 00:02:39,520
worth summarising what Atlas is,
what they're doing differently, 

44
00:02:40,080 --> 00:02:41,800
whether or not you should even 
pay any attention. 

45
00:02:42,480 --> 00:02:45,560
So when this was actually 
launched, Sam Altman described 

46
00:02:45,560 --> 00:02:48,320
Atlas as basically representing 
this like Rare wants in a 

47
00:02:48,680 --> 00:02:51,960
decade, opportunity to rethink 
what browsers could ultimately 

48
00:02:51,960 --> 00:02:54,200
be. 
And the actual product itself 

49
00:02:54,200 --> 00:02:57,600
centres on really 2 main 
capabilities and features, which

50
00:02:57,600 --> 00:03:02,080
is context aware Chat Chi BT 
integration and it's agent mode.

51
00:03:02,280 --> 00:03:05,240
So that context aware piece 
basically means that Chat Chi BT

52
00:03:05,400 --> 00:03:09,000
lives in the sidebar of Atlas, 
which is very similar to what 

53
00:03:09,000 --> 00:03:10,880
Perplexes did with their browser
called Comet. 

54
00:03:11,040 --> 00:03:14,040
And it essentially means that as
you speak to Chat Chi BT in this

55
00:03:14,040 --> 00:03:16,400
sidebar, it understands what's 
on the screen and I guess it 

56
00:03:16,400 --> 00:03:18,480
understands the context of what 
you're talking about. 

57
00:03:18,640 --> 00:03:20,920
So essentially it's just 
removing the need to copy and 

58
00:03:20,920 --> 00:03:22,280
paste. 
So if you're looking at a web 

59
00:03:22,280 --> 00:03:24,760
page and you say and ask a 
question, it already knows what 

60
00:03:24,760 --> 00:03:26,280
you're referring to on that web 
page. 

61
00:03:27,080 --> 00:03:28,480
So it's not entirely new 
technology. 

62
00:03:28,480 --> 00:03:31,800
As I said, Perplexity launched 
it in July, which has many 

63
00:03:31,800 --> 00:03:34,160
similar features. 
But the integration with ChatGPT

64
00:03:34,400 --> 00:03:38,200
obviously means that that 
browser Atlas is now in the 

65
00:03:38,200 --> 00:03:42,000
hands of potentially 800 million
users because that's how many 

66
00:03:42,000 --> 00:03:45,600
people use Open AI's ChatGPT 
every single week. 

67
00:03:45,760 --> 00:03:48,760
Now, Agent mode is where things 
could get both interesting, but 

68
00:03:48,760 --> 00:03:50,720
for me, I found quite 
challenging open air. 

69
00:03:50,720 --> 00:03:54,640
I tried to pitch it as ChatGPT, 
being able to take actions on 

70
00:03:54,640 --> 00:03:57,480
your behalf inside the browser. 
And the examples they included 

71
00:03:57,480 --> 00:04:01,280
are kind of like go do my 
shopping or adding recipe 

72
00:04:01,280 --> 00:04:04,440
ingredients to a cart or placing
specific orders for specific 

73
00:04:04,440 --> 00:04:06,840
things or researching 
competitors and compelling 

74
00:04:06,840 --> 00:04:08,880
insights and putting them into a
document. 

75
00:04:09,400 --> 00:04:12,600
You know, and there's a lot of 
mixed results from people online

76
00:04:12,600 --> 00:04:14,560
and early testers and definitely
from myself. 

