WEBVTT

1
00:00:00.000 --> 00:00:04.880
So you want to use Cloud Code. You want to get the most of it, but you don't know exactly how.

2
00:00:04.880 --> 00:00:10.720
This is a crash course, how to master Cloud Code, and we explain it in the most simple way.

3
00:00:10.720 --> 00:00:15.840
There are thousands, literally thousands of other Cloud Code tutorials on the internet,

4
00:00:15.840 --> 00:00:20.560
but there are none as simple as this. I brought on Professor Ross Mike. He comes on,

5
00:00:20.560 --> 00:00:26.400
and he shares it in the simplest way so that anyone could create jaw dropping startups and

6
00:00:26.400 --> 00:00:31.360
software using Cloud Code. We're going to give you the exact steps for how you can set it up,

7
00:00:31.920 --> 00:00:36.320
thinking about the beginner, how to think about the terminal, how to think about prompting.

8
00:00:36.320 --> 00:00:41.040
But if you stick around to the end of this episode, there's the tips and tricks section,

9
00:00:41.040 --> 00:00:45.520
which I think is super valuable, and I can't wait to see what you build.

10
00:00:46.000 --> 00:00:59.040
We got Ross Mike on the pod by the end of this episode. What are people going to learn?

11
00:00:59.920 --> 00:01:04.400
Hopefully you're going to not feel overwhelmed with Cloud Code. I know the terminal is scary,

12
00:01:04.400 --> 00:01:09.280
and it's a big buggy man, but I'm going to give you the blueprint how to use it. I'm also going

13
00:01:09.280 --> 00:01:14.960
to share, consider this the ultimate crash course on how to use Cloud Code or any agent effectively.

14
00:01:15.920 --> 00:01:22.480
Okay, let's get into it. So I mean, the best way to start these episodes is with sharing our screen.

15
00:01:23.200 --> 00:01:29.920
So when we think of building applications using AI, using some sort of agent like Cloud Code,

16
00:01:29.920 --> 00:01:35.840
or OpenCode, or Codex, whatever it is, there's a couple of things that you always have to keep in mind.

17
00:01:35.840 --> 00:01:41.040
You know, the principles never really change. One thing that it's important for us to understand is,

18
00:01:41.040 --> 00:01:47.440
however good your inputs are, we'll dictate how good your output is. We're getting to a point

19
00:01:47.440 --> 00:01:54.480
where the models are so freakishly good that if you are producing code on code Slop,

20
00:01:54.480 --> 00:01:59.280
it's because you've given it Slop. There was a time where the models weren't good enough. There was

21
00:01:59.280 --> 00:02:05.280
a time where we had serious qualms and issues with the quality of code the models gave us,

22
00:02:05.280 --> 00:02:11.360
but now we're starting to get to a point where even myself, like I'm reviewing a lot more code

23
00:02:11.360 --> 00:02:17.520
than I write and I never thought I'd be able to say that in the early months of 2026. So very

24
00:02:17.520 --> 00:02:22.800
important for us to understand our inputs, how good they are, how precise they are, how articulate

25
00:02:22.800 --> 00:02:28.160
they are, are just as good as our outputs. And we'll dictate just how good our outputs will be.

26
00:02:28.160 --> 00:02:32.800
And the way I want people to think about this is Greg is like imagine you were communicating this

27
00:02:32.800 --> 00:02:39.600
to a human, to a human engineer, right? If you give them sparse instructions and if anyone is in

28
00:02:39.600 --> 00:02:44.560
like client work, you realize that most clients, they tell you one thing, but you have to sort of

29
00:02:44.560 --> 00:02:51.280
extract the deeper thoughts of what it is they want. It's the same way when we work with these

30
00:02:51.280 --> 00:02:57.120
agents, when we work with cloud code, we need to be really, really precise with how we build our

31
00:02:57.120 --> 00:03:04.720
inputs. Now, what do I mean by inputs? What I mean is our PRDs or our to-do list or our plans,

32
00:03:04.720 --> 00:03:09.120
right? Like there's, you know, people are giving you different names. It doesn't really matter.

33
00:03:09.120 --> 00:03:14.320
It's all the same thing, right? And when we think of a PRD, when we think of a to-do list or when

34
00:03:14.320 --> 00:03:20.640
we think of a plan, I want us to think in such a way as this. Let's say I'm trying to build this

35
00:03:20.640 --> 00:03:25.520
product, right? Let's say, I don't know, Greg, any product ideas that we get.

36
00:03:25.520 --> 00:03:30.480
I mean, you have product ideas? Yeah, that's actually the best best best person to ask, right?

