1
00:00:00,120 --> 00:00:04,480
Claude Code has gone viral, one 
of the fastest growing apps of 

2
00:00:04,480 --> 00:00:07,720
all time, reaching over a 
billion dollars in run rate 

3
00:00:07,720 --> 00:00:10,000
revenue within just six months 
of its launch. 

4
00:00:10,160 --> 00:00:13,720
It's so popular even non 
developers are now using it for 

5
00:00:13,720 --> 00:00:17,320
things that Anthropic never ever
anticipated. 

6
00:00:17,640 --> 00:00:20,800
I'm definitely one of those. 
I've started using Claude Code 

7
00:00:20,800 --> 00:00:24,280
in Visual Studio, but just a 
programming interface for tasks 

8
00:00:24,400 --> 00:00:27,880
I never expected, including even
creating documents. 

9
00:00:29,120 --> 00:00:32,680
But for most people, Claude Code
still feels very technical. 

10
00:00:33,400 --> 00:00:36,240
Going into a terminal, running 
commands, staring at a dark 

11
00:00:36,240 --> 00:00:38,480
screen, it's a it's a big 
barrier. 

12
00:00:39,360 --> 00:00:44,720
And so onthropics team who built
Claude Code has now built Claude

13
00:00:44,920 --> 00:00:48,000
Co work. 
So it's the same agent wrapped 

14
00:00:48,000 --> 00:00:51,960
in an interface that everybody 
can use, your head of marketing,

15
00:00:51,960 --> 00:00:55,360
your operations team, probably 
even your dad could use it. 

16
00:00:55,600 --> 00:00:58,680
Today I want to breakdown 
exactly what Claude Co work 

17
00:00:58,720 --> 00:01:02,400
actually is, explain to you how 
it works and why I think this 

18
00:01:02,440 --> 00:01:06,400
might be one of the biggest 
product launches in AI for many 

19
00:01:06,520 --> 00:01:10,640
years and actually what it 
signals about where the entire 

20
00:01:10,640 --> 00:01:13,000
industry is heading. 
This is in the loop with Jack 

21
00:01:13,000 --> 00:01:14,840
on. 
I hope you enjoy the show. 

22
00:01:27,280 --> 00:01:30,120
So let's start with what Quad Co
work actually is. 

23
00:01:30,360 --> 00:01:33,880
Quad Code has been named one of 
the worst name products in AI 

24
00:01:33,880 --> 00:01:38,080
because the Co part just 
actually just distracts from 

25
00:01:38,080 --> 00:01:40,840
what the tool actually does and 
represents. 

26
00:01:41,240 --> 00:01:45,040
It was never just about coding. 
It's a, it's a general purpose 

27
00:01:45,040 --> 00:01:48,240
agent that just happens to be 
really good at coding, but it 

28
00:01:48,240 --> 00:01:52,000
can do a lot, lot more Co work 
isn't just this dumbed down 

29
00:01:52,000 --> 00:01:54,720
version of clawed code. 
It's literally the same agent, 

30
00:01:54,960 --> 00:01:56,760
just with a completely new 
interface. 

31
00:01:56,760 --> 00:02:00,400
Mindset itself has taken a huge 
amount of inspiration from 

32
00:02:00,600 --> 00:02:05,080
clawed code, the way it does its
reasoning, the way it structures

33
00:02:05,080 --> 00:02:07,320
its tools and systems that it 
can access. 

34
00:02:07,440 --> 00:02:10,160
So being able to provide this 
type of technology just with a 

35
00:02:10,160 --> 00:02:12,760
new interface for everyone is 
going to be incredibly powerful 

36
00:02:13,840 --> 00:02:17,280
because ultimately you download 
the desktop app, there's going 

37
00:02:17,280 --> 00:02:20,920
to be a few different tabs when 
you access clod, there's chat, 

38
00:02:21,320 --> 00:02:23,160
there's Co work, and then 
there's code. 

39
00:02:23,440 --> 00:02:27,160
Obviously the code is clock code
and Co work is Clod Co work. 

40
00:02:27,320 --> 00:02:30,080
Now on the face of it, it's 
like, OK, cool, what does it 

41
00:02:30,080 --> 00:02:33,640
actually do? 
Well, it can do a lot of things 

42
00:02:33,640 --> 00:02:38,120
that regular clod cannot do. 
And Clod code can do so it can 

43
00:02:38,120 --> 00:02:41,600
interact with files on your 
computer, it can control your 

44
00:02:41,600 --> 00:02:43,160
browser. 
It can connect to external 

45
00:02:43,160 --> 00:02:46,880
systems, as you know clod does, 
but doing all of that in one 

46
00:02:46,880 --> 00:02:48,640
place is where it becomes very 
powerful. 

47
00:02:49,160 --> 00:02:53,000
In particular, I'll talk about 
why interacting with files is so

48
00:02:53,000 --> 00:02:55,080
damn important. 
So for example, instead of just 

49
00:02:55,560 --> 00:02:59,200
chatting back and forth and then
creating a artifact. 

50
00:02:59,200 --> 00:03:01,320
So if anyone's used clod, it's 
an artifact. 

51
00:03:01,520 --> 00:03:05,560
The equivalent on ChatGPT is a 
canvas where you know, text 

52
00:03:05,560 --> 00:03:07,640
appears or whatever you're 
creating appears. 

