1
00:00:00,000 --> 00:00:01,800
It always will come back to the prompt.

2
00:00:01,840 --> 00:00:05,760
However, intricate or basic the prompt engineering is in the future.

3
00:00:05,780 --> 00:00:10,440
And there will be people who are in charge of making those and maintaining those.

4
00:00:10,480 --> 00:00:14,360
But I think generally that will be a small-ish number of people.

5
00:00:14,360 --> 00:00:17,480
And a lot of people who derive the value will be the people who are interfacing

6
00:00:17,480 --> 00:00:40,360
with them via prompts of service, via forms.

7
00:00:40,360 --> 00:00:43,160
Ladies and gentlemen, boys and girls, children of all ages, dogs, cats, robots

8
00:00:43,160 --> 00:00:46,480
and everybody in between, especially you, those searching for a place for your

9
00:00:46,480 --> 00:00:51,040
prompts to call a home, this is HTTTA, how to talk to AI.

10
00:00:51,280 --> 00:00:53,040
I am your host, Weston Synthline-West.

11
00:00:53,480 --> 00:00:57,960
As always in a realm governed by gigabytes where grand algorithm

12
00:00:57,960 --> 00:01:02,640
valvanize our every move, there emerges a gem, a guiding light in the grander of AI.

13
00:01:03,000 --> 00:01:09,280
Come to the genius of tomorrow, the genuine, the glamorous, this go to go.

14
00:01:09,280 --> 00:01:10,360
Gee, how are you this week?

15
00:01:10,520 --> 00:01:14,640
Hi, I am extremely happy this week.

16
00:01:14,640 --> 00:01:18,600
And I have huge news, which both of us should be sharing.

17
00:01:19,160 --> 00:01:24,960
So last week we talked about our Uplimit course that we closed and

18
00:01:24,960 --> 00:01:27,080
the amazing feedback we received.

19
00:01:27,400 --> 00:01:29,880
However, Uplimit came back to us with what?

20
00:01:29,920 --> 00:01:32,760
13 pages of feedback.

21
00:01:33,080 --> 00:01:34,040
13 pages of feedback.

22
00:01:34,280 --> 00:01:36,600
It was about the most heartwarming stuff I've ever read.

23
00:01:36,600 --> 00:01:40,920
So everyone that was in the course, man, you guys brought it, was so much fun.

24
00:01:40,920 --> 00:01:45,880
We're going to have our next run of AI and chat GPT for everyone January 15th,

25
00:01:46,040 --> 00:01:51,200
kicking off with a soon to be announced follow on course from the team here at

26
00:01:51,200 --> 00:01:54,840
Synthbinds, but again, that was such a professionally rewarding thing.

27
00:01:55,160 --> 00:01:59,000
Thanks to everyone that came out from our listeners, GoTo's viewers,

28
00:01:59,560 --> 00:02:00,840
people who follow us.

29
00:02:01,040 --> 00:02:02,000
It was such a treat.

30
00:02:02,880 --> 00:02:05,320
But speaking of treat, we have one for you today.

31
00:02:05,560 --> 00:02:10,480
So the second one is we have a special guest on our podcast today, Dan.

32
00:02:10,480 --> 00:02:11,680
Hi, Dan Cleary.

33
00:02:11,960 --> 00:02:12,960
Welcome, Dan.

34
00:02:13,120 --> 00:02:14,440
Thank you for joining us.

35
00:02:14,600 --> 00:02:14,840
Yeah.

36
00:02:14,840 --> 00:02:15,560
Thanks for having me.

37
00:02:15,560 --> 00:02:18,760
You know, I get asked a lot, like, why did you want to start your own company?

38
00:02:18,920 --> 00:02:19,520
It's really hard.

39
00:02:19,520 --> 00:02:23,000
It's an uphill battle, but that intro alone, I think the, is there my reason

40
00:02:23,000 --> 00:02:23,760
for the next month?

41
00:02:23,800 --> 00:02:24,720
So yeah, I'm happy to be here.

42
00:02:24,880 --> 00:02:25,240
Cool.

43
00:02:25,640 --> 00:02:29,600
Well, let me give some, our listeners a little bit more context about Dan and

44
00:02:29,600 --> 00:02:34,640
what he has gone on or has accomplished in this world so far from his college

45
00:02:34,640 --> 00:02:37,680
dorm at NYU to the forefront of AI here.

46
00:02:37,680 --> 00:02:41,640
Dan Cleary has always had a knack for building beginning his career by

47
00:02:41,640 --> 00:02:43,280
launching a software development agency.

48
00:02:43,920 --> 00:02:46,880
Dan helped early stage startups with all things software.

49
00:02:47,000 --> 00:02:51,200
Later on, he and his partner Dalton went on to launch their own SaaS bug

50
00:02:51,200 --> 00:02:53,480
reporting tool amidst their agency work.

51
00:02:53,520 --> 00:02:57,760
The tool gained traction, leading to them to a Silicon Valley accelerator pioneer.

52
00:02:58,280 --> 00:03:01,000
They went on to raise funding shortly after their demo day.

53
00:03:01,000 --> 00:03:08,360
And then last November, ChatGPT launched and Dan saw an opportunity that led

54
00:03:08,360 --> 00:03:14,280
Dan and Dalton to launch PromptHub.us, a collaborative platform to test, iterate,

55
00:03:14,480 --> 00:03:16,000
and build on top of prompts.

56
00:03:16,360 --> 00:03:20,600
PromptHub is now used by hundreds of business owners, developers, and people

57
00:03:20,960 --> 00:03:23,200
who are using prompts every day.

58
00:03:23,320 --> 00:03:25,200
So again, Dan, welcome.

59
00:03:25,200 --> 00:03:28,880
We would love for you to take us through PromptHub.us.

60
00:03:29,040 --> 00:03:29,600
Absolutely.

61
00:03:29,600 --> 00:03:30,280
Let's do it.

62
00:03:30,280 --> 00:03:31,360
And I have a question.

63
00:03:31,440 --> 00:03:32,640
How is Silicon Valley?

64
00:03:34,440 --> 00:03:37,400
I would say a lot of what you hear is true.

65
00:03:37,720 --> 00:03:40,640
The conversations that you overhear on the street are about server

66
00:03:40,640 --> 00:03:42,360
architecture and things of that nature.

67
00:03:42,800 --> 00:03:47,200
So it really does feel that way, which is a really fun place to be.

68
00:03:47,200 --> 00:03:50,760
We were out there for a month and it's everybody around you is working

69
00:03:50,760 --> 00:03:53,000
on something that's helpful related.

70
00:03:53,000 --> 00:03:56,120
And so it's a great place to be in the early stages of a company.

71
00:03:56,280 --> 00:03:56,760
Amazing.

72
00:03:56,760 --> 00:03:58,080
Congratulations on that.

73
00:03:58,600 --> 00:03:59,000
Thank you.

74
00:03:59,000 --> 00:04:03,480
Would you like to jump right into the PromptHub or just give us a little

75
00:04:03,480 --> 00:04:08,480
bit of overview of what it is and maybe why those hundreds of

76
00:04:08,480 --> 00:04:09,920
organizations are using it?

77
00:04:09,960 --> 00:04:12,280
Yeah, I'd love to give a quick elevator pitch of it.