37
00:03:33.520 --> 00:03:39.760
Let's say I go on idealbrowser.com. I was just going to it. I was just going to it. Yeah,

38
00:03:39.760 --> 00:03:45.360
pick the idea of the day for my dear browser says it's a diagnostic tool for appliance

39
00:03:45.360 --> 00:03:50.240
techs losing hundreds of repeat visits. See, I have no idea what that means, but let's say I know

40
00:03:50.240 --> 00:03:55.920
what that means. Essentially, when thinking of this idea and looking to build this into a full-fledged

41
00:03:55.920 --> 00:04:04.240
product, generally the way you're going to think is, okay, if product X does Y, Z, A, B, and C,

42
00:04:04.960 --> 00:04:09.200
how I would reach that is I'm going to think of features, right? So let's say there's four core

43
00:04:09.200 --> 00:04:16.320
features to this application that Greg just mentioned. And if I have these four features built

44
00:04:16.320 --> 00:04:24.080
out, we can safely assume that we have said product, right? The way we are to design our PRDs

45
00:04:24.080 --> 00:04:30.160
to do lists and plans is such that we want the agent, the model to build out all these features,

46
00:04:30.160 --> 00:04:36.080
right? Because all these features put together is our product. You see, a lot of times people will

47
00:04:36.080 --> 00:04:42.000
describe a product, not describe features and will be frustrated with AI, like AI is supposed to

48
00:04:42.000 --> 00:04:46.480
magically know what you're thinking about. By the way, Greg, am I making sense so far?

49
00:04:46.480 --> 00:04:53.040
100 percent. I'm with you. Yeah. So we really need to think in features. But here's the cool part.

50
00:04:53.040 --> 00:04:58.640
When developing features, oftentimes the issue with models is like you'll develop a feature,

51
00:04:58.640 --> 00:05:03.040
or like let's say the model develops a feature, we don't know if it works. We don't know if it did

52
00:05:03.040 --> 00:05:08.240
it the right way. That's where with all the cool raft stuff that's happening, we can introduce tests,

53
00:05:08.240 --> 00:05:15.680
right? So let's say the model, the agent builds feature one. Before moving on to feature two,

54
00:05:15.680 --> 00:05:20.160
what I'm going to do is I'm going to get the model to write a test. If that test passes,

55
00:05:20.160 --> 00:05:24.640
then we'll work on the second feature. If that test passes, we work on the third feature, right?

56
00:05:24.640 --> 00:05:31.840
So we're finally entering an era where you can really build something serious with these models.

57
00:05:31.840 --> 00:05:38.000
So instead of telling you about just planning, why don't we do actual planning together? So I'm

58
00:05:38.000 --> 00:05:43.520
going to pop up my terminal. So I know everyone's afraid of the terminal, but in all honesty,

59
00:05:43.520 --> 00:05:48.800
if you don't have to use the terminal, ask AI. It's the simplest thing. And if not, you can even

60
00:05:48.800 --> 00:05:54.480
download the Cloud Code app and go on code section, give it a specific folder you want to work on

61
00:05:54.480 --> 00:05:58.880
and use the app. There's literally no excuse to not use Cloud Code. If you're afraid,

62
00:05:58.880 --> 00:06:03.520
boohoo, just jump into use AI, you have all the tools. That being said, I'm just going to type

63
00:06:03.520 --> 00:06:10.080
in Cloud. And we're going to have Cloud Code open. And usually how people plan is they'll click shift

64
00:06:10.080 --> 00:06:16.320
tab, right? And then you have plan mode on. And you can say, let's say I want to build

65
00:06:16.880 --> 00:06:27.920
take talk, UGC generating app for my marketing agency. I see like these UGC apps everywhere.

66
00:06:28.560 --> 00:06:45.040
Please help me create a plan. Write this in the in PRD.MD file. So this is how most people

67
00:06:45.040 --> 00:06:51.200
have planning set up, right? You'll tell Cloud Code or cursor or whatever agent to do the plan

68
00:06:51.200 --> 00:06:57.040
for you. And you ask it to be in some file. And like it says, it'd be happy to help you plan this

69
00:06:57.040 --> 00:07:02.240
out and it'll ask you some questions, et cetera, et cetera. But I found that there's a better way

70
00:07:02.240 --> 00:07:09.120
to get an even more concise plan. And this way, it actually gets you to think a lot more about

71
00:07:09.120 --> 00:07:15.680
trade offs, concerns, UI, UX decisions. Because most of the time, you're sort of allowing the AI to have

72
00:07:15.680 --> 00:07:20.080
free reign over certain decisions, which I think will lead you with a finished product that you're

73
00:07:20.080 --> 00:07:26.000
not excited about. And that's invoking a special tool. I was going to show you guys the tweet,

74
00:07:26.000 --> 00:07:31.680
but unfortunately, Twitter is down right now. But Cloud Code has a specific tool called Ask User

75
00:07:31.680 --> 00:07:38.000
Question Tool. And essentially what this tool does, it starts to interview about the specifics

76
00:07:38.000 --> 00:07:43.360
of your plan, right? So I'm going to drop this prompt where it says read this plan file,

77
00:07:43.360 --> 00:07:47.520
interview me in detail using the Ask User Question Tool about literally anything,

78
00:07:47.520 --> 00:07:52.400
technical implementation, UI, UX concerns, and trade offs. I spelled implementation wrong,

79
00:07:52.480 --> 00:07:58.320
do not judge me. And what this is going to do is it's going to go past the plan that we have

80
00:07:58.320 --> 00:08:03.040
and start to ask us about minute details. So let's finish off this plan first. I'm just going to

81
00:08:03.040 --> 00:08:12.240
accept this is internal use, tech SAP will use React. I just want core features. We'll submit answers.