77
00:04:14,720 --> 00:04:17,440
For example, one security 
researcher called Simon 

78
00:04:17,440 --> 00:04:21,720
Williamson wrote that actually 
watching agent mode A felt like 

79
00:04:22,040 --> 00:04:24,680
watching a toddler try and use a
computer for the same first 

80
00:04:24,680 --> 00:04:30,040
time, but actually realized that
you could very easily inject 

81
00:04:30,040 --> 00:04:34,800
prompts and sabotage the browser
and the agent whilst it's 

82
00:04:34,800 --> 00:04:36,560
performing tasks. 
So it's actually a big security 

83
00:04:36,560 --> 00:04:39,520
concern. 
Also there's many reports on how

84
00:04:39,520 --> 00:04:42,000
it's just incredibly slow, it 
gets stuck in loops and 

85
00:04:42,000 --> 00:04:44,440
sometimes just fails at very 
simple tasks that would have 

86
00:04:44,440 --> 00:04:47,680
taken a human less than 30 
seconds to complete themselves. 

87
00:04:47,840 --> 00:04:49,600
And that's definitely been my 
experience. 

88
00:04:49,960 --> 00:04:52,680
I've not really been a fan of it
to be honest so far. 

89
00:04:53,520 --> 00:04:55,920
I hope it gets better because it
definitely represents some 

90
00:04:55,920 --> 00:04:58,720
really cool ideas. 
But as I've said in previous 

91
00:04:58,720 --> 00:05:02,040
episodes, it taking control of 
the screen just feels a little 

92
00:05:02,040 --> 00:05:05,080
bit pointless when you just use 
APIs. 

93
00:05:05,080 --> 00:05:08,960
So transferring data behind the 
scenes for you between agents, 

94
00:05:08,960 --> 00:05:11,160
it's just quicker. 
It's the type of experience that

95
00:05:11,160 --> 00:05:14,080
I look for. 
Not this, you know, clunky 

96
00:05:14,080 --> 00:05:17,040
browser that pops up with an 
agent confused trying to 

97
00:05:17,040 --> 00:05:19,840
navigate websites. 
But that that's quite a 

98
00:05:19,840 --> 00:05:21,960
pessimistic view on this new 
capability. 

99
00:05:21,960 --> 00:05:25,800
Of course, they've also added a 
memory feature Atlas. 

100
00:05:25,800 --> 00:05:30,120
So the Browser Atlas now starts 
to track your browsing behaviour

101
00:05:30,320 --> 00:05:32,560
and of course, build up more 
context over time. 

102
00:05:32,560 --> 00:05:36,240
And according to Open AI, this 
will enable you to search and 

103
00:05:36,240 --> 00:05:40,080
say, find all the job postings I
looked at last week or summarize

104
00:05:40,080 --> 00:05:42,520
all the industry trends and 
reports that I've been reading 

105
00:05:42,520 --> 00:05:44,840
over the past month. 
And obviously that means it's 

106
00:05:44,840 --> 00:05:48,960
just very personalized and the 
trade off obviously being 

107
00:05:49,080 --> 00:05:52,000
massive privacy implications. 
I know I've said this again and 

108
00:05:52,000 --> 00:05:55,000
I keep on saying it. 
I have a tinfoil hat at times 

109
00:05:55,000 --> 00:05:58,920
and I'm going to put it on when 
ChatGPT is using my conversation

110
00:05:58,920 --> 00:06:00,600
history, some of which might be 
private. 

111
00:06:01,040 --> 00:06:03,640
It's using it's memory about all
the things that I've talked 

112
00:06:03,640 --> 00:06:05,560
about. 
It's maybe using context I 

113
00:06:05,560 --> 00:06:08,200
provided to it in the form of 
files and documents. 

114
00:06:09,000 --> 00:06:12,080
I don't want it to use that to 
sell more things to me and 

115
00:06:12,080 --> 00:06:14,160
advertise more things to me, to 
be honest. 

116
00:06:14,320 --> 00:06:17,320
Now, if we actually then look at
what's working with today's 

117
00:06:17,320 --> 00:06:20,840
early users. 
So the context aware assistance 

118
00:06:21,520 --> 00:06:22,960
definitely helps with some 
things. 

119
00:06:22,960 --> 00:06:24,920
I've definitely enjoyed using it
in some tasks. 

120
00:06:24,920 --> 00:06:26,800
Now it's only available for Mac 
users. 

121
00:06:26,800 --> 00:06:28,800
So sorry everyone that's on 
Windows. 