78
00:04:12,280 --> 00:04:16,120
And I think Wes did a really great job, but basically what we wanted to build

79
00:04:16,120 --> 00:04:20,240
with PromptHub was a place to work on prompts collaboratively and use them

80
00:04:20,240 --> 00:04:22,320
for whatever reason you may have.

81
00:04:22,440 --> 00:04:27,960
The genesis of it was that we were building an AI feature into our bug

82
00:04:27,960 --> 00:04:32,040
reporting tool, SaaS company, and the process of getting the prompt to work

83
00:04:32,040 --> 00:04:35,800
well, iterating on it, testing it with data, testing it across different

84
00:04:35,800 --> 00:04:39,560
models and stuff, it was all really challenging and we're left to a lot of

85
00:04:39,560 --> 00:04:40,760
different manual solutions.

86
00:04:40,760 --> 00:04:42,800
And so we thought, Hey, there's got to be a better way.

87
00:04:43,000 --> 00:04:46,680
But since the industry, so Mason, we knew there were going to be a lot of good

88
00:04:46,680 --> 00:04:50,520
tools out there and we decided to spend something up and we're like, Hey, this

89
00:04:50,520 --> 00:04:52,920
is pretty useful for our use case.

90
00:04:52,920 --> 00:04:56,120
So then we went to a couple other founders that we knew, ran it past them

91
00:04:56,120 --> 00:04:58,680
and got some pretty good feedback.

92
00:04:58,680 --> 00:05:01,160
And so we decided to kind of go all in with it.

93
00:05:01,320 --> 00:05:04,680
There's a couple of different use cases because prompts are so ubiquitous,

94
00:05:04,840 --> 00:05:08,120
you could be building them into your full scale product.

95
00:05:08,440 --> 00:05:09,960
You could be just using them every day.

96
00:05:09,960 --> 00:05:14,120
You could be using them to make your company more efficient and work at the

97
00:05:14,120 --> 00:05:15,000
center of all that.

98
00:05:15,080 --> 00:05:18,120
We want to help it make get from A to B faster with prompts.

99
00:05:18,120 --> 00:05:20,680
And then whatever you want to do with them, you always have a place where

100
00:05:20,680 --> 00:05:22,920
you can come back and work on them or deploy them.

101
00:05:23,160 --> 00:05:25,160
And that's how we see ourselves in the ecosystem.

102
00:05:25,160 --> 00:05:29,880
I think the first time I cried it and immediately reminded me of GitHub in the

103
00:05:29,880 --> 00:05:34,200
sense that it functions like a collaborative tool where you still have the

104
00:05:34,280 --> 00:05:37,560
initial prompt that is pushed or posted.

105
00:05:37,720 --> 00:05:42,600
And then you can iterate every little bit of changes or certain changes or

106
00:05:42,760 --> 00:05:44,680
rewind the tape a little bit.

107
00:05:44,760 --> 00:05:49,000
I found that as a very, very inventive way, definitely better than a bunch of

108
00:05:49,080 --> 00:05:52,120
TXT files or a spreadsheet.

109
00:05:52,120 --> 00:05:56,920
I just wanted to say that because I made a video and I don't know if you've seen

110
00:05:57,240 --> 00:06:03,400
that I implemented API of GPD 3.5 into Google Sheets.

111
00:06:03,640 --> 00:06:08,440
And the primary use for me was to test prompts, you know, just like drag and drop,

112
00:06:08,440 --> 00:06:09,720
play with variables.

113
00:06:09,720 --> 00:06:14,280
And now I have prompt hub in front of me and I'm like, yeah, this is way

114
00:06:14,280 --> 00:06:17,000
smoother, way smarter solution.

115
00:06:17,080 --> 00:06:17,240
Yeah.

116
00:06:17,240 --> 00:06:18,520
And I wish we caught you then.

117
00:06:18,520 --> 00:06:22,520
But yeah, that's definitely the main use case and our initial kind of saying out

118
00:06:22,520 --> 00:06:24,760
of the gate was, yeah, we want it to be like GitHub for prompts.

119
00:06:24,760 --> 00:06:28,440
And so the versioning and collaboration was really at the core of what we were

120
00:06:28,440 --> 00:06:32,920
going to do because yeah, we found testing was particularly difficult.

121
00:06:32,920 --> 00:06:36,680
Even just trying to compare outputs side by side is kind of impossible in any of

122
00:06:36,680 --> 00:06:37,800
the current playgrounds.

123
00:06:37,800 --> 00:06:41,080
And so we wanted to make that that easy so you could really kind of tweak

124
00:06:41,080 --> 00:06:43,960
whatever you want, be really meticulous about that.

125
00:06:43,960 --> 00:06:45,880
And that was a big thing for us out of the gate.

126
00:06:45,880 --> 00:06:46,680
Fantastic.

127
00:06:46,680 --> 00:06:51,800
I found that I've tried probably half a dozen or dozen prompt IDEs and for our

128
00:06:51,800 --> 00:06:56,840
listeners and viewers and IDE is a software term, meaning integrated

129
00:06:56,840 --> 00:06:58,040
development environment.

130
00:06:58,040 --> 00:07:03,000
So there's plenty of these, if you search for prompt IDE, there's dozens of

131
00:07:03,000 --> 00:07:07,320
little pop-ups, but they inherently are very technical.

132
00:07:07,320 --> 00:07:10,680
Or at least they put a very technical, they're looking at, they're presenting

133
00:07:10,680 --> 00:07:12,920
themselves through this very technical lens.

134
00:07:12,920 --> 00:07:16,280
What I thought immediately was appealing about prompt hub is even if you have no

135
00:07:16,280 --> 00:07:20,120
idea what GitHub is, but you're creating prompts, you're collaborating on prompts,

136
00:07:20,120 --> 00:07:21,640
it was very, very approachable.

137
00:07:21,640 --> 00:07:23,640
So that was immediately appreciated too.

138
00:07:23,640 --> 00:07:24,120
Yeah.

139
00:07:24,120 --> 00:07:25,560
And I think that's a really good point.

140
00:07:25,560 --> 00:07:29,960
And it's actually something we wanted to focus on too, because we think with

141
00:07:29,960 --> 00:07:33,640
Chatchi BD, with these AI models coming out, the amount of people who can build

142
00:07:33,640 --> 00:07:35,960
something is going to be continued to increase.

143
00:07:35,960 --> 00:07:38,200
You don't need to be a developer anymore.

144
00:07:38,200 --> 00:07:42,440
And so we want to focus on that, that segment that is going to be a good

145
00:07:42,440 --> 00:07:44,040
segment, a growing segment.

146
00:07:44,040 --> 00:07:46,600
And so a user-friendly product was really big out of the gate.

147
00:07:46,600 --> 00:07:50,440
How about we go to prompt hub.us?

148
00:07:50,440 --> 00:07:50,680
Yes.

149
00:07:50,680 --> 00:07:52,280
I just want to make it clear.

150
00:07:52,280 --> 00:07:54,680
And that's the direction you should go.

151
00:07:54,680 --> 00:07:59,880
And just, if you could just walk us through the whole interface and options

152
00:07:59,880 --> 00:08:00,760
and features.

153
00:08:00,760 --> 00:08:01,080
Cool.

154
00:08:01,080 --> 00:08:05,720
And for listeners, under the episode, you will find the link where you can

155
00:08:05,720 --> 00:08:08,360
actually enjoy this episode in a video format.