82
00:08:12.800 --> 00:08:17.200
And then Cloud Code you'll see might ask us a few more questions, but this will generally be

83
00:08:18.000 --> 00:08:25.360
the plan, right? So it's not just the plan, it's the right plan, right? Like to what you were saying,

84
00:08:25.360 --> 00:08:30.800
like go back, scroll back up here, the features and yeah, the features and tests.

85
00:08:31.520 --> 00:08:36.880
Like the way I think about this, and I don't know if you agree, is like if you ask Cloud Code to

86
00:08:36.880 --> 00:08:41.440
build you a car, it doesn't really know what a car is. It doesn't understand like you need a steering

87
00:08:41.440 --> 00:08:48.960
wheel and a radio and you need wheels. So the hard part is trying to figure out is basically

88
00:08:48.960 --> 00:08:54.960
explaining what those things are in a really succinct and clear way. And that's what this

89
00:08:54.960 --> 00:08:59.520
interview is basically doing. It's explaining each of them. And then we're going to test each of

90
00:08:59.520 --> 00:09:04.640
those features. Exactly. Like think of think of it this like a simple example. Let's say you

91
00:09:04.640 --> 00:09:10.960
ask the agent to build you a specific feature, right? How is it going to present that specific

92
00:09:11.040 --> 00:09:15.840
feature? Did you want it in a dashboard? Did you want it to be a model? Did it have to be a separate

93
00:09:15.840 --> 00:09:21.840
page? Like when you don't specify these minute details, it will make the assumption for you. And

94
00:09:21.840 --> 00:09:26.240
with Ralph loops and all these type of things, like you might have a whole application built out and

95
00:09:26.240 --> 00:09:32.640
it's not exactly to the liking or the expectations you had, right? So let me continue. I'll just make

96
00:09:32.640 --> 00:09:39.760
some selections here just so we can move on and then hit submit. And then I'm going to pause this

97
00:09:39.760 --> 00:09:45.840
planning here and then I'm going to paste this. I'm going to say read this plan file and I'm going

98
00:09:45.840 --> 00:09:54.160
to tag the plan file. It's called prd.md. We have that right here. And I'm going to say interview me

99
00:09:54.160 --> 00:09:58.560
the details about this question or I don't even need to tag it because it has it in its context.

100
00:09:58.560 --> 00:10:04.320
But I just want to show you how annoyingly annoying this is going to get meaning. It's going to

101
00:10:04.320 --> 00:10:13.440
keep asking me questions about said plan or said app idea. So notice how it says round one core

102
00:10:13.440 --> 00:10:17.520
workflow and technical foundation, right? And some of the questions that might even ask you are

103
00:10:17.520 --> 00:10:21.120
things that you might not know about because you're not technical. So what do I do when I don't know

104
00:10:21.120 --> 00:10:25.360
something Greg? I'm going to copy this and I'm going to go to the chatbot of my choice whether it's

105
00:10:25.360 --> 00:10:31.360
clawed, chadgbt, whatever. And I'm going to ask it questions. So if you remember earlier, it asked

106
00:10:31.360 --> 00:10:35.760
me generic questions about the app. Now it's saying what's your ideal workflow for generating

107
00:10:35.760 --> 00:10:42.000
UGC video from start to finish? Like notice how the questions are even more specific now. So it says

108
00:10:42.000 --> 00:10:48.960
linear step by step template based batch processing iterative conversational. So let's say I select

109
00:10:48.960 --> 00:10:54.560
that and then it says how should the app handle agent API costs and usage. So now it's talking about

110
00:10:54.560 --> 00:10:59.600
costs, right? Again, most of the times when you just have a basic plan, this is not included in the

111
00:11:00.160 --> 00:11:05.200
plan, right? Let's say we want to have a hard part budget. What database and hosting approach do

112
00:11:05.200 --> 00:11:09.840
you want to use? Most of you probably watching this have no idea. So I can copy this over, go to

113
00:11:09.840 --> 00:11:14.800
chatgbt and ask what's the best decision? This is my current situation. And then you keep going,

114
00:11:14.800 --> 00:11:20.880
you keep going, and you submit answers. So when you use this ask user question tool, the questions

115
00:11:20.880 --> 00:11:26.080
become more granular. So it asks me about core workflow and technical foundation. Now it's going

116
00:11:26.080 --> 00:11:32.560
to ask me about UI UX and script generation. If you notice the first plan that it came up with

117
00:11:32.560 --> 00:11:38.880
of the default plan for cloud code, it was pretty basic. Now it's asking me, okay, what AI do I want

118
00:11:38.880 --> 00:11:44.160
to use for the script generation? I'll use cloud. What UI style aesthetic are you going for? Minimal

119
00:11:44.160 --> 00:11:50.800
clean, dashboard heavy, creative tool feel, chat first, right? So hopefully Greg, I'm making sense with

120
00:11:50.800 --> 00:11:57.600
like how much more questions I'm being asked when I'm invoking this ask user question tool.