53
00:03:07,640 --> 00:03:10,200
So it's where the chat just 
moves to the left and creates 

54
00:03:10,200 --> 00:03:13,960
what looks like a kind of 
document on the right hand side 

55
00:03:13,960 --> 00:03:16,000
that you can interact with. 
And a lot of companies have 

56
00:03:16,000 --> 00:03:18,480
spent a lot of time trying to 
define what this is going to be.

57
00:03:18,480 --> 00:03:20,400
And I think it's really 
interesting here, and I'll come 

58
00:03:20,400 --> 00:03:24,120
on to this later, that actually 
perhaps just the winner of the 

59
00:03:24,120 --> 00:03:28,560
new artifact or asset that AI 
creates and then you share with 

60
00:03:28,560 --> 00:03:33,120
others is not a artifact or a 
canvas or a whatever thing that 

61
00:03:33,120 --> 00:03:35,960
people are naming it. 
It could just be a local file 

62
00:03:35,960 --> 00:03:37,520
that's downloaded straight onto 
your desktop. 

63
00:03:37,720 --> 00:03:40,640
So yeah, called Co work and then
open those files on your 

64
00:03:40,640 --> 00:03:44,400
desktop, rename them, create 
spreadsheets, navigate websites.

65
00:03:44,400 --> 00:03:47,800
It could send an e-mail and you 
might go, OK, cool. 

66
00:03:47,800 --> 00:03:49,600
That's what a lot of these tools
seem to do. 

67
00:03:49,600 --> 00:03:52,040
But there are big differences. 
Now I'll come onto exactly what 

68
00:03:52,040 --> 00:03:54,040
those are and why it matters 
later. 

69
00:03:54,440 --> 00:03:57,320
So to help understand why this 
is also important, I'm going to 

70
00:03:57,320 --> 00:04:00,080
first start with how it actually
works, like how you could use 

71
00:04:00,080 --> 00:04:04,000
this, what you'd use it for. 
Because really one of the things

72
00:04:04,000 --> 00:04:07,000
I'd love for people to do is go 
and test this and really use it.

73
00:04:07,000 --> 00:04:09,200
It's not, you know, like a new 
browser that's kind of 

74
00:04:09,200 --> 00:04:12,040
interesting, but probably 
unlikely to make you change. 

75
00:04:12,880 --> 00:04:16,200
This is actually something 
really, really powerful if you 

76
00:04:16,200 --> 00:04:18,440
can systematize it and get it 
across all your company. 

77
00:04:36,530 --> 00:04:39,010
So when you open Co work, you're
starting I guess with a blank 

78
00:04:39,010 --> 00:04:41,200
slate. 
You know, it can't see any 

79
00:04:41,200 --> 00:04:43,760
files, it can't access the 
browser, it can't do anything. 

80
00:04:44,080 --> 00:04:46,200
You have to grant access 
explicitly. 

81
00:04:46,440 --> 00:04:49,920
So for files, you can give it 
access to specific folders, just

82
00:04:49,920 --> 00:04:51,040
the ones that you want to work 
with. 

83
00:04:51,560 --> 00:04:53,960
For the browser, you can 
obviously you can install a 

84
00:04:53,960 --> 00:04:56,640
Chrome extension that lets you 
see and interrupt with any web 

85
00:04:56,640 --> 00:05:00,680
app for external services like 
Slack or you know, whatever 

86
00:05:00,680 --> 00:05:03,880
system that you're using, you 
can just connect it through 

87
00:05:03,880 --> 00:05:05,720
MCPS. 
So it's just like the cloud 

88
00:05:05,720 --> 00:05:07,600
interface today. 
And anytime it wants to do 

89
00:05:07,600 --> 00:05:10,880
something, so change a file, 
interact with something, it will

90
00:05:10,880 --> 00:05:13,000
come and ask you first and you 
can always say no. 

91
00:05:14,320 --> 00:05:15,880
So you can stay in a lot of 
control. 

92
00:05:16,040 --> 00:05:19,240
So I'll give you a, the most 
simple example, which doesn't 

93
00:05:19,240 --> 00:05:21,440
really make it seem like super 
exciting, but then I'm going to 

94
00:05:21,440 --> 00:05:25,000
build on that a little bit. 
So you might have a shed load of

95
00:05:25,000 --> 00:05:27,840
receipts in your folder because 
you know, you've been travelling

96
00:05:27,840 --> 00:05:31,000
with work, you've been doing 
stuff and you've taken 

97
00:05:31,000 --> 00:05:33,200
screenshots, you've got it on 
your PC or your phone or 

98
00:05:33,200 --> 00:05:35,160
whatever it is. 
And essentially it's just a a 

99
00:05:35,160 --> 00:05:37,760
bunch of files with very 
unhelpful names like receipt, 

100
00:05:37,760 --> 00:05:42,400
one dot, PDF, March, GPG, you 
know, whatever it might be. 

101
00:05:42,560 --> 00:05:47,200
You can then give Claude access 
to that folder, giving Co work 

102
00:05:47,200 --> 00:05:50,360
access just to that specific 
location where everything's 

103
00:05:50,360 --> 00:05:53,120
stored. 
And now you can ask, I have a 

104
00:05:53,120 --> 00:05:55,240
receipts folder. 
Can you just rename all those 

105
00:05:55,240 --> 00:05:58,000
files to say match the dates on 
the receipts? 