156
00:08:08,360 --> 00:08:13,560
Dan, if I could ask you, if you can just use as much descriptive language as

157
00:08:13,560 --> 00:08:16,600
possible for the listeners of the podcast.

158
00:08:16,600 --> 00:08:17,720
Yeah, absolutely.

159
00:08:17,720 --> 00:08:17,960
Yeah.

160
00:08:17,960 --> 00:08:21,560
Like I've kind of mentioned, prompt hub is a home for your prompts.

161
00:08:21,560 --> 00:08:25,960
And so you can see all of my prompts kind of laid out in different tiles here.

162
00:08:25,960 --> 00:08:30,600
And this is from all of my team and they range from content creation to different

163
00:08:30,600 --> 00:08:35,160
types of changing the shape of different formats, some stuff we built for clients

164
00:08:35,160 --> 00:08:35,880
back in the day.

165
00:08:35,880 --> 00:08:41,800
So everything that we use in our business in some way or another is listed in here.

166
00:08:41,800 --> 00:08:44,840
And so I'm going to click on one of those.

167
00:08:44,840 --> 00:08:45,480
Let's see.

168
00:08:45,480 --> 00:08:48,120
We'll go to our LinkedIn one is a good one.

169
00:08:48,840 --> 00:08:52,840
So we write and post a lot of content on our blog.

170
00:08:52,840 --> 00:08:57,160
And so whenever we do that, we'll end up posting a LinkedIn post as well.

171
00:08:57,160 --> 00:09:00,440
And so I created a little prompt basically to help with that process.

172
00:09:00,440 --> 00:09:04,040
But so when you click on a prompt and you come into the testing page, you all

173
00:09:04,040 --> 00:09:08,520
have access to the system message and the prop, which is gives you an extra boost

174
00:09:08,520 --> 00:09:11,560
in terms of getting the model to do what you want, I would say a little bit more.

175
00:09:11,560 --> 00:09:15,240
And then you have access to all the parameters as well that come along with

176
00:09:15,240 --> 00:09:17,400
the different Oakland AI models that are available.

177
00:09:17,400 --> 00:09:21,960
So for listeners system message in this case, as a reference would be, you know,

178
00:09:21,960 --> 00:09:25,640
what more people know now is custom instructions on charge BT.

179
00:09:25,640 --> 00:09:26,600
Exactly.

180
00:09:26,600 --> 00:09:27,160
Yes.

181
00:09:27,160 --> 00:09:30,680
You know, so I can come in here and I've written a prompt to help with this

182
00:09:30,680 --> 00:09:34,600
LinkedIn creation LinkedIn post creation, and you'll see if you're watching this,

183
00:09:34,600 --> 00:09:37,080
I have a couple of blue pieces of text in here.

184
00:09:37,080 --> 00:09:41,880
These are just variables and these allow me to plug in different things into

185
00:09:41,880 --> 00:09:43,080
the prompt and easily use it.

186
00:09:43,080 --> 00:09:47,000
So I can use the same prompt as the content of the article changes that we

187
00:09:47,000 --> 00:09:48,440
write every single week.

188
00:09:48,440 --> 00:09:51,320
I just fill in the variables here.

189
00:09:51,880 --> 00:09:54,840
I replace the content with whatever article that we've written and then I

190
00:09:54,840 --> 00:09:56,120
can run the prompt.

191
00:09:56,120 --> 00:09:59,320
And so it helps to keep it a little bit more clean rather than seeing really

192
00:09:59,320 --> 00:10:01,400
big pieces of text kind of floating around.

193
00:10:01,400 --> 00:10:05,640
And so what Wes, what I was kind of mentioning before and piggyback off

194
00:10:05,640 --> 00:10:08,680
what Gota said, we, what we really wanted to make was the process of iterating

195
00:10:08,680 --> 00:10:09,960
testing prompt slot easier.

196
00:10:09,960 --> 00:10:14,360
And so with every test you run, you're comparing your last output in red to

197
00:10:14,360 --> 00:10:15,800
your new output in green.

198
00:10:15,800 --> 00:10:20,520
And so a classic flow is to come in, write a prompt, tweak it a little bit,

199
00:10:20,520 --> 00:10:22,760
and then commit the changes.

200
00:10:22,760 --> 00:10:25,160
You know, maybe you updated the temperature, maybe you added a different

201
00:10:25,160 --> 00:10:28,600
example, maybe you tried out a different method that you've been using.

202
00:10:28,600 --> 00:10:32,040
Maybe you tried out a different method that you picked up from the Uplimit

203
00:10:32,040 --> 00:10:33,560
course, whatever it might be.

204
00:10:33,560 --> 00:10:34,840
Then you can just make a note.

205
00:10:34,840 --> 00:10:39,800
This is the GitHub inspired aspect of it.

206
00:10:39,800 --> 00:10:44,600
You get a nice kind of version history that you can always fall back to.

207
00:10:44,600 --> 00:10:49,000
So again, for our listeners, you have this, the system message block, which

208
00:10:49,000 --> 00:10:52,120
again is the person that the text behind the curtain that's pulling the

209
00:10:52,120 --> 00:10:53,560
levers, you have your input.

210
00:10:53,560 --> 00:10:59,160
You have elegant ways to make the prompt reusable with variables pop up screen.

211
00:10:59,160 --> 00:11:01,640
And then on the right hand side, you get all the little levers.

212
00:11:01,640 --> 00:11:04,200
That's something you don't get when you're using ChatGPT.

213
00:11:04,200 --> 00:11:06,520
You only use the Playground or some of these other tools.

214
00:11:06,520 --> 00:11:10,680
You can adjust your tokens, your temperature, top B, presence,

215
00:11:10,680 --> 00:11:13,560
frequency, penalty, and change your model.

216
00:11:13,560 --> 00:11:17,080
You want to tell us about some of the models that you have integrated right now?

217
00:11:17,080 --> 00:11:20,600
Yeah, so right now we offer access to all the OpenAI models.

218
00:11:20,600 --> 00:11:24,280
In the near future, hopefully within the next 10 days or so, we'll be integrating

219
00:11:24,280 --> 00:11:25,160
Anthropics models.

220
00:11:25,160 --> 00:11:28,520
So you can do these really cool side-by-side tests between the providers.

221
00:11:29,080 --> 00:11:35,000
Because even as great as OpenAI is, sometimes a different model is going to be better.

222
00:11:35,000 --> 00:11:38,840
Just for context window reasons or for performance reasons for specific tasks.

223
00:11:39,560 --> 00:11:40,680
It's good to explore that.

224
00:11:40,680 --> 00:11:42,440
So we'll be offering that for our users soon.

225
00:11:42,440 --> 00:11:48,920
And even seeing what model does the best for the job at hand, comparing it one to the other.

226
00:11:48,920 --> 00:11:49,880
That's great.

227
00:11:49,880 --> 00:11:51,400
Yeah. Yeah, exactly.

228
00:11:51,400 --> 00:11:54,680
GPT-4 is considerably more expensive and it's great.

229
00:11:54,680 --> 00:11:58,280
But if you're doing a basic classification test, you might not need it.

230
00:11:58,280 --> 00:12:04,520
Yeah, I've been testing Open Interpreter and running on GPT-4, two days, 20 bucks.