121
00:11:59.200 --> 00:12:03.920
Yeah, make complete sense. You're also, you're also going to use less tokens in the end, right?

122
00:12:04.480 --> 00:12:12.560
Because the thing is the better your plan, the better your input, the better the initial set of

123
00:12:12.560 --> 00:12:18.640
documents that you give the model, the better the outcome. And if the better the outcome,

124
00:12:18.640 --> 00:12:23.040
there's no back and forth, right? Most people will have a route through running. It'll be a basic

125
00:12:23.040 --> 00:12:27.680
plan and it'll do what you told it to do, but you aren't specific. So now you're going back and then

126
00:12:27.680 --> 00:12:32.000
maybe you're running another loop or you're going back and doing all these changes. But if you get

127
00:12:32.000 --> 00:12:38.960
it done right, if you invest the time in the planning stage, I 100% believe you'll save a lot more

128
00:12:38.960 --> 00:12:44.000
money. And this will help you clear up a lot of ideas. So like, for example, this idea that we

129
00:12:44.000 --> 00:12:50.720
just had this TikTok UGC farm, how do we want it set up? Do we want it to be flat with search?

130
00:12:50.720 --> 00:12:56.880
Do we want it to be client campaign? There's a lot of like these minute details that you're not

131
00:12:56.880 --> 00:13:01.040
thinking about and because you're not thinking about it, you're allowing cloud code to make those

132
00:13:01.040 --> 00:13:06.000
assumptions for you, right? Which at the end, after it's burnt through a ton of tokens, now you're

133
00:13:06.000 --> 00:13:12.560
going back to change, right? We can save so much headache if we do the proper planning from the

134
00:13:12.560 --> 00:13:21.120
beginning. And hopefully, people see value in this ask user question tool. Make sure you specify

135
00:13:21.120 --> 00:13:27.840
it in your prompt. And hopefully, Greg, that that made sense. It does. So I would say step number one

136
00:13:27.840 --> 00:13:33.520
for this cloud crash course is I would get good at planning. I would get really, really good at

137
00:13:33.520 --> 00:13:39.120
planning. I would get good at generating these, right? Like, look, it keeps on asking me questions.

138
00:13:39.120 --> 00:13:43.840
If you notice the very first plan that we generated with cloud, it was two sets of questions and

139
00:13:43.840 --> 00:13:48.880
it was ready to build. But with this, it's asking me, do I have basic avatars, custom avatars,

140
00:13:48.880 --> 00:13:54.400
multi-seemed videos? How do I want to handle storage? Do I want to download the videos instantly?

141
00:13:54.400 --> 00:14:00.240
Cloud storage, external storage, like, there's so much to software engineering. And I think in our

142
00:14:00.240 --> 00:14:05.040
last video, you, someone shared this on Twitter. I don't know if it was you or someone else. Like,

143
00:14:05.040 --> 00:14:10.800
software building personal software is easy. But building software others are going to use is very,

144
00:14:10.800 --> 00:14:16.560
very difficult. And if you don't have the audacity or the decency to, to set up a little time,

145
00:14:16.560 --> 00:14:20.880
a little extra time to plan, then I guarantee you whatever your generator is going to be AI

146
00:14:20.880 --> 00:14:26.080
flop. And you might blame the model, but really the problem is you. So invest in your plans,

147
00:14:26.080 --> 00:14:33.680
spend time using planning. Don't use the generic plan mode that cursor or cloud code has. I would

148
00:14:33.680 --> 00:14:39.360
use cloud code and then I would specify the ask user question tool. It's going to continue to

149
00:14:39.360 --> 00:14:44.720
know you with questions like it keeps asking, right? Because until it knows exactly what it is you

150
00:14:44.720 --> 00:14:52.160
want, it won't start building. So I would say that step number one to building with cloud code.

151
00:14:52.160 --> 00:14:58.080
Step number two, and everyone's talking about Ralph and it's exciting. But I wouldn't use it.

