1
00:00:00,600 --> 00:00:03,080
Speaker 1: At nine am on a Tuesday and you're working on

2
00:00:03,120 --> 00:00:06,440
Speaker 1: an important project and decide to use Jenai for some help.

3
00:00:07,120 --> 00:00:09,840
Speaker 1: But what if I told you that relying on just

4
00:00:10,039 --> 00:00:13,920
Speaker 1: one Jenai is like asking just one single expert for

5
00:00:14,000 --> 00:00:17,360
Speaker 1: advice when you can have an entire panel of specialists.

6
00:00:18,400 --> 00:00:21,400
Speaker 1: For the last two years, I've been using jeni daily,

7
00:00:21,440 --> 00:00:24,600
Speaker 1: and one of the biggest game changes has been treating

8
00:00:24,680 --> 00:00:28,280
Speaker 1: AI like a think tank rather than a single assistant.

9
00:00:29,000 --> 00:00:33,840
Speaker 1: Different AI models have unique strengths, and by strategically combining them,

10
00:00:33,880 --> 00:00:37,600
Speaker 1: you can dramatically enhance the quality of your work. And

11
00:00:37,600 --> 00:00:41,920
Speaker 1: that's what today's episode is all about. I am joined

12
00:00:41,960 --> 00:00:46,960
Speaker 1: by Inventium's resident AI expert, Neo Afflin to unpack how

13
00:00:47,040 --> 00:00:50,639
Speaker 1: you can leave ridge multiple Genai tools to turn them

14
00:00:50,640 --> 00:00:54,800
Speaker 1: into your personal advisory board. So whether you're writing, strategizing,

15
00:00:54,920 --> 00:00:58,760
Speaker 1: or problem solving, this approach could completely transform the way

16
00:00:58,800 --> 00:01:03,120
Speaker 1: you use Genai. Just quickly before we get into today's episode,

17
00:01:03,200 --> 00:01:05,840
Speaker 1: if you want to go DeFi and learn how to

18
00:01:05,880 --> 00:01:09,760
Speaker 1: work with AI instead of worrying about being replaced by it,

19
00:01:10,280 --> 00:01:13,080
Speaker 1: I have something that I think you will love, but

20
00:01:13,160 --> 00:01:17,759
Speaker 1: my consultancy you invent him. We've created the Jenai productivity system.

21
00:01:17,880 --> 00:01:20,640
Speaker 1: It's a self paced online course designed to help you

22
00:01:20,680 --> 00:01:25,119
Speaker 1: boost productivity and save literally hours every week by using

23
00:01:25,200 --> 00:01:28,319
Speaker 1: Jenai a whole lot smarter. In just a few hours,

24
00:01:28,319 --> 00:01:32,960
Speaker 1: you'll go from Jenai dabbler to Jenai expert with practical

25
00:01:32,959 --> 00:01:36,959
Speaker 1: tools you can start using immediately. Head to inventium dot

26
00:01:37,000 --> 00:01:40,640
Speaker 1: com dot au forward slash Jnai to check it out

27
00:01:40,680 --> 00:01:48,320
Speaker 1: and get started today. Welcome to How I Work, a

28
00:01:48,400 --> 00:01:52,400
Speaker 1: show about habits, rituals, and strategies for optimizing your day.

29
00:01:53,000 --> 00:01:57,040
Speaker 1: I'm your host, doctor Amantha imber So. A lot of

30
00:01:57,240 --> 00:02:01,280
Speaker 1: people out there just rely on one NAI tool like

31
00:02:01,440 --> 00:02:05,800
Speaker 1: Copilot or chat GPT. I'm curious as to why you

32
00:02:05,880 --> 00:02:08,840
Speaker 1: think that might actually limit the results that someone is

33
00:02:08,880 --> 00:02:10,040
Speaker 1: getting from the Genai.

34
00:02:10,880 --> 00:02:14,040
Speaker 2: The first thing I'd say to that is the Genais

35
00:02:14,080 --> 00:02:16,280
Speaker 2: are brilliant, and you're getting great results out of a

36
00:02:16,280 --> 00:02:19,040
Speaker 2: lot of these different Genais. So take this as a

37
00:02:19,120 --> 00:02:21,920
Speaker 2: next level thing. But you'll find that each of the

38
00:02:21,960 --> 00:02:26,240
Speaker 2: Ai tools have their own abilities, and some of them

39
00:02:26,280 --> 00:02:29,919
Speaker 2: are better at certain things than others, and so using

40
00:02:30,000 --> 00:02:32,720
Speaker 2: more than one Genai tool will give you more than

41
00:02:32,760 --> 00:02:33,440
Speaker 2: one perspective.

42
00:02:34,320 --> 00:02:36,400
Speaker 1: I know that something that you do a lot of

43
00:02:36,600 --> 00:02:41,160
Speaker 1: is using multiple genais as a feedback loop. Can you

44
00:02:41,240 --> 00:02:44,000
Speaker 1: tell me a bit more about what you're doing here.

45
00:02:44,720 --> 00:02:49,000
Speaker 2: When I'm writing something, or I'm doing deep analysis, or

46
00:02:49,160 --> 00:02:52,359
Speaker 2: I'm working with a client on something, then I want

47
00:02:52,360 --> 00:02:54,520
Speaker 2: to get the best result. And so what I'm doing

48
00:02:54,600 --> 00:02:58,040
Speaker 2: is I'm working deeply with one GENAI to get one perspective,

49
00:02:58,120 --> 00:03:01,480
Speaker 2: and generally it's more about industry knowledge or structure or

50
00:03:01,520 --> 00:03:04,720
Speaker 2: problem solving or which things could be included. And then

51
00:03:04,760 --> 00:03:08,240
Speaker 2: I'll do exactly the same work on a completely different GENAI.