231
00:12:04,520 --> 00:12:08,760
And I was like, maybe for organizing my files, I don't need four.

232
00:12:08,760 --> 00:12:12,120
I just wanted to go and test with the best of the best.

233
00:12:12,840 --> 00:12:18,360
Then we also, what we are seeing now, this window, what we are looking is under the test.

234
00:12:18,360 --> 00:12:21,720
But there is reports, feedback and forms.

235
00:12:21,720 --> 00:12:23,800
Could you walk us through that?

236
00:12:23,800 --> 00:12:25,960
Yeah, I'll jump over to the form.

237
00:12:25,960 --> 00:12:29,240
It's our newest, I would say most powerful feature yet.

238
00:12:29,240 --> 00:12:34,040
And so what this allows you to do is turn your prompt into a form.

239
00:12:34,040 --> 00:12:35,160
So what does that mean?

240
00:12:35,160 --> 00:12:40,200
It makes it accessible to anybody on your team to run that prompt just from a link.

241
00:12:40,760 --> 00:12:43,880
And so I clicked on the form tab now and I've come over.

242
00:12:44,440 --> 00:12:45,800
I can add some instructions.

243
00:12:45,800 --> 00:12:47,880
I can do some customizations here.

244
00:12:47,880 --> 00:12:52,760
And basically I'm previewing what is a form for this LinkedIn post generator.

245
00:12:52,760 --> 00:12:58,520
And so now my team can come in and they can just input the content for whatever post they

246
00:12:58,520 --> 00:13:00,040
want to be created.

247
00:13:00,040 --> 00:13:04,280
And then they'll be able to generate a LinkedIn post using the prompt that I slave nights

248
00:13:04,280 --> 00:13:05,240
and weekends over.

249
00:13:05,240 --> 00:13:10,440
And so it's a really easy way to distribute the value of whatever type of prompt you're

250
00:13:10,440 --> 00:13:15,080
writing to people on your team, anybody that you want to make available to it.

251
00:13:15,080 --> 00:13:19,240
Lead magnets, there's the use cases for forms are really kind of up to your imagination.

252
00:13:19,240 --> 00:13:23,240
So forms essentially functions as prompts as a service.

253
00:13:23,240 --> 00:13:25,640
Would you kind of classify it as that?

254
00:13:25,640 --> 00:13:26,680
Yeah, definitely.

255
00:13:26,680 --> 00:13:29,160
It's the easiest way to deploy a prompt.

256
00:13:29,160 --> 00:13:30,680
And yeah, prompts is a service.

257
00:13:30,680 --> 00:13:34,760
I think it's something we already see and we're going to continue to see more and more

258
00:13:34,760 --> 00:13:38,440
I'd be curious to hear, I mean, obviously just by virtue of you having this feature

259
00:13:38,440 --> 00:13:43,480
inside of Prompt Hub, it must be something that you guys are thinking about and talking

260
00:13:43,480 --> 00:13:47,480
about on the regular as the way the market is kind of heading.

261
00:13:47,480 --> 00:13:51,560
Our most viral prompt, Professor Synapse, that's been viewed over a hundred thousand

262
00:13:51,560 --> 00:13:55,000
times and used and shared, ostensibly does the same thing.

263
00:13:55,000 --> 00:14:00,760
It takes the need for you to be a prompt engineer out of the equation and just ask you questions

264
00:14:00,760 --> 00:14:02,440
about what you're trying to do.

265
00:14:02,440 --> 00:14:05,720
Do you think the future is prompt less, so to speak?

266
00:14:05,720 --> 00:14:09,960
And when this does hit the masses, they're going to be interfacing with prompts as a

267
00:14:09,960 --> 00:14:14,760
service because someone like us set up the site to generate their resume cover letter

268
00:14:14,760 --> 00:14:15,880
or their LinkedIn post?

269
00:14:15,880 --> 00:14:17,160
Yeah, I think so.

270
00:14:17,160 --> 00:14:20,680
I think it always will come back to the prompt.

271
00:14:20,680 --> 00:14:26,440
However intricate or basic the prompt engineering is in the future, but it will always come

272
00:14:26,440 --> 00:14:26,920
back to that.

273
00:14:26,920 --> 00:14:31,880
And there will be people who are in charge of making those and maintaining those and

274
00:14:31,880 --> 00:14:35,320
making sure they're great, you know, like following logs, make sure there's no malicious

275
00:14:35,320 --> 00:14:36,200
behavior.

276
00:14:36,200 --> 00:14:40,360
But I think generally that will be a small-ish number of people.

277
00:14:40,360 --> 00:14:44,120
And a lot of people who derive the value will be the people who are interfacing with them

278
00:14:44,120 --> 00:14:48,920
via prompts as a service, via forms, via things of that nature.

279
00:14:48,920 --> 00:14:52,520
I resonate with sentiments so much.

280
00:14:52,520 --> 00:14:56,840
I had a business call and basically one of the partners was like, oh, I watched your

281
00:14:56,840 --> 00:14:58,440
video, Professor Synapse.

282
00:14:58,440 --> 00:14:59,240
So that's it.

283
00:14:59,240 --> 00:15:00,680
Prompt engineering is dead.

284
00:15:00,680 --> 00:15:06,120
And I was like, yeah, but I'm not sure if I'm going to be able to do that.

285
00:15:06,120 --> 00:15:09,880
But someone had to create, you know, Professor Synapse.

286
00:15:09,880 --> 00:15:14,680
So this person, that person's value just grew incredibly.

287
00:15:14,680 --> 00:15:19,240
And then, Wes, you also shared this research that just came out about productivity and

288
00:15:19,240 --> 00:15:24,040
efficiency and they measured people who actually know how to use AI.

289
00:15:24,040 --> 00:15:29,720
And basically results probably did not surprise anyone, but that becomes your kind of super

290
00:15:29,720 --> 00:15:31,080
employees.

291
00:15:31,080 --> 00:15:35,480
So if you have a team member who can actually develop prompts, I'm thinking, let's say

292
00:15:35,480 --> 00:15:40,440
marketing, develop prompts, distribute it to the team and then reiterate based on

293
00:15:40,440 --> 00:15:41,720
campaign.

294
00:15:41,720 --> 00:15:46,200
This was some research done by the Boston Consulting Group and they published a paper

295
00:15:46,200 --> 00:15:47,720
about it, but no surprise.

296
00:15:47,720 --> 00:15:53,400
They measured productivity and efficiency and completion across 18 different tasks for

297
00:15:53,400 --> 00:15:54,760
hundreds of different people.

298
00:15:54,760 --> 00:16:00,280
And, you know, I think it's almost finished 12% more completed tasks, 25% faster,

299
00:16:00,280 --> 00:16:05,400
produce 40% higher results when they're able to leverage AI tools.

300
00:16:05,400 --> 00:16:10,120
Thesis maintains there's nothing more powerful than an expert human equipped with some AI

301
00:16:10,120 --> 00:16:11,160
who knows how to use it.

302
00:16:11,160 --> 00:16:12,920
Yeah, I totally agree.

303
00:16:12,920 --> 00:16:15,160
You still need to know how to use the underlying tools.

304
00:16:15,160 --> 00:16:19,240
In the same way that software development has gotten easier in us, we moved away from

305
00:16:19,240 --> 00:16:21,240
machine code all the way to what we have today.