152
00:14:58.560 --> 00:15:04.240
I wouldn't use Ralph. And the reason I wouldn't use Ralph, if I was just starting out great,

153
00:15:04.240 --> 00:15:12.320
is because how are you going to like imagine this, like imagine not knowing how to drive,

154
00:15:12.320 --> 00:15:19.280
but then buying a Tesla for like the self-driving stuff, like cool in theory. But maybe it's a great

155
00:15:19.280 --> 00:15:24.160
idea to know how to drive, how to steer, how to hit the corners, how to maybe yell at someone

156
00:15:24.240 --> 00:15:30.640
when they cut you off before you get the full automated version. I say this to say because when

157
00:15:30.640 --> 00:15:37.840
you get good at developing plans and then working with the AI to build each feature and testing

158
00:15:37.840 --> 00:15:45.760
each feature, you start to develop the sense on product building on like you know, even I heard

159
00:15:45.760 --> 00:15:51.760
someone call vibe QA testing, you get the sense by going one on one yourself. And this is why a lot

160
00:15:51.760 --> 00:15:57.040
of people who were fighting with cloud code all these months are really, really good at using it now

161
00:15:57.040 --> 00:16:02.720
because they spent the time building without using these crazy automation. So if you're using

162
00:16:02.720 --> 00:16:08.800
cloud code for the first time or you're just getting into it, good plan. Number one, and number two,

163
00:16:08.800 --> 00:16:14.480
get your reps in by not using Ralph. So develop the features one by one. Now that you have your plan,

164
00:16:14.480 --> 00:16:20.560
you can literally tell cloud code, hey, okay, let's build the first feature, you know, go ahead and do

165
00:16:20.720 --> 00:16:24.800
it. And then once the feature is done, you can test it or ask it, how can I test this? How can I run

166
00:16:24.800 --> 00:16:34.160
this app? I wouldn't jump into using Ralph right away. Build without Ralph. But let's say you've built

167
00:16:34.160 --> 00:16:40.880
these reps now and you're comfortable with cloud code. Now you hear about all these things skills,

168
00:16:40.880 --> 00:16:50.880
mcp prompt.md agent.md. What else is there? Something.md. You hear all these conventions plugins.

169
00:16:51.920 --> 00:16:57.360
You have Ralph, all these things. So what do I need to perfectly build something

170
00:16:58.160 --> 00:17:03.600
using cloud code or any agent? I'll be honest with you, most of these things are all the same.

171
00:17:04.080 --> 00:17:12.160
prompt.md and agent.md are just markdown files plugins are skills with, you know, a little bit extra.

172
00:17:12.720 --> 00:17:18.720
What you need to build successfully using these agents is first of all, you need a good plan,

173
00:17:19.840 --> 00:17:25.440
right, which are documents, which is the prd.md we just generated. And then you need

174
00:17:26.160 --> 00:17:34.720
to document the progress that's being made. For anyone who's familiar with Ralph, you know what I'm

175
00:17:34.720 --> 00:17:40.720
talking about? For those who aren't, what's cool about a Ralph loop is as follows. A Ralph loop is

176
00:17:40.720 --> 00:17:48.160
basically, you have a list of things that need to get done, the, what you might call it, the prd.md

177
00:17:48.160 --> 00:17:54.080
or the plan. You give it to the AI model. The model works on the first task. It finishes,

178
00:17:54.080 --> 00:18:01.360
it then documents it in another file, and then it goes again, and it stops until it's completed

179
00:18:01.360 --> 00:18:07.200
the whole list. Now, this isn't anything special, but the reason why it's now super powerful is

180
00:18:07.200 --> 00:18:13.280
because the models are getting so, so good. But here is the issue. If you have a terrible plan,

181
00:18:13.280 --> 00:18:18.720
if you have a terrible prd, this doesn't matter. You're just donating money to Anthropic,

182
00:18:18.720 --> 00:18:23.440
and I wish you the best of luck if that's what you want to do. But if you want to make sure that

183
00:18:23.440 --> 00:18:30.240
your tokens are not wasted, you're going to invest in a good prd.md file or a good plan file.

184
00:18:31.840 --> 00:18:38.640
Greg, am I making sense before? 100%. You're driving the point home. Yes. So I'll talk a little bit

185
00:18:38.640 --> 00:18:47.600
about Ralph now. So with Mr. Ralph Wiggum, how do we use this? Now, there's a lot of different

186
00:18:47.600 --> 00:18:51.600
iterations like people are coming with their own style. I'm going to share with you my Ralph set

187
00:18:51.760 --> 00:18:58.320
up in a second. Greg, one thing I will say is Cloud Code has a plugin, a Ralph Wiggum plugin. I

188
00:18:58.320 --> 00:19:02.720
wouldn't use that. And the reason I wouldn't use that is even the person who invented the whole

189
00:19:02.720 --> 00:19:09.760
Ralph system is against it. It's not the best use of Ralph, but I just want to share this concept

190
00:19:09.760 --> 00:19:15.440
of how Ralph works. It's essentially going to go through our plan, and it's going to build out

191
00:19:15.520 --> 00:19:22.800
each feature step by step. And it's not going to stop until it's done. This is cool when your plan

192
00:19:22.800 --> 00:19:30.480
rocks. If your plan sucks, then it's terrible. It doesn't matter. Now, in terms of how to set up

193
00:19:31.280 --> 00:19:37.280
Ralph Wiggum, I have my own setup. And I don't want anyone to think I'm shilling my own setup

194
00:19:37.280 --> 00:19:42.720
for any reason. But the reason why I built my own setup is there's a couple of things my Ralph

195
00:19:42.800 --> 00:19:49.360
loop does. The first thing is it makes sure that there's a plan, a PRD.MD file, and there's a

196
00:19:49.360 --> 00:19:57.600
progress.txt file. But it also, every feature it builds, it then writes a test, and it then lints.