106
00:05:58,000 --> 00:06:01,040
And it's super simple, but 
actually quite helpful for 

107
00:06:01,040 --> 00:06:04,840
anyone that's up to, to do 
invoicing and, and get paid by 

108
00:06:04,840 --> 00:06:06,840
their company for the things 
that they've already bought at 

109
00:06:06,840 --> 00:06:10,040
work. 
But anyway, so, so the agent can

110
00:06:10,040 --> 00:06:12,920
just open that folder, look at 
each receipt, read all the 

111
00:06:12,920 --> 00:06:15,920
content and information inside 
it to find the correct date, and

112
00:06:15,920 --> 00:06:19,040
it just then renames the files 
on your local device on your 

113
00:06:19,040 --> 00:06:20,600
laptop. 
If one of the dates might be 

114
00:06:20,600 --> 00:06:22,840
missing a receipt, instead of 
guessing or doing something kind

115
00:06:22,840 --> 00:06:26,120
of weird, it stops and asks, you
know, this receipt doesn't have 

116
00:06:26,120 --> 00:06:26,960
a date. 
What should I do? 

117
00:06:27,080 --> 00:06:29,760
And it's because Anthropic, 
they've spent a lot of time on 

118
00:06:29,760 --> 00:06:33,160
trying to get the model to ask 
for clarification or this agent 

119
00:06:33,160 --> 00:06:36,760
to ask for clarification when 
it's uncertain rather than 

120
00:06:36,760 --> 00:06:39,600
making assumptions that are just
often crap. 

121
00:06:40,320 --> 00:06:42,160
And it sounds like a small 
thing, but it makes a massive 

122
00:06:42,160 --> 00:06:45,040
difference in practice. 
Now, you could then go further 

123
00:06:45,080 --> 00:06:47,400
with this receipt example. 
You could then say, can you put 

124
00:06:47,400 --> 00:06:49,880
this into a spreadsheet? 
And it will be able to actually 

125
00:06:49,880 --> 00:06:51,440
do that. 
It creates an Excel file with 

126
00:06:51,440 --> 00:06:54,360
all the receipt data, dates, 
amounts, whatever was on there 

127
00:06:54,360 --> 00:06:56,040
organize into columns just like 
that. 

128
00:06:56,200 --> 00:06:58,480
And then maybe you don't want 
this local file just sitting on 

129
00:06:58,480 --> 00:07:01,640
your computer. 
Maybe you want it in Google 

130
00:07:01,640 --> 00:07:04,040
Sheets so you can share it with 
your team and then everyone can 

131
00:07:04,040 --> 00:07:06,640
access it from anywhere. 
So you say actually go make this

132
00:07:06,640 --> 00:07:10,360
a Google Sheets instead, and the
agent opens the browser and can 

133
00:07:10,360 --> 00:07:13,040
navigate to Google Sheets, 
create a new spreadsheet and 

134
00:07:13,200 --> 00:07:15,480
start populating it with the 
information. 

135
00:07:16,440 --> 00:07:19,840
You can watch it type around, 
control the interface, and now 

136
00:07:19,840 --> 00:07:22,040
it's not quick, it's not 
instant, but you can watch it 

137
00:07:22,040 --> 00:07:23,720
working through every single 
step. 

138
00:07:23,720 --> 00:07:27,480
And at one point it might notice
it didn't do something correct, 

139
00:07:27,480 --> 00:07:29,240
so it has to go back and it 
fixes it. 

140
00:07:29,640 --> 00:07:31,400
So it's checking its own work as
it goes. 

141
00:07:31,480 --> 00:07:34,680
There's been a lot of progress 
made on computer use in the last

142
00:07:34,680 --> 00:07:37,840
kind of 6 to 12 months. 
And then you say now, OK, go and

143
00:07:37,840 --> 00:07:42,520
take this and send it to this 
person, let's say Amy, and the 

144
00:07:42,520 --> 00:07:46,040
agent will be able to open 
Gmail, Click to compose or 

145
00:07:46,040 --> 00:07:49,640
create new e-mail, type in Amy 
in the receipts field and 

146
00:07:49,640 --> 00:07:53,520
actually pull from contacts. 
Now an objection many people say

147
00:07:53,520 --> 00:07:56,040
is that this is just very slow. 
It's cool, but very slow. 

148
00:07:56,200 --> 00:07:58,920
I could do all of it myself. 
And this is where cloud cowork 

149
00:07:58,920 --> 00:08:01,880
gets a massive amount of 
insiration from cloud code. 

150
00:08:02,000 --> 00:08:04,200
Because yes, you could do 
something like that for a single

151
00:08:04,200 --> 00:08:06,680
task faster. 
And you're definitely faster 

152
00:08:06,680 --> 00:08:10,280
than most agents right now for 
doing tasks on the Internet, 

153
00:08:10,280 --> 00:08:11,960
let's say. 
But the real shift is doing 

154
00:08:11,960 --> 00:08:15,160
things in parallel because while
cowork is doing one thing, you 

155
00:08:15,160 --> 00:08:17,000
could be just doing something 
else. 

156
00:08:17,120 --> 00:08:20,360
So this is something that called
code got really, really right. 

157
00:08:20,360 --> 00:08:22,960
You can create many agents that 
are doing many different things 

158
00:08:22,960 --> 00:08:26,160
all at the same time. 
For example, you could have one 

159
00:08:26,160 --> 00:08:29,160
creating a plan, 1 executing 
part of the plan that you're 

160
00:08:29,160 --> 00:08:34,360
happy with, one writing tests to
assess the quality of the either

161
00:08:34,360 --> 00:08:37,840
plan or the things that have 
been written, the code all on 

162
00:08:37,840 --> 00:08:41,240
one project, whilst another 
agent is then planning your next

163
00:08:41,240 --> 00:08:43,799
task. 
So that's what you can do now. 