306
00:16:21,240 --> 00:16:22,520
You still need to have an understanding.

307
00:16:22,520 --> 00:16:25,160
You still need to be able to know the tooling will work.

308
00:16:25,160 --> 00:16:27,960
And so, yeah, those numbers definitely don't surprise me.

309
00:16:27,960 --> 00:16:31,960
It's something we hear a lot talking to people, talking to business owners especially.

310
00:16:31,960 --> 00:16:35,560
Everyone's trying to figure out what to do with AI.

311
00:16:35,560 --> 00:16:39,720
And there's always a few people in the company that are chat-chickity power users and know

312
00:16:39,720 --> 00:16:40,760
how to use this stuff.

313
00:16:40,760 --> 00:16:44,840
But a lot are still kind of lacking or still learning just because it's so new.

314
00:16:45,400 --> 00:16:50,120
And so, the ability to distribute that kind of knowledge and value across the company,

315
00:16:50,120 --> 00:16:52,760
I think, is going to be really big, especially in the near future.

316
00:16:52,760 --> 00:16:59,480
So the 1% who's going to be creating prompts and everyone else who is going to need to

317
00:16:59,480 --> 00:17:01,080
know how to use it properly.

318
00:17:01,080 --> 00:17:01,640
Exactly.

319
00:17:01,640 --> 00:17:06,120
In the same way that I could build a dashboard to look at all of our back-end data and share

320
00:17:06,120 --> 00:17:08,840
that with my marketing team so that they could have up-to-date data.

321
00:17:08,840 --> 00:17:11,880
They don't need to know how to build this stuff, but they benefit from having those

322
00:17:12,440 --> 00:17:13,720
tools and infrastructure in place.

323
00:17:14,280 --> 00:17:19,320
There'll be a point here too where a lot of these things are just optimized away for us.

324
00:17:19,320 --> 00:17:24,920
Even if you don't know how to use these things extra well, there'll be enough data that

325
00:17:24,920 --> 00:17:27,480
kind of goes, okay, I think they're asking about this.

326
00:17:27,480 --> 00:17:30,280
Give them what they want, which is both good and bad.

327
00:17:30,280 --> 00:17:34,120
So one thing I would love to touch on too, I encourage folks, even if they don't have

328
00:17:34,120 --> 00:17:39,400
a direct need for a place for their prompts, their lonely prompts stored in some Google

329
00:17:39,400 --> 00:17:43,720
seat somewhere or your downloads folder in a TXT file, zipped away, never to be looked

330
00:17:43,720 --> 00:17:44,040
at again.

331
00:17:44,680 --> 00:17:49,240
Even if they don't have a need for those prompts to find a home, Dan and his team put out some

332
00:17:49,240 --> 00:17:52,520
excellent content on their blog about every week or two.

333
00:17:52,520 --> 00:17:57,400
I point people there all the time for some of your beginner's guides, breaking down how

334
00:17:57,400 --> 00:18:00,920
you talk about temperature and top P, some of those settings.

335
00:18:00,920 --> 00:18:03,800
I really encourage folks to check that out.

336
00:18:03,800 --> 00:18:08,200
Is there any one of these posts that has gotten the best reception or are you excited about

337
00:18:08,200 --> 00:18:09,880
something that you're working on right now?

338
00:18:09,880 --> 00:18:14,840
Yeah, I think our most, the one that's got the best reception is the multi-persona prompting

339
00:18:14,840 --> 00:18:19,160
one, which I think is just very interesting to think about because I think it may be

340
00:18:19,160 --> 00:18:22,200
mimics how you would expect humans to come to a good answer.

341
00:18:22,200 --> 00:18:26,920
And so for some quick context, multi-persona prompting basically is when you tell a model,

342
00:18:26,920 --> 00:18:30,040
hey, get a group of people together to solve the problem.

343
00:18:30,040 --> 00:18:33,640
And then they, the model kind of leads this brainstorming session and it's really cool.

344
00:18:33,640 --> 00:18:34,760
You can watch it happen.

345
00:18:35,320 --> 00:18:38,040
And then they collaborate and they come to a final answer.

346
00:18:38,040 --> 00:18:40,920
And that one's definitely gotten a lot of reception.

347
00:18:40,920 --> 00:18:48,600
We just launched a new one on based on a research article from DeepMind about LLMs for optimizing

348
00:18:48,600 --> 00:18:50,280
prompts that I thought was really interesting.

349
00:18:50,280 --> 00:18:51,880
So that was the last one that came out.

350
00:18:51,880 --> 00:18:58,040
I saw it, the one where AIs optimize themselves and the winning one is take a deep breath.

351
00:18:58,040 --> 00:18:59,080
Yeah, that's this one.

352
00:18:59,080 --> 00:19:00,280
It was like laughing.

353
00:19:00,280 --> 00:19:00,760
Yeah.

354
00:19:00,760 --> 00:19:04,840
The interesting part about that too is if you scroll down, Wes, is actually that I have

355
00:19:04,840 --> 00:19:09,480
the graph in here, but that take a deep breath part is related to, yeah, there it is.

356
00:19:10,200 --> 00:19:11,720
It's related to the Palm model.

357
00:19:11,720 --> 00:19:16,120
But if you look at the best top instruction for GPT-4, it's now five sentences or five

358
00:19:16,120 --> 00:19:16,920
lines there.

359
00:19:16,920 --> 00:19:21,240
And so the difference in terms of what the top instruction was for each model, it was

360
00:19:21,240 --> 00:19:26,680
a huge variance, which is like a kind of a head scratcher, but another just like really

361
00:19:26,680 --> 00:19:29,720
interesting intricacy across different models even.

362
00:19:29,720 --> 00:19:35,880
So this again just proves the value of Promo Hub, what you showed then once you add different

363
00:19:35,880 --> 00:19:40,680
models, you will be able to see what is valid put and how it differs.

364
00:19:40,680 --> 00:19:41,400
Exactly.

365
00:19:41,400 --> 00:19:41,880
Yeah.

366
00:19:41,880 --> 00:19:45,400
And I think that's going to be really huge and really fun to mess around with.

367
00:19:45,400 --> 00:19:49,560
So this blog basically lives and breathes on Promo Hub.

368
00:19:49,560 --> 00:19:49,880
Yes.

369
00:19:49,880 --> 00:19:52,280
Like we said, I encourage people to check this out.

370
00:19:52,280 --> 00:19:56,760
I think that's very funny, but if you think about it too, if you're looking for the best

371
00:19:56,760 --> 00:20:02,600
response, I'm sure at some point to get those responses, there's enough cross pathways in

372
00:20:02,600 --> 00:20:06,360
that neural network that goes, all right, well, the people that are producing the best

373
00:20:06,360 --> 00:20:08,760
responses are relaxed and we've relaxed them.

374
00:20:08,760 --> 00:20:13,560
Or there's something, and it's just so much a reflection of a perfect articulation of how

375
00:20:13,560 --> 00:20:18,600
these models are a reflection of us, who we are as humans, because that is inherently

376
00:20:18,600 --> 00:20:22,440
a very human statement to yield a top construction outcome.

377
00:20:23,320 --> 00:20:26,040
You've been around this kind of world for a while.