197
00:19:57.600 --> 00:20:02.800
And basically what this does is it makes sure that every feature that's built actually works,

198
00:20:02.800 --> 00:20:06.720
right? Because there's no point on working on feature two. A feature one doesn't work.

199
00:20:06.720 --> 00:20:11.360
A feature one doesn't work. If the test fails, guess what the AM model is going to do. It's going to

200
00:20:11.360 --> 00:20:17.840
go back to working on feature one. And once the test passes, we work on feature two. And then once

201
00:20:17.840 --> 00:20:23.600
feature two test passes, we work on features three, right? All this is awesome. But I'm going to go

202
00:20:23.600 --> 00:20:30.960
back to the same point. If your plan sucks, then the Ralph loop won't matter. Now, in order to set up

203
00:20:30.960 --> 00:20:36.560
this loop, you can find the get up here. How to set it up? You honestly, I'm not even going to

204
00:20:37.520 --> 00:20:41.760
explain it. Greg, people can literally copy the link, pass it to clot, and then be like,

205
00:20:41.760 --> 00:20:47.440
I want to run this Ralph loop. And it will tell you exactly what to do. That's how good the models

206
00:20:47.440 --> 00:20:54.720
have become. But I'll show you an example of this running. So I have a simple PRD.MD file. It's

207
00:20:54.720 --> 00:20:59.360
nothing crazy. It's just to show you the point. But basically, there are a couple of tasks here.

208
00:20:59.360 --> 00:21:04.560
I want to build a basic server that has some basic endpoints. And I just want to show you how my

209
00:21:04.640 --> 00:21:10.640
Ralph loop works. So when I run this Ralph loop, and again, if you don't know how to run this,

210
00:21:10.640 --> 00:21:17.440
the you paste the GitHub URL in Cloud code in your agent and ask it and it will tell you how to do it.

211
00:21:17.440 --> 00:21:22.560
I have a few different configurations. I can use open code if I want. I can use codex if I want.

212
00:21:22.560 --> 00:21:27.280
But I'm just going to use Cloud code. And I'm just going to run the script. And basically,

213
00:21:27.280 --> 00:21:34.240
what it's going to start doing is it's going to start running through each task as you can see.

214
00:21:34.240 --> 00:21:38.640
And it's going to update the PRD. And it's just going to continue to work. Now,

215
00:21:38.640 --> 00:21:48.800
I can go and leave. I can go about my day, hang with Greg. And this loop will continue to work.

216
00:21:48.800 --> 00:21:52.880
And I'm going to see that at some point, whether it's five minutes, three minutes, ten minutes,

217
00:21:52.880 --> 00:21:57.440
however long this is, this is going to finish all the tasks. I'm going to have a working product

218
00:21:58.160 --> 00:22:06.000
and all this is cool, but it doesn't matter if I'm going to go back to the original document

219
00:22:06.000 --> 00:22:11.440
if the plan isn't good. Now, skills are great. MCPs are great. All these different

220
00:22:11.440 --> 00:22:21.760
markdown files are great. You would do yourself a serious service if your plan is good. So the key

221
00:22:21.760 --> 00:22:27.280
to successfully building with Cloud code is you have an absolutely great plan. And if you use the

222
00:22:27.280 --> 00:22:32.960
ask user question tool, you will spend so much time on the plan where it starts to get annoying.

223
00:22:32.960 --> 00:22:38.320
It doesn't get fun. But those of us who focus on this will end up having better outputs.

224
00:22:39.600 --> 00:22:45.760
Let's continue. If you notice here, my Ralph loop is continuing to go. And it took care of the first

225
00:22:45.760 --> 00:22:52.320
task. I can see some files already generated. If I go to the progress.txt file, you can see Greg,

226
00:22:52.320 --> 00:22:57.360
it started to make some progress. It's documenting that. And this is just going to continue to work.

227
00:22:57.360 --> 00:23:02.160
This is just going to continue to run. So people have different iterations. I know the AMP code people

228
00:23:02.160 --> 00:23:06.720
have their own iteration and different people have their own iteration. It doesn't really matter.

229
00:23:06.720 --> 00:23:11.440
Right. Someone's Ralph is could be better. Someone's can be worse. Someone's gonna be all of that is

230
00:23:11.440 --> 00:23:18.160
cool. But don't get stuck in the weeds. The main source is how you can are

231
00:23:18.160 --> 00:23:24.000
particularly perfectly in a beautiful presentation. Create the perfect input because if you create

232
00:23:24.000 --> 00:23:30.160
the perfect input, we have reached a point where the models will give you perfect outputs. So that's

233
00:23:30.160 --> 00:23:38.000
my main tip crash course for people. Use the ask user question tool build without using Ralph.