164
00:08:43,799 --> 00:08:46,280
You can do that on Claude today 
with many, many different tabs, 

165
00:08:47,320 --> 00:08:51,000
but cloud Co work is trying to 
bring the ability the cloud code

166
00:08:51,000 --> 00:08:53,800
got so right into the main 
application. 

167
00:08:53,960 --> 00:08:56,080
And what's really important 
here, and this is what we'll 

168
00:08:56,080 --> 00:08:58,680
come onto and spend a bit of 
time on, is everything it 

169
00:08:58,680 --> 00:09:03,680
creates, say every file it 
creates, it persists because it 

170
00:09:03,680 --> 00:09:07,160
can be saved as a local file on 
your device and you can just 

171
00:09:07,160 --> 00:09:09,880
point it at that local file. 
You don't need to make AGPT or a

172
00:09:09,880 --> 00:09:13,680
cloud project, or you might 
create loads of, let's say, 

173
00:09:13,720 --> 00:09:16,920
artifacts or canvases that you 
might put in a Google Doc and is

174
00:09:16,920 --> 00:09:20,480
then lost in many ways for 
contexts that another future use

175
00:09:20,480 --> 00:09:23,680
of AI could have benefited from.
Instead, you can literally every

176
00:09:23,680 --> 00:09:27,040
file, save it, put it in the 
folder, organize it in a 

177
00:09:27,040 --> 00:09:30,240
specific way, and suddenly the 
agent will then access it 

178
00:09:30,240 --> 00:09:32,480
anytime. 
And this overcomes many, many, 

179
00:09:32,480 --> 00:09:34,240
many big problems, which I'll 
come onto in a minute. 

180
00:09:35,480 --> 00:09:38,600
Now developers have been doing 
this for years, creating and 

181
00:09:38,600 --> 00:09:42,680
storing files on their local 
device, committing it to the 

182
00:09:42,680 --> 00:09:44,560
entire project. 
This is a very normal thing. 

183
00:09:44,560 --> 00:09:48,040
And This is why I think non 
developers are about to go 

184
00:09:48,040 --> 00:09:50,800
through what developers have 
been doing for many years and in

185
00:09:50,800 --> 00:09:52,880
particular with AI over the last
12 months. 

186
00:09:52,880 --> 00:09:55,080
And they're going to do a break 
net speed. 

187
00:09:56,040 --> 00:09:58,760
So we're going to see a huge 
amount of change in how people 

188
00:09:58,760 --> 00:10:01,200
generally work because I really 
think the difference between 

189
00:10:01,200 --> 00:10:04,120
say, chatting with normal cloud 
and working with coworkers, the 

190
00:10:04,120 --> 00:10:07,120
difference between having 
conversations and continually 

191
00:10:07,120 --> 00:10:09,360
building a system. 
And it's a real mental shift 

192
00:10:09,360 --> 00:10:14,040
that everybody must really take 
on board because the ability to 

193
00:10:14,040 --> 00:10:16,840
create files and save them 
straight on your device 

194
00:10:16,840 --> 00:10:18,560
overcomes, oh, many big 
problems. 

195
00:10:18,560 --> 00:10:21,720
So for anyone that's been using 
normal cloud, you might end up 

196
00:10:21,720 --> 00:10:24,720
hitting context window problems.
And you do the same on GP2. 

197
00:10:24,720 --> 00:10:27,320
So you've put too many files in,
you've had too many 

198
00:10:27,320 --> 00:10:29,160
conversations, you've had too 
many iterations. 

199
00:10:29,160 --> 00:10:32,000
It says I've hit the file 
context in it, which is a pain 

200
00:10:32,000 --> 00:10:34,640
because you might be on a really
good track and suddenly you have

201
00:10:34,640 --> 00:10:39,600
to go make a new chart. 
All of that can go away because 

202
00:10:39,600 --> 00:10:43,480
you can suddenly just create a 
chart and all the files you've 

203
00:10:43,480 --> 00:10:45,280
saved and assets. 
Not if I've put it into a 

204
00:10:45,280 --> 00:10:49,760
folder, point cowork at that 
folder quickly and it's done so 

205
00:10:49,760 --> 00:10:52,920
super, super quick. 
Another huge benefit outside of 

206
00:10:52,920 --> 00:10:57,000
context windows here is the way 
you can create a giant system. 

207
00:10:58,040 --> 00:11:00,120
So a good example here with how 
we work. 

208
00:11:00,240 --> 00:11:04,160
So essentially if you create the
right folder structure and give 

209
00:11:04,280 --> 00:11:08,480
Claude access to certain folders
to run instructions, and you 

210
00:11:08,480 --> 00:11:12,400
tell it to always follow this 
folder full of the instructions,

211
00:11:13,360 --> 00:11:15,680
it will always do that. 
So it's like creating a giant 

212
00:11:15,680 --> 00:11:19,960
project, but the agent won't 
just be stuck on one set of 

213
00:11:19,960 --> 00:11:22,080
instructions. 
So we've got a folder which 

214
00:11:22,080 --> 00:11:26,200
essentially is an agent because 
when people go, what is an agent

215
00:11:26,200 --> 00:11:27,840
or why the what's the value of 
an agent? 