378
00:20:26,040 --> 00:20:30,440
I mean, you still had two other companies, you've done the Silicon Valley thing.

379
00:20:30,440 --> 00:20:35,880
I would love to get your perspective or your advice for people that maybe have a good idea

380
00:20:35,880 --> 00:20:39,960
that they don't know where to begin, or maybe they do know where to begin, but they don't

381
00:20:39,960 --> 00:20:41,480
know how to get to that next step.

382
00:20:41,480 --> 00:20:44,120
Can you share a little bit of your perspective on that?

383
00:20:44,120 --> 00:20:49,560
I think the biggest thing is that there's always like a million reasons to not do something.

384
00:20:49,560 --> 00:20:54,760
And so to always just jump into it, because the worst case scenario is nobody cares.

385
00:20:55,640 --> 00:20:58,440
Nobody uses like you're saying, or like nobody goes too long.

386
00:20:58,440 --> 00:20:59,560
Watches your videos?

387
00:20:59,560 --> 00:21:01,080
No one watches your videos.

388
00:21:01,080 --> 00:21:05,480
I'm sure you probably made a million in the beginning where the views were in the 10

389
00:21:05,480 --> 00:21:08,920
digits or something small, but you got to start somewhere.

390
00:21:08,920 --> 00:21:13,560
I think people a lot of times maybe spin their wheels on like the perfect idea or I need

391
00:21:13,560 --> 00:21:15,560
the perfect team or I need to raise money first.

392
00:21:15,560 --> 00:21:17,800
And it's like, you don't need to do any of that stuff.

393
00:21:17,800 --> 00:21:21,720
The only thing you need is a product and hopefully some users.

394
00:21:21,720 --> 00:21:25,240
But a product with no users is actually maybe even more valuable because then you know,

395
00:21:25,240 --> 00:21:26,920
you should work on something else.

396
00:21:26,920 --> 00:21:29,800
Especially in AI these days, you need a seven page white paper.

397
00:21:30,920 --> 00:21:32,520
That's good for about 75 mil.

398
00:21:33,160 --> 00:21:34,200
Yeah, exactly.

399
00:21:34,200 --> 00:21:36,120
Not pointing fingers anywhere, Wes.

400
00:21:36,120 --> 00:21:36,440
Right.

401
00:21:36,440 --> 00:21:36,760
No.

402
00:21:36,760 --> 00:21:37,480
Just bring it up.

403
00:21:37,480 --> 00:21:39,160
That's the thing.

404
00:21:39,160 --> 00:21:41,960
It's so funny that now there's so much interest in that.

405
00:21:41,960 --> 00:21:45,720
That hopefully is not a bubble in some ways, but I'm sure a lot of people are going to

406
00:21:45,720 --> 00:21:50,600
lose a lot of money because they're just going, okay, you said GPT, I'm in.

407
00:21:50,600 --> 00:21:52,360
Now here's a hundred million dollar check.

408
00:21:52,360 --> 00:21:53,080
Yeah, it is.

409
00:21:53,080 --> 00:21:57,800
The fundraising landscape is certainly, you know, it's coming off a weird period where

410
00:21:57,800 --> 00:22:01,080
there was a kind of a bubble before this and then there's this new platform shift.

411
00:22:01,080 --> 00:22:04,920
And so the coalescing of the two has been really interesting to watch and I think will

412
00:22:04,920 --> 00:22:07,560
be for the next couple of years going forward.

413
00:22:07,560 --> 00:22:08,680
But yeah, I don't know.

414
00:22:08,680 --> 00:22:09,720
It's fun to watch.

415
00:22:09,720 --> 00:22:15,880
With your other two products, Tether and Modus, they both have a different kind of use.

416
00:22:15,880 --> 00:22:18,680
Was raising money for those or launching those?

417
00:22:18,680 --> 00:22:23,880
How were those inherently different from Prompt Hub's origin, so to speak?

418
00:22:23,880 --> 00:22:25,240
So Modus was the agency.

419
00:22:25,240 --> 00:22:27,240
So Modus was, we did client work.

420
00:22:27,240 --> 00:22:31,880
We would help mostly early stage companies and we built a lot of different things.

421
00:22:31,880 --> 00:22:37,480
And so it was through our work at Modus and our goal was always eventually to do our own

422
00:22:37,480 --> 00:22:39,240
product as much as we loved having clients.

423
00:22:39,240 --> 00:22:41,000
We wanted to try our own route at things.

424
00:22:41,000 --> 00:22:46,040
And so we then started to work on Tether from there to make bug reporting easier for our

425
00:22:46,040 --> 00:22:46,520
clients.

426
00:22:47,160 --> 00:22:51,240
And then, oh, again, and that was solving a problem that we ran into, which is what we

427
00:22:51,240 --> 00:22:55,320
also did with Prompt Hub, which is another piece of advice I would say is that if you

428
00:22:55,320 --> 00:22:58,840
want to try and start something, but you like, maybe don't have an idea, just look for whatever

429
00:22:58,840 --> 00:23:02,680
annoys you the most and that's like a good place for inspiration.

430
00:23:02,680 --> 00:23:07,160
Yeah, Tether came from a problem we had to help our clients and then people started to

431
00:23:07,160 --> 00:23:07,960
like that.

432
00:23:07,960 --> 00:23:12,040
And then we kind of dope-head first once we got into the accelerator that allowed us to

433
00:23:12,040 --> 00:23:16,520
go in, drop our clients that we're currently working with and dive in with Tether.

434
00:23:16,520 --> 00:23:18,520
And then eventually that led us to Prompt Hub.

435
00:23:18,520 --> 00:23:23,720
Are you guys actively out there kind of seeking funding?

436
00:23:23,720 --> 00:23:25,960
Are you looking for investors?

437
00:23:25,960 --> 00:23:32,840
I know we have a great list of onboarders, a waiting list that will share an exciting

438
00:23:32,840 --> 00:23:37,320
code that Dan has set up for our listeners and viewers to get a little earlier access

439
00:23:37,320 --> 00:23:38,280
to Prompt Hub.

440
00:23:38,280 --> 00:23:42,920
But what's kind of in the three to six month roadmap for Prompt Hub?

441
00:23:42,920 --> 00:23:43,960
What are you hoping to achieve?

442
00:23:43,960 --> 00:23:47,080
Yes, I mean, on the fundraising front, we aren't raising right now.

443
00:23:47,080 --> 00:23:49,080
We may in later on in the year.

444
00:23:49,080 --> 00:23:51,560
But right now we're currently just heads down in the product.

445
00:23:51,560 --> 00:23:56,280
And looking forward, adding different models, making the testing suite a little bit more

446
00:23:56,280 --> 00:23:58,440
built out is something we're focusing on.

447
00:23:59,160 --> 00:24:02,520
And then what we really want to do is we want to make it easier to deploy prompts.

448
00:24:02,520 --> 00:24:06,840
And so the kind of prompts is a service model, which we have in its current form.

449
00:24:06,840 --> 00:24:12,680
But we want to make it more easier to access, have a really developed API that people can

450
00:24:12,680 --> 00:24:16,280
integrate stuff into whatever type of workflow that they may have.

451
00:24:16,280 --> 00:24:21,000
Because the thing with AI and people trying to leverage it is every situation is a little

452
00:24:21,000 --> 00:24:21,560
bit different.