234
00:23:38.000 --> 00:23:42.480
And if you are going to use Ralph, understand if your plan sucks, you're just donating

235
00:23:42.480 --> 00:23:46.960
money to a topic. And I think a topic has enough money that they don't need your money being

236
00:23:46.960 --> 00:23:55.120
donated to them. Amen. Is there anything else people need to know? Like little tips and tricks? I

237
00:23:55.120 --> 00:24:02.880
notice, you know, you're not using the Mac terminal. You're using ghosty. Yes. Yes. So honestly,

238
00:24:02.880 --> 00:24:08.080
it's all preference, right? So like the terminal you use and all this stuff is all preference.

239
00:24:08.080 --> 00:24:16.640
Here's what I would say. Like let's have tips and tricks list tips and tricks. So first I would

240
00:24:16.640 --> 00:24:24.400
say is my goodness spelling today. First, I would say is use the ask what was the specific tool.

241
00:24:24.400 --> 00:24:30.080
I just want to make sure I don't forget ask user questions tool. Slept on. I don't know why no one's

242
00:24:30.080 --> 00:24:36.000
not talking about it. It literally I saw the tweet from the entropic team. 100% I would use that

243
00:24:36.000 --> 00:24:48.640
when planning. Number two, don't over obsess, obsess on MCP skills, etc, etc. I'm not saying don't

244
00:24:48.640 --> 00:24:53.600
get into these. I'm not saying don't read about them. I'm not saying don't use them. But I can

245
00:24:53.600 --> 00:24:59.360
almost guarantee you these things are not the reason why your product isn't working, right?

246
00:24:59.360 --> 00:25:07.040
Most of the time it's your plan sucks, right? That's number two. Number three, I would use Ralph

247
00:25:08.160 --> 00:25:14.480
after I built something without. And the reason being is again, listen, if you are a baller shot

248
00:25:14.480 --> 00:25:18.720
caller and you have all the money to blow and you don't care and you want to donate money to

249
00:25:18.720 --> 00:25:24.720
entropic, go ahead and use Ralph. But if we were to sit here eye to eye and you haven't built

250
00:25:24.720 --> 00:25:29.920
anything, deployed anything, there isn't a URL that I myself or Greg can click on the truth

251
00:25:29.920 --> 00:25:35.600
built. You have no business using Ralph. You literally have no business using Ralph. I would first

252
00:25:35.600 --> 00:25:41.440
get good at prompting and building something using a plan, whether it's whatever ageing one cloud

253
00:25:41.440 --> 00:25:46.880
code, open code, whatever. Once you have something deployed to Versel or like there's a URL and we can

254
00:25:46.880 --> 00:25:56.800
use it, then you can use Ralph. Number four, this is a little in the weeds, but context is more

255
00:25:56.800 --> 00:26:02.240
important than ever. And a lot of times cloud code or even cursor will tell you what percent of

256
00:26:02.240 --> 00:26:10.560
context has been used. I generally wouldn't go over 50%. Meaning like the entropic model Opus 4.5

257
00:26:10.560 --> 00:26:17.280
has a 200,000 token context limit. The moment in my opinion, you've got over 100,000 tokens,

258
00:26:17.280 --> 00:26:21.920
meaning you're using the same session. It starts to sort of deteriorate. That's when you have

259
00:26:21.920 --> 00:26:26.560
people, Greg, who say, oh, like I started off good, but it started going bad. That's because you've

260
00:26:26.560 --> 00:26:31.840
filled it with so much context. And the best way to think about this is like yourself, right? Like

261
00:26:31.840 --> 00:26:38.080
let's say we went to some English class and or some, you know, whatever class and the professor

262
00:26:38.080 --> 00:26:42.240
just kept dumping information, information, information. At some point, we're going to feel

263
00:26:42.240 --> 00:26:46.640
overwhelmed and we're going to actually start forgetting stuff. And I'm not saying that's how

264
00:26:46.640 --> 00:26:52.080
the models work, but that's how the models act, right? So context is very much important. The moment

265
00:26:52.080 --> 00:26:59.840
you see 50% or even 40%, I would start a new session. And last but not least, have audacity.

266
00:27:00.480 --> 00:27:06.240
And what I mean by that is software development is starting to become easy. The software engineering

267
00:27:06.240 --> 00:27:12.160
is very, very hard. And what do I mean by that? To architect software, to make sure things are

268
00:27:12.160 --> 00:27:17.840
usable, to create great UX UI, to have great taste, to make something that people actually use,

269
00:27:17.840 --> 00:27:22.720
requires time. And in order to spend time, it requires audacity. I know the models are good,

270
00:27:22.720 --> 00:27:28.640
and you can clone a six billion dollar software. But if all of us can do it, now what makes software

271
00:27:28.640 --> 00:27:33.680
different, I think thinking about those things and thinking about the art of building products and

272
00:27:33.760 --> 00:27:38.000
building something that's tasteful is very, very important. And I think anyone who uses these five

273
00:27:38.720 --> 00:27:48.080
tips should kick cheeks in 2025, 2026, sorry. I agree on the audacity thing. I think for me,

274
00:27:48.080 --> 00:27:54.880
it's about creating scroll stopping software. There's so many people and there's a lot of tutorials

275
00:27:54.880 --> 00:27:59.760
about this like cloning billion dollar software. I cloned to four billion dollar software,

276
00:27:59.760 --> 00:28:07.440
look at me. But that's not the type of software that's going to work in 2026, right? I saw this,

277
00:28:08.640 --> 00:28:15.360
let me just share real quick. I saw this guy who created a running app based on how you're feeling.