216
00:11:28,120 --> 00:11:33,000
It's literally just a series of 
prompts that fire at the perfect

217
00:11:33,000 --> 00:11:36,600
moment to provide an LLM with 
the perfect amount of context to

218
00:11:36,600 --> 00:11:40,040
complete that task. 
So what you're doing by creating

219
00:11:40,040 --> 00:11:42,640
a folder structure is 
essentially pointing an LLM at 

220
00:11:42,640 --> 00:11:45,920
folders full of information. 
You might have a folder called 

221
00:11:46,080 --> 00:11:48,720
planner, you might have a folder
called researcher, you might 

222
00:11:48,720 --> 00:11:51,880
have a planner called 
Presentation Maker, whatever it 

223
00:11:51,880 --> 00:11:54,200
might be. 
And inside that there is a set 

224
00:11:54,200 --> 00:11:57,000
of instructions, which is 
essentially just a document 

225
00:11:57,240 --> 00:12:01,160
saved as an MD file. 
An MD file is just the best 

226
00:12:01,160 --> 00:12:04,400
format for an LLM to read. 
But you know, imagine it's 

227
00:12:04,400 --> 00:12:07,560
equivalent as a Doc X. 
You can download a file as an MD

228
00:12:07,560 --> 00:12:10,480
file. 
So you create a folder on your 

229
00:12:10,480 --> 00:12:12,800
local device with a set of 
instructions, which are 

230
00:12:12,800 --> 00:12:16,080
essentially your agents. 
Because when you say to an LLM 

231
00:12:16,680 --> 00:12:19,480
essentially go read that folder,
it's going to use that context. 

232
00:12:20,080 --> 00:12:22,160
You might then create another 
one called skills. 

233
00:12:22,240 --> 00:12:24,720
Skills are processes that this 
agent can follow. 

234
00:12:25,560 --> 00:12:28,680
So you might have one called a 
product requirement document 

235
00:12:28,680 --> 00:12:31,320
creator. 
You might have another one 

236
00:12:31,320 --> 00:12:37,080
called Python Writer, you might 
have another one called e-mail 

237
00:12:37,080 --> 00:12:39,040
writer. 
So these are skills, which is 

238
00:12:39,040 --> 00:12:41,960
essentially a document which 
lays out in a specific format 

239
00:12:42,680 --> 00:12:45,160
how an LLM should follow a set 
of instructions. 

240
00:12:46,080 --> 00:12:48,840
So as soon as it then 
understands that this person is 

241
00:12:48,840 --> 00:12:52,280
asking for a certain task, it 
will be able to essentially 

242
00:12:52,440 --> 00:12:56,160
access that folder with that 
skill and read it and then go, 

243
00:12:56,160 --> 00:12:58,600
OK, I know I need to use this 
skill in this particular way. 

244
00:12:58,600 --> 00:13:00,160
And it will fold the process 
that you set out. 

245
00:13:00,320 --> 00:13:02,800
And all you can start to do here
is continually build this 

246
00:13:02,800 --> 00:13:04,440
system. 
And every time you create a new 

247
00:13:04,440 --> 00:13:07,280
file, you can commit it to the 
right folder and an agent can 

248
00:13:07,280 --> 00:13:09,760
then read it further and further
over and over again. 

249
00:13:09,760 --> 00:13:11,480
So you're creating a giant 
system here. 

250
00:13:12,280 --> 00:13:15,760
And this is what we've done it 
from a engineering perspective. 

251
00:13:16,160 --> 00:13:17,560
It's what we're doing across the
company. 

252
00:13:17,560 --> 00:13:21,080
It is literally going to 
transform how people work. 

253
00:13:21,160 --> 00:13:24,600
Another major mental shift with 
this here is just this idea of 

254
00:13:24,600 --> 00:13:27,760
having many, many, many 
different claws or AIS all the 

255
00:13:27,760 --> 00:13:30,000
time. 
See, it is like tending to your 

256
00:13:30,000 --> 00:13:34,440
claws or tending to your AI 
because instead of going super 

257
00:13:34,440 --> 00:13:38,000
deep in one thing, you have many
things going on at once. 

258
00:13:38,120 --> 00:13:40,760
You're going to check in on each
one of them because it might be 

259
00:13:40,760 --> 00:13:43,360
going for three minutes, 5 
minutes, 12 minutes, thirty 

260
00:13:43,360 --> 00:13:46,960
minutes on a specific task and 
you check in to make sure it 

261
00:13:46,960 --> 00:13:49,560
doesn't need to be redirected 
for some reason. 

262
00:13:49,640 --> 00:13:51,240
And you're just doing this 
across all your different 

263
00:13:51,240 --> 00:13:53,200
clauses. 
And that's how programmers have 

264
00:13:53,200 --> 00:13:56,200
been coding for the last 12 
months with all these new AI 

265
00:13:56,200 --> 00:13:57,680
tools. 
It's how I've been using 

266
00:13:57,680 --> 00:14:01,040
different systems and now Co 
workers bringing that to 

267
00:14:01,120 --> 00:14:03,080
everybody. 
One thing again, I mentioned it 

268
00:14:03,080 --> 00:14:05,080
earlier about skills. 
I think this is going to make 

269
00:14:05,200 --> 00:14:07,600
things increasingly better for 
people over time. 