122
00:06:29,040 --> 00:06:32,040
Windows is rubbish anyway. 
So, you know, it's a, it's a 

123
00:06:32,040 --> 00:06:34,440
lesson there, but there's been 
lots of people on Twitter 

124
00:06:34,720 --> 00:06:38,400
basically getting Atlas to help 
with improving landing pages, 

125
00:06:38,600 --> 00:06:42,000
suggesting improved thumbnails 
or without kind of switching 

126
00:06:42,000 --> 00:06:43,520
context. 
So just setting it a task in the

127
00:06:43,520 --> 00:06:46,680
browser and it going doing it. 
Another Twitter user called Raza

128
00:06:46,680 --> 00:06:50,600
Martin actually compared both 
Atlas from Open AI and Comet 

129
00:06:50,600 --> 00:06:54,480
from Perplexity on specific 
tasks such as tracking her 

130
00:06:55,160 --> 00:06:57,600
daughter's school assignments 
across many different classes 

131
00:06:57,880 --> 00:07:00,880
and apparently reported that 
Atlas completed it much more 

132
00:07:00,880 --> 00:07:03,160
accurately whilst Comet really 
struggled. 

133
00:07:03,360 --> 00:07:05,520
Now, I do recommend you guys go 
test it. 

134
00:07:05,760 --> 00:07:08,520
And if you're going to go test 
it, I'd approach it somewhat 

135
00:07:08,520 --> 00:07:11,120
systematically. 
It's quite a big ask to change 

136
00:07:11,120 --> 00:07:12,560
browsers actually. 
It's a weird one. 

137
00:07:12,560 --> 00:07:14,560
It's like trying to change 
mobile phone providers. 

138
00:07:15,080 --> 00:07:18,120
There's definitely a, for me, a 
resistance to doing it. 

139
00:07:18,120 --> 00:07:21,520
I'm so used to using certain 
things over others, even as 

140
00:07:21,520 --> 00:07:24,400
somebody that's quite an early 
adopter of new technologies. 

141
00:07:24,720 --> 00:07:27,400
But if you do want to test it, 
I'd recommend trying it for a 

142
00:07:27,400 --> 00:07:30,520
few days to maybe a week and 
trying to focus on specific 

143
00:07:30,520 --> 00:07:34,520
things that you would use that 
context aware sidebar for. 

144
00:07:35,040 --> 00:07:38,520
So without touching agent mode, 
you know, draft social media 

145
00:07:38,520 --> 00:07:43,120
posts using this information, IE
the web page or, or use this 

146
00:07:43,120 --> 00:07:46,880
page to write a summary e-mail 
and send it to John or improve 

147
00:07:46,880 --> 00:07:48,760
my presentation that I'm working
on right now. 

148
00:07:48,960 --> 00:07:51,400
You could also maybe use it for 
learning, you know, on YouTube, 

149
00:07:51,400 --> 00:07:53,880
you can ask questions about 
what's on the YouTube video. 

150
00:07:54,720 --> 00:07:56,640
You could ask clarifying 
questions about what you're 

151
00:07:56,640 --> 00:07:59,480
reading without obviously having
to explain or put web pages and 

152
00:07:59,480 --> 00:08:01,840
provide any context. 
And then obviously later I'd 

153
00:08:01,840 --> 00:08:05,880
recommend trying agents probably
for low friction or quite simple

154
00:08:05,880 --> 00:08:09,640
tasks. 
So form filling basic research 

155
00:08:09,640 --> 00:08:12,880
asks like find the top five 
articles about X, maybe maybe 

156
00:08:12,880 --> 00:08:15,160
some form of data collection 
across different websites or 

157
00:08:15,160 --> 00:08:18,640
simple comparisons. 
And you want to be watching for 

158
00:08:18,640 --> 00:08:21,960
how quickly the agent can 
complete that task versus you 

159
00:08:21,960 --> 00:08:25,440
just doing it yourself. 
So yeah, have a test of it and 

160
00:08:25,440 --> 00:08:28,480
trying to take away learnings 
about a, the future direction of

161
00:08:28,480 --> 00:08:31,640
this space, but also whether you
can imagine users not using this

162
00:08:31,640 --> 00:08:45,410
on a a day-to-day basis 
Anyway, now we're going to I 

163
00:08:45,410 --> 00:08:48,610
guess what ChatGPT is the new 
features and capabilities. 