453
00:24:21,560 --> 00:24:23,640
And so we need to be a really malleable solution.

454
00:24:24,200 --> 00:24:28,360
And so we'll be really pushing on that so that we can integrate us anywhere, however

455
00:24:28,360 --> 00:24:28,920
you want.

456
00:24:28,920 --> 00:24:34,440
I totally could see, for example, what you shared with LinkedIn post generator, like

457
00:24:35,080 --> 00:24:41,080
maybe marketing package where it integrates the buffer or distribution thing that you

458
00:24:41,080 --> 00:24:45,320
generate a prompt and you can immediately schedule that on your social media.

459
00:24:45,320 --> 00:24:47,960
It doesn't need to that you develop scheduling.

460
00:24:47,960 --> 00:24:52,440
It can be just plug and play with some of the platforms.

461
00:24:52,440 --> 00:24:56,040
I think we may need to make that com integration is what I'm reading from that.

462
00:24:56,040 --> 00:24:59,000
Yes, make.com.

463
00:24:59,000 --> 00:25:01,080
Yeah, let's go to the new spot set.

464
00:25:01,080 --> 00:25:05,880
Yeah, I don't think I should be announcing it before I make a video.

465
00:25:05,880 --> 00:25:07,800
Oh, we'll do this in a couple of weeks.

466
00:25:07,800 --> 00:25:10,760
Now I need to run make video, but yeah, make.com.

467
00:25:10,760 --> 00:25:11,800
We partnered.

468
00:25:11,800 --> 00:25:15,320
I'm very excited to show some of the things that our team built.

469
00:25:15,320 --> 00:25:20,600
And I'm looking to plug some integrations for myself and the whole social media building

470
00:25:20,600 --> 00:25:23,000
brand is constantly running in my head.

471
00:25:23,000 --> 00:25:24,600
And I know that it's full time job.

472
00:25:24,600 --> 00:25:29,160
So I'm just looking how to make it way smarter and smoother and run it with AI.

473
00:25:29,160 --> 00:25:34,280
So I actually use things that I talk about at full scope and scale.

474
00:25:35,080 --> 00:25:37,400
Yeah, no, that sounds like a great use case.

475
00:25:37,400 --> 00:25:40,200
I mean, yeah, we have a bunch of people who were using make.

476
00:25:40,200 --> 00:25:44,440
Obviously the comp is safe here, but it seems like it's growing a lot more, at least in

477
00:25:44,440 --> 00:25:45,480
my circles.

478
00:25:45,480 --> 00:25:46,200
So that's really great.

479
00:25:46,200 --> 00:25:47,000
Congrats on that.

480
00:25:47,000 --> 00:25:47,800
Thank you.

481
00:25:47,800 --> 00:25:54,680
I don't know if it's good time and place, but as Wes asked you about what's coming in

482
00:25:54,680 --> 00:26:01,160
next three months, I know that you have something exciting coming where you actually bring the

483
00:26:01,160 --> 00:26:07,240
best minds in prompt engineering together, you know, to discuss all these research papers

484
00:26:07,240 --> 00:26:08,360
and development.

485
00:26:08,360 --> 00:26:11,160
Do you want to walk us a bit through that?

486
00:26:11,160 --> 00:26:12,120
Yeah, definitely.

487
00:26:12,120 --> 00:26:17,480
So in early October here, just about a month away, October 12th, I am helping co-league,

488
00:26:17,480 --> 00:26:19,720
co-run a prompt engineering conference.

489
00:26:19,720 --> 00:26:22,440
And the domain is promptengineering.rocks.

490
00:26:22,440 --> 00:26:26,680
I'm doing it alongside a few other people and it's all, you know, not for profit and

491
00:26:26,680 --> 00:26:30,680
just for people to get together to learn about different types of prompting methods, learn

492
00:26:30,680 --> 00:26:34,840
about how people are using prompts in their, you know, their daily lives.

493
00:26:34,840 --> 00:26:36,120
And it could be really good.

494
00:26:36,120 --> 00:26:40,360
We have a bunch of really good profound speakers coming in, a professor from Oxford, some people

495
00:26:40,360 --> 00:26:41,560
from Microsoft.

496
00:26:41,560 --> 00:26:45,240
A lot of different use cases that it's all, it's very practical.

497
00:26:45,240 --> 00:26:48,200
It's not, you don't have to learn, we're not going to be telling you what a neural network

498
00:26:48,200 --> 00:26:48,920
is or anything like that.

499
00:26:48,920 --> 00:26:52,840
We're kind of trying to get our hands dirty with it and show you how you can walk away

500
00:26:52,840 --> 00:26:57,320
with something that's helpful for you and online free one day.

501
00:26:57,320 --> 00:26:59,480
And it could be, yeah, it should be a lot of fun.

502
00:26:59,480 --> 00:27:00,120
Fantastic.

503
00:27:00,680 --> 00:27:02,920
I'm looking forward to that, right, Wes?

504
00:27:02,920 --> 00:27:03,640
Yeah, I think so.

505
00:27:03,640 --> 00:27:07,400
We may have a little bit of involvement in that process.

506
00:27:07,400 --> 00:27:14,520
So stay tuned for more announcements from your friends at HTTTA about our participation

507
00:27:14,520 --> 00:27:18,520
in the first annual prompt engineering conference, October 12th.

508
00:27:18,520 --> 00:27:23,400
And Wes, I know that you slightly mentioned and we may be brushed through that, but this

509
00:27:23,400 --> 00:27:26,920
a little gift to the listeners to try Promt Hub.

510
00:27:26,920 --> 00:27:28,440
Let's dive into that.

511
00:27:28,440 --> 00:27:35,000
So yes, folks can go to prompthub.us, click on the join waitlist.

512
00:27:35,000 --> 00:27:42,280
And if they use the code HTTTAI, here's the VIP section, ladies and gentlemen.

513
00:27:42,280 --> 00:27:46,200
We'll let you come on down to the red carpet and give Promt Hub a try.

514
00:27:46,200 --> 00:27:47,400
Give Dan some feedback.

515
00:27:47,400 --> 00:27:49,560
Give them a follow on LinkedIn.

516
00:27:49,560 --> 00:27:54,120
I mean, even if you don't have any need, like I said, do yourself a favor, follow the blog.

517
00:27:54,120 --> 00:27:54,920
Oh my gosh.

518
00:27:54,920 --> 00:28:00,200
We cited your blog three times in our record setting AI and chat GPT for everyone course,

519
00:28:00,200 --> 00:28:03,000
because it's just, I can't describe it better than this.

520
00:28:03,000 --> 00:28:05,160
Come read Dan's blog, come watch Dan's video.

521
00:28:05,160 --> 00:28:06,040
I appreciate that.

522
00:28:06,040 --> 00:28:08,920
I did, there is one note on that for the access front.

523
00:28:08,920 --> 00:28:10,680
I'm actually, you have to go to a different link.

524
00:28:10,680 --> 00:28:12,680
So it's not the join waitlist.

525
00:28:12,680 --> 00:28:13,180
Oops.

526
00:28:14,600 --> 00:28:17,320
If you go, different red carpet.

527
00:28:17,320 --> 00:28:22,680
Like, yeah, if you just go to our site and hit the login button and then there'll be

528
00:28:22,680 --> 00:28:24,920
a sign up button beneath that.