278
00:28:15.360 --> 00:28:21.600
So it's like, how are you feeling stressed? Angry. And it's an AI assisted running app that

279
00:28:21.600 --> 00:28:25.920
interprets your current emotions to generate a personalized route. And I just thought it was

280
00:28:25.920 --> 00:28:30.880
interesting. You know what I mean? Like, I had never seen an app like this. And I think that like,

281
00:28:30.880 --> 00:28:37.440
as you know, you call it audacity, I think this is an audacious app, right? It's scroll stopping.

282
00:28:37.440 --> 00:28:45.200
You haven't seen it before. So I think push, you want to push cloud code to like get you to this,

283
00:28:45.200 --> 00:28:51.360
basically. And this is why like, so pro people not using rough, if they haven't built anything,

284
00:28:51.440 --> 00:28:56.320
fully because like, now we're people are getting to a point where they want the model to think for

285
00:28:56.320 --> 00:29:00.480
them, right? Where like, if you look at the app, you just share the animations and how things

286
00:29:00.480 --> 00:29:05.360
were floating. And like, even the colors used for the different emotions, like that required

287
00:29:05.360 --> 00:29:10.960
thought, right? And that's what stops people now. Like if building the AI chat interfaces easy,

288
00:29:10.960 --> 00:29:15.120
what's going to make your app different? I think a little bit of audacity, a little bit of thought

289
00:29:15.120 --> 00:29:21.040
and care. And a little bit of taste goes a long way nowadays. And more than the models getting

290
00:29:21.040 --> 00:29:24.240
better, because it's going to get easier, it's going to get better, it's going to get faster.

291
00:29:24.800 --> 00:29:29.920
But unfortunately, if you don't change, then it doesn't all matter. Yeah, and don't be afraid to

292
00:29:29.920 --> 00:29:35.280
use pen and paper like this, this person literally just like started sketching out the features.

293
00:29:36.160 --> 00:29:41.920
How did this thing work? Yeah, how should I do it? And I love it. I love it, right? And this is why

294
00:29:42.000 --> 00:29:46.080
the app, if I don't know the metrics, but I'm willing to bet it's doing really, really well,

295
00:29:46.080 --> 00:29:50.720
because all this stuff matters. Like we could clone something like this feature wise,

296
00:29:50.720 --> 00:29:55.760
but I'm willing to bet like the feel, the animations, the colors, we would not be able to get it

297
00:29:55.760 --> 00:30:03.600
exactly like this. 100%. All right, man. Thanks for coming on. You got me fired up. Actually,

298
00:30:03.600 --> 00:30:09.920
I didn't know about that interview tool. So thanks for sharing that with me. Yeah, just a heads up,

299
00:30:09.920 --> 00:30:13.680
it will ask a lot of questions. I shared it with a couple of friends and a couple of people got

300
00:30:13.680 --> 00:30:19.200
annoyed, but it's worth it, right? Especially if you want it to build something end-to-end or you're

301
00:30:19.200 --> 00:30:26.240
building a very like very minute detailed feature, then it's really, really worth it. I wouldn't

302
00:30:26.240 --> 00:30:31.520
use the general plan personally. So just a heads up, but it's really, really worth it. And I would

303
00:30:31.520 --> 00:30:37.200
love to hear people's feedback in the comments. Sounds good. We'll be in the comments. You got to

304
00:30:37.200 --> 00:30:42.640
come back on in a few months or whenever people want you. It's always an absolute privilege to have

305
00:30:42.640 --> 00:30:50.080
you here. I'll include links where you can follow. And you should follow. Monsieur Rass Mike,

306
00:30:50.960 --> 00:30:57.200
as YouTube channels X, I'll include the link to Ralphie, even though if you're a beginner,

307
00:30:57.200 --> 00:31:03.600
don't even click that link. I know there's maybe some degenerates to do, but I highly suggest you

308
00:31:03.600 --> 00:31:10.320
don't. If you haven't even built without it, there's no point. I have some world-power folks. Come

309
00:31:10.320 --> 00:31:14.720
on. Don't click the link, but I'm putting it in there because I want to see who's tempted.

310
00:31:16.880 --> 00:31:21.680
Thanks again for coming on. I'll see you. I'm coming to Toronto in April. We'll see each other

311
00:31:21.680 --> 00:31:25.760
then. And again, as always, it's a pleasure. Thank you so much for bringing me on.

312
00:31:25.760 --> 00:31:27.760
Of course. Later. Have a good one.