164
00:08:49,010 --> 00:08:52,130
Let's spend a minute discussing,
I guess that strategic context 

165
00:08:53,010 --> 00:08:56,130
about because I think 
understanding why Atlas exists 

166
00:08:56,130 --> 00:08:58,570
and I guess the constraint it's 
going to have to operate under 

167
00:08:58,810 --> 00:09:02,880
will give us quite a good 
indication about how successful 

168
00:09:02,880 --> 00:09:04,040
it is. 
And you'll probably start to see

169
00:09:04,040 --> 00:09:09,000
why I think what I think around 
open AI not becoming the go to 

170
00:09:09,000 --> 00:09:12,160
browser and dethroning Chrome. 
So from a technical point of 

171
00:09:12,160 --> 00:09:15,520
view, Cronium, so the 
infrastructure provided and 

172
00:09:15,520 --> 00:09:18,880
built by Google to allow other 
companies to build web browsers,

173
00:09:18,880 --> 00:09:24,120
that's Cronium powers about 83% 
of global browsers as of 2025 S 

174
00:09:24,120 --> 00:09:28,360
Atlas runs on Cronium, so does 
Edge, Brave, Opera, Arc. 

175
00:09:28,600 --> 00:09:31,680
The exceptions are literally 
only Safari, which has 16% 

176
00:09:31,680 --> 00:09:35,560
market share and Firefox at 
2.37% market share. 

177
00:09:35,680 --> 00:09:38,000
Well, I think it's quite 
interesting is that Open AI have

178
00:09:38,000 --> 00:09:41,400
hired Ben Guja as Atlas's head 
of engineering. 

179
00:09:41,880 --> 00:09:45,880
And for context of why this is 
important, Guja built Firefox at

180
00:09:45,880 --> 00:09:47,480
Mozilla and then Chrome at 
Google. 

181
00:09:47,880 --> 00:09:50,840
So his third big browser 
project, he chose to build 

182
00:09:50,920 --> 00:09:54,600
Chrome on top of Cronium. 
So Google's open source 

183
00:09:54,600 --> 00:09:57,280
infrastructure for building new 
web browsers rather than 

184
00:09:57,280 --> 00:09:59,920
creating a new engine. 
You know, even Microsoft 

185
00:09:59,920 --> 00:10:01,640
couldn't justify fighting this 
fight. 

186
00:10:01,800 --> 00:10:05,480
They spent many years developing
what's called Edge HTML, which 

187
00:10:05,480 --> 00:10:07,920
is, I guess, their own 
proprietary browser 

188
00:10:07,920 --> 00:10:10,600
infrastructure. 
And you can guess what happened 

189
00:10:10,720 --> 00:10:15,440
in in December 2018, Microsoft 
announced they were giving up on

190
00:10:15,440 --> 00:10:17,800
the Edge HTML project and 
switching to Chronium. 

191
00:10:18,120 --> 00:10:20,240
And by January 2020, that 
transition is done. 

192
00:10:20,680 --> 00:10:24,280
So that's one of the biggest, 
wealthiest technology providers 

193
00:10:24,280 --> 00:10:28,000
in the world with almost 
unlimited resources, really 

194
00:10:28,000 --> 00:10:30,280
looked at that competitive 
landscape and say we're not 

195
00:10:30,280 --> 00:10:32,080
really going to invest here. 
We don't think we can win. 