529
00:28:24,920 --> 00:28:27,480
And you know, sign up today will be what the link there says.

530
00:28:27,480 --> 00:28:31,080
And you'll be able to type in that code and we all give you a direct link as well.

531
00:28:31,080 --> 00:28:34,760
But that would be the fastest way to go log in and then sign up today.

532
00:28:34,760 --> 00:28:35,000
Yeah.

533
00:28:35,000 --> 00:28:39,400
We will include everything in description and their video newsletter.

534
00:28:40,440 --> 00:28:44,920
Sign up today and their app.prompt hub.us slash log in.

535
00:28:44,920 --> 00:28:48,440
You know, like if people want to try it out, could you have it?

536
00:28:48,440 --> 00:28:53,960
If you're comfortable to talk about the pricing model, like how these things actually run on

537
00:28:53,960 --> 00:28:56,680
your side versus the user side.

538
00:28:56,680 --> 00:28:57,080
Yeah.

539
00:28:57,080 --> 00:28:58,280
So great news there.

540
00:28:58,280 --> 00:28:59,320
Everything's free right now.

541
00:28:59,320 --> 00:29:01,800
Bring your API key and you're good to go.

542
00:29:01,800 --> 00:29:06,120
We will be rolling out monetization plans in the next month or so, but there will always

543
00:29:06,120 --> 00:29:07,320
be a free plan.

544
00:29:07,320 --> 00:29:11,880
But as of now, everything we've talked about today from forms to batch testing,

545
00:29:11,880 --> 00:29:13,160
everything's included.

546
00:29:13,160 --> 00:29:15,080
You can have as many team members as you want.

547
00:29:15,080 --> 00:29:19,160
And like I said, we really have a big goal with our blog and with the product is to help

548
00:29:19,160 --> 00:29:20,040
educate people.

549
00:29:20,040 --> 00:29:22,520
And so there will always be a free plan so you can access.

550
00:29:22,520 --> 00:29:23,000
Oh yeah.

551
00:29:23,000 --> 00:29:23,960
I love this.

552
00:29:23,960 --> 00:29:26,840
Like we're, I sense why we're very much aligned.

553
00:29:26,840 --> 00:29:33,960
In just a previous podcast with Joseph Rosenbaum, we talked exactly about this, why we gave

554
00:29:33,960 --> 00:29:39,560
professors synapse on YouTube free before course starts, because to educate people and

555
00:29:39,560 --> 00:29:45,000
actually give them value and to see the curiosity spark up out of that.

556
00:29:45,720 --> 00:29:50,600
And when people start learning and then we have all these testimonials now that we actually

557
00:29:50,600 --> 00:29:56,360
know what happens when people learn, but also apply and get value immediately at their work.

558
00:29:56,360 --> 00:29:58,760
That's the most fascinating to me.

559
00:29:58,760 --> 00:30:02,440
I appreciate you and your team's work trying to cut through the noise.

560
00:30:02,440 --> 00:30:05,000
That's what we try to do ourselves.

561
00:30:05,000 --> 00:30:09,960
I can't deal with another Forbes article that's like, he's there and chat GBT prompts will

562
00:30:09,960 --> 00:30:10,520
change your life.

563
00:30:10,520 --> 00:30:15,720
And it's one sentence prompt that I'm like, come on, don't just put your toe in the pool.

564
00:30:15,720 --> 00:30:18,040
What it was six, eight months now.

565
00:30:18,040 --> 00:30:21,560
Why don't you know it's like all the news.

566
00:30:21,560 --> 00:30:23,560
Everyone knows, but it's not good stuff.

567
00:30:23,560 --> 00:30:27,720
It's all one today was like five chat, JPT prompts to feel invincible at work.

568
00:30:27,720 --> 00:30:29,160
I'm like, what's that even mean?

569
00:30:29,160 --> 00:30:32,040
You mean you don't like being told that you have no idea how you're using chat,

570
00:30:32,040 --> 00:30:36,920
GPT and that the only way is to become in the top 2% is to take this prompt sheet that I

571
00:30:36,920 --> 00:30:37,080
have.

572
00:30:37,080 --> 00:30:38,680
You don't, you don't like those codes.

573
00:30:38,680 --> 00:30:43,960
You need Dan's top 3000 prompts for only the low price of 9.99.

574
00:30:43,960 --> 00:30:48,920
Thank you so much for jumping on the pod today with us, Dan, aside from shooting over

575
00:30:48,920 --> 00:30:54,520
to prompt hub that us where and checking out prompt engineering dot rocks for the upcoming

576
00:30:54,520 --> 00:30:55,560
prompt engineering conference.

577
00:30:55,560 --> 00:30:57,720
Where else can folks find what you got going on?

578
00:30:57,720 --> 00:31:01,560
Yeah, you can connect with me on LinkedIn, Dan Cleary and we'll include a link, but yeah,

579
00:31:01,560 --> 00:31:03,080
connect with me on Twitter as well.

580
00:31:03,080 --> 00:31:04,920
Dan underscore Cleary, but we'll link there.

581
00:31:04,920 --> 00:31:08,520
But yeah, if you have any questions about prompt engineering, any of this stuff, I would

582
00:31:08,520 --> 00:31:09,560
always love to chat more.

583
00:31:09,560 --> 00:31:11,640
And so those are the two places I'm most active.

584
00:31:11,640 --> 00:31:18,440
And for our listeners and alumni from the course, can we ping you on the synth minds

585
00:31:18,440 --> 00:31:19,320
discord?

586
00:31:19,320 --> 00:31:20,040
Yeah, of course.

587
00:31:20,040 --> 00:31:21,320
I'm right in there with you guys.

588
00:31:21,320 --> 00:31:26,760
I'm happy and it's been great to collaborate with y'all and I go to, like you said, I think

589
00:31:26,760 --> 00:31:29,240
we share a lot of values in terms of being open.

590
00:31:29,240 --> 00:31:33,800
And I think your discord is a good example of Matt in all tides or tides rate, high tides

591
00:31:33,800 --> 00:31:37,400
rates, all ships or whatever they're saying as it goes, as we teach people and they can

592
00:31:37,400 --> 00:31:38,440
apply it more.

593
00:31:38,440 --> 00:31:39,720
We all move forward.

594
00:31:39,720 --> 00:31:39,960
Yeah.

595
00:31:39,960 --> 00:31:42,440
Feel free to reach out and send my discord as well.

596
00:31:42,440 --> 00:31:43,800
This is perfect.

597
00:31:43,800 --> 00:31:45,000
No better words.

598
00:31:45,000 --> 00:31:47,320
Yes, I have nothing to add.

599
00:31:47,320 --> 00:31:47,960
There you go.

600
00:31:47,960 --> 00:31:54,200
So for go to go, Dan Cleary, I am West Synth Mind saying happy prompting everybody.

601
00:31:54,200 --> 00:31:56,360
Happy prompting everybody.

602
00:31:56,360 --> 00:31:57,640
Then you can also do this.

603
00:31:57,640 --> 00:31:58,680
Happy prompting everyone.

604
00:32:06,840 --> 00:32:12,200
As always, you can check out the show notes and links at howtotalk2.ai.

605
00:32:13,160 --> 00:32:14,680
That's all for this week's episode.

606
00:32:14,680 --> 00:32:17,800
Happy prompting everyone.