270
00:14:08,000 --> 00:14:12,600
Essentially, skills is when the 
agent creates, let's say a 

271
00:14:12,640 --> 00:14:16,200
spreadsheet, it can load up 
something called an Excel skill 

272
00:14:16,200 --> 00:14:18,120
behind the scenes, which is 
essentially a set of 

273
00:14:18,120 --> 00:14:22,520
instructions that teach claw the
best way to create a Excel 

274
00:14:22,520 --> 00:14:24,920
spreadsheet. 
Now Anthropic provides 

275
00:14:24,920 --> 00:14:28,240
prepackaged skills. 
We as mindset, we build skills 

276
00:14:28,240 --> 00:14:30,120
all the time. 
But what you can also do is 

277
00:14:30,120 --> 00:14:33,000
build your own set of skills 
with MD files so you can start 

278
00:14:33,000 --> 00:14:35,440
to build these skills that are 
very particular to your team and

279
00:14:35,440 --> 00:14:37,520
process. 
Now that said, if you're just 

280
00:14:37,520 --> 00:14:39,800
getting started, of course, 
don't over customize, don't go 

281
00:14:39,800 --> 00:14:42,160
too deep. 
Start simple. 

282
00:14:42,480 --> 00:14:47,360
Go online and search for Co work
file structure, for example, and

283
00:14:47,360 --> 00:14:50,560
look online for the best way to 
create and set up the system. 

284
00:14:51,640 --> 00:14:53,760
You know, Anthropic have 
actually run competitions for 

285
00:14:53,760 --> 00:14:56,360
programmers on the best system 
design and that's where we've 

286
00:14:56,360 --> 00:14:59,560
got a lot of inspiration from. 
So yeah, start very simple. 

287
00:14:59,560 --> 00:15:03,480
Install the Chrome extensions, 
install the app so you can 

288
00:15:03,480 --> 00:15:05,920
actually interact, you know, 
with web as and do things and 

289
00:15:06,040 --> 00:15:07,760
justice. 
Start using it and force 

290
00:15:07,760 --> 00:15:10,800
yourself into this new way of 
working because I promise you 

291
00:15:10,800 --> 00:15:12,720
it's going to be infinitely 
better. 

292
00:15:13,080 --> 00:15:15,920
There's a reason Cloud Code has 
been so successful. 

293
00:15:33,080 --> 00:15:36,160
So yeah, I think let's move on 
to a few interesting use cases 

294
00:15:36,160 --> 00:15:38,760
and what I think is practical 
tips here outside of how this 

295
00:15:38,760 --> 00:15:40,520
all works and why it's a bit of 
a mental shift. 

296
00:15:40,680 --> 00:15:42,840
I think some of the really 
interesting use cases here is 

297
00:15:42,840 --> 00:15:48,160
large processing of content. 
So there was someone called 

298
00:15:48,200 --> 00:15:51,240
Lenin Rychisky. 
So he's a big podcaster in the 

299
00:15:51,240 --> 00:15:53,160
space, in the product management
space. 

300
00:15:53,840 --> 00:15:59,040
He basically gave access to 320 
podcast scripts on his laptop, 

301
00:15:59,880 --> 00:16:04,080
which is about 450 to 600 hours 
of content, so a huge amount 

302
00:16:04,080 --> 00:16:08,000
that would take forever in let's
say if you were uploading that 

303
00:16:08,000 --> 00:16:11,840
to ChatGPT or clot because you 
can't drag 320 documents into a 

304
00:16:11,840 --> 00:16:14,720
chat window badly. 
With Co work you can just 

305
00:16:14,720 --> 00:16:17,560
literally point it at a folder 
done as another use case. 

306
00:16:17,560 --> 00:16:21,440
One of the fantastic things 
about accessing your actual 

307
00:16:21,440 --> 00:16:25,440
local files is the ability to 
produce lots of local files and 

308
00:16:25,440 --> 00:16:27,200
then get AI to organise them 
better. 

309
00:16:28,360 --> 00:16:31,240
So you might constantly be 
creating folders and systems. 

310
00:16:31,240 --> 00:16:34,160
And This is why looking online 
for a really good initial system

311
00:16:34,160 --> 00:16:35,440
to structure this around is 
good. 

312
00:16:35,920 --> 00:16:38,600
But as you create more files, 
you download it, you save it, 

313
00:16:38,680 --> 00:16:42,000
add it to new folders, you can 
then just ask Claude to go and 

314
00:16:42,080 --> 00:16:46,120
reorganize it, tidy up and make 
it efficient for whatever system

315
00:16:46,120 --> 00:16:49,000
you're trying to design it. 
So again, using AI to prune the 

316
00:16:49,040 --> 00:16:50,640
AI system that it's going to be 
following. 

317
00:16:50,800 --> 00:16:53,440
Another big use case here, I 
think is actually starting to 

318
00:16:53,440 --> 00:16:56,360
like unlock things like APIs 
without being someone that's 

319
00:16:56,360 --> 00:17:01,160
technical because most vendors, 
most technology platforms offer 

320
00:17:01,160 --> 00:17:03,720
APIs that allow you to do really
interesting stuff, pull 

321
00:17:03,920 --> 00:17:07,880
interesting data straight into 
your conversation. 

322
00:17:08,200 --> 00:17:10,079
It allows you to just do a lot 
of different things. 

323
00:17:10,480 --> 00:17:13,839
And for most people, they cannot
use those unless they're a 

324
00:17:13,839 --> 00:17:16,160
developer. 
So suddenly a big use case, it's

325
00:17:16,160 --> 00:17:19,000
emerging and it started because 
of non-technical people using 

326
00:17:19,000 --> 00:17:21,920
cloud codes or any coding tool 
for that matter. 

327
00:17:22,599 --> 00:17:25,079
But suddenly with Co work, 
because it's the same underlying

328
00:17:25,079 --> 00:17:27,400
agent, it's going to be very 
good at doing that as well. 