196
00:10:32,240 --> 00:10:35,440
You know, building a browser 
engine is difficult and very 

197
00:10:35,440 --> 00:10:40,080
expensive and arguably difficult
to have any real IP or 

198
00:10:40,080 --> 00:10:42,440
competitive advantage by owning 
that infrastructure layer. 

199
00:10:42,520 --> 00:10:45,040
You have to build rendering 
pipelines, JavaScript 

200
00:10:45,040 --> 00:10:48,800
interpreters in preventing 
thousands and many thousands of 

201
00:10:48,800 --> 00:10:52,840
web standards, handling security
updates, maintaining, 

202
00:10:53,240 --> 00:10:55,920
maintaining compatibility across
different devices and our 

203
00:10:55,920 --> 00:10:58,640
operating systems. 
You know, the cost is just 

204
00:10:58,640 --> 00:11:02,680
astronomical and the use cases 
that you have to get right is 

205
00:11:02,680 --> 00:11:05,080
also astronomical. 
And Microsoft, obviously the 

206
00:11:05,080 --> 00:11:07,320
Marfan for there's no point 
doing this. 

207
00:11:07,480 --> 00:11:09,720
Now coming back to Ben Gudger 
and why I think this is 

208
00:11:09,720 --> 00:11:12,280
especially interesting on top of
the fact that he chose to use 

209
00:11:12,280 --> 00:11:14,240
Cronium. 
Now there are many different 

210
00:11:14,240 --> 00:11:17,440
ways to interpret this decision.
One perspective might be that 

211
00:11:17,640 --> 00:11:21,000
building a browser from or 
browser infrastructure from 

212
00:11:21,000 --> 00:11:23,680
scratch would require hundreds 
of millions of dollars in years 

213
00:11:23,680 --> 00:11:26,920
development and deliver very 
limited additional value to any 

214
00:11:26,920 --> 00:11:29,240
users. 
But by building on Chronium, 

215
00:11:29,240 --> 00:11:33,320
Open AI can just focus on the AI
integration, which is the actual

216
00:11:33,320 --> 00:11:36,920
innovation here and that's 
arguably smart resource 

217
00:11:36,920 --> 00:11:38,960
allocation. 
Another perspective is that it 

218
00:11:38,960 --> 00:11:42,080
creates a big structural 
dependency on Google's 

219
00:11:42,080 --> 00:11:44,640
infrastructure. 
Well, really just trying to 

220
00:11:44,640 --> 00:11:48,520
compete at an application layer.
Every Cronium update would 

221
00:11:48,520 --> 00:11:51,000
favour Chrome, which obviously 
is a massive strategic 

222
00:11:51,000 --> 00:11:53,760
vulnerability. 
And I think both perspectives 

223
00:11:53,760 --> 00:11:55,560
capture something real and 
important. 

224
00:11:56,280 --> 00:11:59,640
And the real question is which 
dynamic will prove more 

225
00:11:59,640 --> 00:12:03,120
important to success. 
You know, And I think it was 

226
00:12:03,160 --> 00:12:06,280
definitely an argument to say 
that Open AI financial situation

227
00:12:06,280 --> 00:12:10,560
is a driver here because Open AI
are going to lose $44 billion 

228
00:12:10,880 --> 00:12:14,480
between now and 2029 before 
maybe reaching profitability. 

229
00:12:14,640 --> 00:12:17,840
And only about 5% of their 800 
million users actually pay for 

230
00:12:17,840 --> 00:12:20,920
subscriptions. 
And some Altman in a Fortune 

231
00:12:20,920 --> 00:12:23,080
interview actually acknowledge 
that they're losing money on a 

232
00:12:23,080 --> 00:12:26,480
$200 a month Pro subscription 
because the amount of usage 

233
00:12:26,480 --> 00:12:28,720
exceeds their initial 
projections. 

234
00:12:28,840 --> 00:12:32,600
So another interpretation is 
that Open AI is investing very 

235
00:12:32,600 --> 00:12:36,280
heavily in its growth phase, 
similar to how Amazon operated 

236
00:12:36,520 --> 00:12:39,400
at a loss for years and Uber and
many big companies. 

237
00:12:39,800 --> 00:12:42,120
And obviously therefore Atlas 
could be a very clever 

238
00:12:42,120 --> 00:12:45,560
distribution strategy to 800 
million users. 

239
00:12:45,560 --> 00:12:47,680
But you could also argue that 
the amount of financial pressure

240
00:12:47,680 --> 00:12:51,280
on open AI is leading to 
strategies which are a bit 

241
00:12:51,280 --> 00:12:53,640
pointless. 
A good example of this is Sam 

242
00:12:53,640 --> 00:12:57,120
Altman described advertising as 
a complete last resort at a 

243
00:12:57,120 --> 00:12:59,360
Harvard Business School 
interview in May 2024. 

244
00:12:59,520 --> 00:13:04,280
Yet by 2025 of this year, he's 
been telling structures. 

245
00:13:04,280 --> 00:13:07,560
So an advertising media platform
led by someone called Ben 

246
00:13:07,560 --> 00:13:10,400
Thompson saying that he loves 
Instagram ads. 

247
00:13:10,960 --> 00:13:15,440
An ad week reported that Open AI
hired one of Google's ex head of

248
00:13:15,440 --> 00:13:18,800
search advertising as the VP in 
May 2025. 

249
00:13:18,800 --> 00:13:22,680
And apparently Open AI targeting
about $1 billion in advertising 

250
00:13:22,680 --> 00:13:25,680
revenue by 2026. 
Maybe this just shows that 

251
00:13:25,680 --> 00:13:29,160
they're burning so much cash, 
they're maybe pivoting to other 

252
00:13:29,160 --> 00:13:33,440
revenue models to cover costs. 
It wouldn't be that surprising. 

253
00:13:33,440 --> 00:13:36,200
As a consumer platform scales, 
it often adds advertising 

254
00:13:36,200 --> 00:13:40,000
revenues to its business model. 
And time will tell which 

255
00:13:40,000 --> 00:13:42,560
interpretation matters. 
But also there are some points 

256
00:13:42,560 --> 00:13:46,000
on data collection and context 
capturing that maybe highlight 

257
00:13:46,000 --> 00:13:49,600
why they've done this as well. 
You know, the Perplexity CEO has

258
00:13:49,600 --> 00:13:52,640
said very explicitly that the 
reason that AI companies are 

259
00:13:52,640 --> 00:13:55,720
producing browsers is because 
they want more context and data 

260
00:13:55,720 --> 00:13:58,440
on you because they want to know
what hotels you're buying, what 

261
00:13:58,440 --> 00:14:01,640
restaurants are going to, what 
you look at on the Internet 

262
00:14:02,720 --> 00:14:06,440
because it tells them more about
you and they say to personalized

263
00:14:06,440 --> 00:14:08,560
interactions. 
And it coming down to the killer

264
00:14:08,560 --> 00:14:11,240
point of why I don't think they 
can be that successful is just 

265
00:14:11,240 --> 00:14:12,520
looking at that market 
structure. 

266
00:14:12,520 --> 00:14:16,120
The fact that Chrome has almost 
70% of market share and the 

267
00:14:16,120 --> 00:14:18,760
companies that Google are 
competing with are basically in 

268
00:14:18,760 --> 00:14:21,760
Google's pocket. 
You know, Safari has about 16% 

269
00:14:21,760 --> 00:14:25,520
mainly because they own iPhones 
and Firefox has 2.45%. 

270
00:14:25,680 --> 00:14:29,680
And we watch Firefox go from 
about 31% of market share in 

271
00:14:29,680 --> 00:14:33,520
2009 to 2.3% now. 
Google has invested very 

272
00:14:33,520 --> 00:14:37,880
cleverly in bundling Chrome into
all Android devices and 

273
00:14:37,880 --> 00:14:40,720
throughout the Google ecosystem,
they've obviously put huge 

274
00:14:40,720 --> 00:14:43,280
amounts of investment into 
Chrome itself. 