329
00:17:27,520 --> 00:17:30,040
Another big use case area is 
browser based tasks. 

330
00:17:30,360 --> 00:17:33,360
So it's really the team has 
really tried to be good at using

331
00:17:33,360 --> 00:17:34,840
the browser. 
I think there's going to be an 

332
00:17:34,840 --> 00:17:37,800
emerging space to try and create
general purpose agents. 

333
00:17:38,040 --> 00:17:40,640
So an agent that will sometimes 
use skills, sometimes use 

334
00:17:40,640 --> 00:17:46,360
browsers to essentially be able 
to go and complete specific 

335
00:17:46,360 --> 00:17:49,000
tasks, e-mail things to people 
and do things for people. 

336
00:17:49,440 --> 00:17:51,560
And again, it can be slow and 
cumbersome. 

337
00:17:51,880 --> 00:17:54,880
But if you don't see it as a 
function of speed all the time, 

338
00:17:54,880 --> 00:17:58,160
but having many different tasks 
all happening at once and you're

339
00:17:58,160 --> 00:18:02,120
just checking in on your clouds 
or your AIS, then I think this 

340
00:18:02,120 --> 00:18:03,920
becomes and is more powerful 
over time. 

341
00:18:04,040 --> 00:18:06,040
And finally, when it comes to 
research, I think it's really 

342
00:18:06,040 --> 00:18:09,760
interesting because you know, if
you do research on different, 

343
00:18:09,760 --> 00:18:13,200
you know, AI systems, you 
download the file, you might use

344
00:18:13,200 --> 00:18:15,200
that somewhere else. 
You have to upload it now. 

345
00:18:15,200 --> 00:18:18,120
You could just get it to 
research many different things 

346
00:18:18,120 --> 00:18:21,400
but smaller tasks and just 
download it all into a project 

347
00:18:21,400 --> 00:18:25,680
called Research on xtopic and 
then start a new chart with 0 

348
00:18:25,680 --> 00:18:28,200
context. 
So the context window is big, 

349
00:18:28,640 --> 00:18:31,320
which means you can do lots of 
work on that research and point 

350
00:18:31,320 --> 00:18:34,800
it straight at the folder. 
So again, I hope you're starting

351
00:18:34,800 --> 00:18:39,160
to see the compounding impact of
doing task saving files, 

352
00:18:39,160 --> 00:18:43,840
organizing the files with AI, 
and then starting new chats and 

353
00:18:43,840 --> 00:18:46,120
doing it again. 
But pointing at that folder that

354
00:18:46,120 --> 00:18:48,200
you've created just then. 
It's really powerful. 

355
00:19:05,240 --> 00:19:07,880
So finally, I guess you can see 
that I'm relatively excited by 

356
00:19:07,880 --> 00:19:10,840
this, but it's going to why I 
think this overall is going to 

357
00:19:10,840 --> 00:19:14,160
be a really big deal. 
So I think #1 it's a platform 

358
00:19:14,160 --> 00:19:18,880
moment, and it really showcases 
the power of, I guess, nobody 

359
00:19:18,880 --> 00:19:20,520
really knowing where things are 
going to go. 

360
00:19:20,520 --> 00:19:23,160
You know, when the App Store 
launched, the initial apps were 

361
00:19:23,160 --> 00:19:25,800
beer drinking simulators. 
You know, where you got the iPod

362
00:19:25,800 --> 00:19:27,080
and you pretend you're drinking 
a beer. 

363
00:19:27,240 --> 00:19:31,600
Nobody predicted Uber or Tiktok,
you know, nobody had an idea 

364
00:19:31,600 --> 00:19:34,040
what that killer use case was. 
And I think we're at the exact 

365
00:19:34,040 --> 00:19:36,840
moment with agents Cloud code 
was created for programmers, a 

366
00:19:36,840 --> 00:19:40,960
very specific thing and suddenly
everybody lived it. 

367
00:19:40,960 --> 00:19:45,080
And so they suddenly had to 
diverse and move out of just 

368
00:19:45,080 --> 00:19:46,880
coding. 
See, I think it's going to spawn

369
00:19:46,880 --> 00:19:49,480
a load of new potential ways of 
doing things. 

370
00:19:49,600 --> 00:19:51,520
And I think the other big area 
of the reason I think it's 

371
00:19:51,520 --> 00:19:54,280
really exciting is the 
exponential impact of this is 

372
00:19:54,280 --> 00:19:56,080
real. 
You know, a year ago people at 

373
00:19:56,080 --> 00:20:00,320
Anthropic were saying by the end
of 2025, most people won't be 

374
00:20:00,320 --> 00:20:02,160
writing code. 
And I think it sounded very 

375
00:20:02,160 --> 00:20:03,880
aggressive. 
I actually commented that it 

376
00:20:03,880 --> 00:20:06,920
sound quite aggressive, but many
engineers are having most of 

377
00:20:06,920 --> 00:20:09,760
their code written by Claude. 
It's certainly the case that 

378
00:20:10,480 --> 00:20:13,440
most of the code that is 
produced at Mindset is now with 

379
00:20:13,440 --> 00:20:15,760
AI, and I think that's the 
trajectory for coding. 

380
00:20:15,760 --> 00:20:18,640
But the same thing is going to 
happen across all forms of work.