275
00:14:43,320 --> 00:14:46,760
Google also pays 10's and 10s of
billions of dollars to Apple to 

276
00:14:46,760 --> 00:14:48,480
have Chrome on every single 
device. 

277
00:14:48,560 --> 00:14:52,520
And up to 70% of Firefox's 
revenue is coming from Google to

278
00:14:52,520 --> 00:14:55,320
use Google services. 
And even Microsoft, who've got 

279
00:14:55,320 --> 00:14:59,480
lots of Enterprise Products, 
lots of bundling of Edge into 

280
00:14:59,480 --> 00:15:02,040
Windows products, lots of 
marketing, they've only reached 

281
00:15:02,040 --> 00:15:04,720
about 5% after years of trying 
hard. 

282
00:15:04,880 --> 00:15:08,440
Now, this seems very doom and 
gloom because I do think that 

283
00:15:08,440 --> 00:15:12,080
Atlas could have maybe 1 to 4% 
market share, which you could 

284
00:15:12,080 --> 00:15:15,040
say seems very small. 
But that's about 25 to 100 

285
00:15:15,040 --> 00:15:17,440
million users. 
And by many metrics, that's a 

286
00:15:17,440 --> 00:15:19,360
successful product. 
So yeah, maybe it becomes a 

287
00:15:19,360 --> 00:15:22,680
specialized tool for 100 million
people around the world. 

288
00:15:22,880 --> 00:15:25,640
Maybe it doesn't, but they've 
got a massive distribution 

289
00:15:25,640 --> 00:15:28,320
advantage because they've got 
800 million weekly active users 

290
00:15:28,320 --> 00:15:31,360
there and converting just a few 
percent would put them into 10's

291
00:15:31,360 --> 00:15:34,280
and 10s of millions of users, 
which they could monetize and 

292
00:15:34,280 --> 00:15:36,600
generate 10s of millions or 
billions of dollars. 

293
00:15:36,760 --> 00:15:39,040
Now I find it quite depressing. 
I wish we were moving away from 

294
00:15:39,040 --> 00:15:41,760
an advertising model. 
It seems that it's creeping back

295
00:15:41,760 --> 00:15:55,080
as per usual. 
But anyway, let's pull this 

296
00:15:55,080 --> 00:15:57,000
together. 
I think open our eyes Atlas 

297
00:15:57,000 --> 00:16:00,240
browser is unlikely, as I said, 
to achieve this revolutionary 

298
00:16:00,240 --> 00:16:03,080
market transformation. 
I think mainly because I think 

299
00:16:03,080 --> 00:16:06,320
just market dynamics, historical
precedent, many structural 

300
00:16:06,320 --> 00:16:09,040
constraints all point towards 
that outcome. 

301
00:16:09,920 --> 00:16:11,840
But I don't think that means 
it's not worth engaging with 

302
00:16:12,240 --> 00:16:14,120
market share and individual 
utility. 

303
00:16:14,120 --> 00:16:17,360
I think two different things. 
Maybe Atlas serves up to 100 

304
00:16:17,360 --> 00:16:21,000
million users and has and 
becomes a niche product. 

305
00:16:21,000 --> 00:16:23,440
Maybe it doesn't. 
But for you as an individual, 

306
00:16:23,440 --> 00:16:25,680
market outcome shouldn't really 
matter too much. 

307
00:16:25,680 --> 00:16:29,000
What matters is whether Atlas 
solves a pain point that you 

308
00:16:29,000 --> 00:16:31,280
have today more than your 
current setup. 

309
00:16:32,280 --> 00:16:35,680
And the only way to learn and 
find that out is to go test. 

310
00:16:36,120 --> 00:16:38,920
So give it one to two to three 
weeks of testing and see what 

311
00:16:38,920 --> 00:16:41,320
you think. 
Anyway, that's it for today. 

312
00:16:42,120 --> 00:16:44,920
I hope you enjoyed the episode 
and I'll see you next week.