381
00:20:19,080 --> 00:20:21,880
If you listen to last week's 
episode on what trends are going

382
00:20:21,880 --> 00:20:25,520
to be there for 2026? 
A couple of them were everybody 

383
00:20:25,520 --> 00:20:28,720
at knowledge workers becoming an
AI manager and knowledge work 

384
00:20:28,720 --> 00:20:32,280
having a transformation akin to 
what programmers has. 

385
00:20:32,960 --> 00:20:36,920
And this is a big reason why. 
See, I think often we think in a

386
00:20:37,160 --> 00:20:40,000
they think progress is going to 
happen in a linear way, but 

387
00:20:40,000 --> 00:20:41,760
sometimes it can happen 
exponentially. 

388
00:20:41,760 --> 00:20:43,960
I think this is going to be one 
of those instances. 

389
00:20:44,120 --> 00:20:46,720
I think another big reason this 
is going to be so impactful is 

390
00:20:46,720 --> 00:20:48,400
that everybody is going to 
follow. 

391
00:20:49,120 --> 00:20:51,080
Open Eye is obviously going to 
copy this. 

392
00:20:51,080 --> 00:20:53,200
Google is going to do something,
Microsoft is going to do 

393
00:20:53,200 --> 00:20:55,720
something, probably built into 
Windows and distributed to 

394
00:20:55,720 --> 00:20:58,600
billions of people. 
Apple are going to incorporate 

395
00:20:58,600 --> 00:21:01,000
it in some way over the next few
months. 

396
00:21:01,520 --> 00:21:04,480
It's going to be in added into 
things like Adobe and other 

397
00:21:04,480 --> 00:21:07,520
forms of systems like that. 
The entire industry is just 

398
00:21:07,520 --> 00:21:08,600
going to converge on this 
pattern. 

399
00:21:08,600 --> 00:21:11,720
And what I think is really 
interesting it actually that 

400
00:21:11,800 --> 00:21:15,000
there's a lot of investment into
things like the idea of a canvas

401
00:21:15,000 --> 00:21:19,440
or an artefact or a new, I 
guess, file type that happens 

402
00:21:19,440 --> 00:21:23,760
within the AI system that you 
can create things with 

403
00:21:23,760 --> 00:21:27,320
collaboratively with AI. 
And what's funny is that I think

404
00:21:27,320 --> 00:21:30,920
maybe what we've just seen is 
that simply downloading a local 

405
00:21:30,920 --> 00:21:33,440
file onto your laptop, the 
oldest thing in the book, I 

406
00:21:33,640 --> 00:21:36,400
think that's been around for the
longest, might just be that 

407
00:21:36,400 --> 00:21:37,400
winner. 
I think this is going to be 

408
00:21:37,400 --> 00:21:39,320
massive. 
I think that's going to reshape 

409
00:21:39,320 --> 00:21:41,640
much of the industry. 
I think this is going to lead to

410
00:21:41,640 --> 00:21:43,720
a big transformation of how 
products are going to be built 

411
00:21:43,720 --> 00:21:46,520
and designed. 
In short, I think the genie is 

412
00:21:46,520 --> 00:21:49,720
out of the bottle. 
So yeah, in conclusion, I think 

413
00:21:50,200 --> 00:21:53,880
Co work is clawed code without 
the terminal, the same agent. 

414
00:21:53,880 --> 00:21:56,920
It's accessible to everyone, it 
works on your files, it controls

415
00:21:56,920 --> 00:21:59,240
your browser, run tasks in 
parallel. 

416
00:21:59,240 --> 00:22:01,920
So you can have many different 
Claws or AIS going at once. 

417
00:22:02,200 --> 00:22:05,000
And I think this shift is going 
to be massive. 

418
00:22:05,040 --> 00:22:07,760
And in particular, I'm really 
excited to see people start to 

419
00:22:07,760 --> 00:22:11,640
think about files that you save 
in folders on your laptop being 

420
00:22:11,640 --> 00:22:13,640
something that you spend a lot 
of time thinking about the 

421
00:22:13,640 --> 00:22:17,440
system and structure of that, 
and whether that is going to 

422
00:22:17,440 --> 00:22:19,360
just persist for those that 
live. 

423
00:22:19,400 --> 00:22:23,200
I guess it might feel for some 
people somewhat technical, even 

424
00:22:23,200 --> 00:22:26,360
though it's really not, because 
it's slightly different to the 

425
00:22:26,360 --> 00:22:29,520
way they work today. 
Or whether this becomes the de 

426
00:22:29,520 --> 00:22:31,000
facto standard across the 
industry. 

427
00:22:31,000 --> 00:22:34,080
Or whether some people try to 
replicate the idea of files 

428
00:22:34,480 --> 00:22:37,760
being saved in that format but 
on an application that feels 

429
00:22:37,760 --> 00:22:41,200
more like AG drive. 
But anyway, I think the industry

430
00:22:41,200 --> 00:22:44,520
is going to converge on this 
pattern and the people who build

431
00:22:44,520 --> 00:22:48,640
intuition and new processes 
around this tool now will just 

432
00:22:48,640 --> 00:22:50,720
have a huge advantage to how 
they work. 

433
00:22:50,840 --> 00:22:54,120
So yeah, download the app, 
connect your folders, start 

434
00:22:54,120 --> 00:22:57,440
simple, but please do start. 
This is in the Libra Jackal and 

435
00:22:57,440 --> 00:22:59,440
I hope you enjoyed this week's 
episode. 

436
00:23:00,440 --> 00:23:03,640
I had a lot of fun going through
it, getting really stuck into 

437
00:23:03,640 --> 00:23:08,160
Cloud Co work and I look forward
to speaking to you next week.