52
00:03:08,400 --> 00:03:10,560
Speaker 2: So I might use chat GPT for one of them,

53
00:03:10,800 --> 00:03:12,600
Speaker 2: and then I'll go to Claude and do the same

54
00:03:12,600 --> 00:03:16,000
Speaker 2: amount of work on Claude and get Claude's perspective. And

55
00:03:16,040 --> 00:03:18,200
Speaker 2: what I'll find is Claude will come up with things

56
00:03:18,240 --> 00:03:19,560
Speaker 2: that chat TIPT may not.

57
00:03:20,160 --> 00:03:22,400
Speaker 1: Do you remember what inspired you to take this approach

58
00:03:22,400 --> 00:03:23,160
Speaker 1: in the first place.

59
00:03:23,720 --> 00:03:25,600
Speaker 2: Maybe it's because I'm a bit of a GENAI geek

60
00:03:25,600 --> 00:03:27,520
Speaker 2: and I want to just try all the different genis.

61
00:03:28,160 --> 00:03:31,040
Speaker 2: But what I found is that going to each of

62
00:03:31,080 --> 00:03:34,520
Speaker 2: them gives me a different perspective, not just on knowledge.

63
00:03:34,520 --> 00:03:37,520
Speaker 2: By the way, I will say this on language. Going

64
00:03:37,560 --> 00:03:40,520
Speaker 2: to one, from one to the other, you might find

65
00:03:40,560 --> 00:03:43,160
Speaker 2: better ways to phrase a sentence or phrase a paragraph,

66
00:03:43,280 --> 00:03:47,120
Speaker 2: So it's same principles apply even with writing.

67
00:03:47,120 --> 00:03:51,160
Speaker 1: It's great. I've heard you compare using multiple genais to

68
00:03:51,240 --> 00:03:55,000
Speaker 1: having a room full of expert consultants. Could you just

69
00:03:55,440 --> 00:03:57,000
Speaker 1: expand on what you mean by that.

70
00:03:57,280 --> 00:03:59,920
Speaker 2: Yeah, it's almost like having two rooms. Like you've got

71
00:04:00,120 --> 00:04:03,360
Speaker 2: one room of consultants because you've got to CHATCHAPT and

72
00:04:03,360 --> 00:04:07,200
Speaker 2: then get a completely different room of consultants with Claude

73
00:04:07,280 --> 00:04:09,720
Speaker 2: or with Gemini or any of the others, because they've

74
00:04:09,760 --> 00:04:13,320
Speaker 2: got different ways that they've absorbed their knowledge and different

75
00:04:13,320 --> 00:04:16,039
Speaker 2: ways that they kind of work through that knowledge. So

76
00:04:16,080 --> 00:04:19,360
Speaker 2: you get two different sets of opinions back, which I love.

77
00:04:19,920 --> 00:04:24,240
Speaker 1: Now. I've heard you say that different GENAI models have

78
00:04:24,320 --> 00:04:28,640
Speaker 1: different strengths and weaknesses, and I imagine that people listening

79
00:04:29,240 --> 00:04:31,600
Speaker 1: have their favorite, or they go to one or maybe

80
00:04:31,640 --> 00:04:33,920
Speaker 1: the one the only one that they're not blocked from

81
00:04:34,040 --> 00:04:37,080
Speaker 1: using at their workplace. I would love to know if

82
00:04:37,080 --> 00:04:40,839
Speaker 1: you could briefly break down with the major genais that

83
00:04:40,880 --> 00:04:44,760
Speaker 1: people are using, which what like, what are the things

84
00:04:44,800 --> 00:04:46,799
Speaker 1: that each one is particularly good at.

85
00:04:47,160 --> 00:04:50,160
Speaker 2: So let's start with a big granddaddy CHATCHPT, which is

86
00:04:50,720 --> 00:04:57,800
Speaker 2: great for analysis logic general, it's an awesome everyday workhorse,

87
00:04:58,200 --> 00:05:01,760
Speaker 2: fantastic for that. It's also got very good logic tools

88
00:05:01,760 --> 00:05:04,039
Speaker 2: if you want to use the one and the three models.

89
00:05:04,240 --> 00:05:06,960
Speaker 2: So if you need analysis and maths and stuff like that,

90
00:05:07,160 --> 00:05:10,800
Speaker 2: brilliant for that. If you're a writer, go for Claude.

91
00:05:10,960 --> 00:05:13,560
Speaker 2: Claude just out of the box, sounds a lot more

92
00:05:13,640 --> 00:05:17,440
Speaker 2: like a human than an AI and I'm particularly impressed

93
00:05:17,440 --> 00:05:19,360
Speaker 2: what they do on safety as well. So with writing,

94
00:05:19,400 --> 00:05:23,880
Speaker 2: they're just great. Also with coding, it does coding and

95
00:05:23,920 --> 00:05:28,000
Speaker 2: development problems a lot better than others. Gemini, I've got

96
00:05:28,040 --> 00:05:30,400
Speaker 2: to say, is really impressed me. They've come up with

97
00:05:30,440 --> 00:05:34,840
Speaker 2: a Google Gemini. They've come up with a lot more,

98
00:05:34,880 --> 00:05:39,640
Speaker 2: more tools, better, better models. It's not perfect yet, but

99
00:05:39,920 --> 00:05:43,520
Speaker 2: don't discount googling this one. This race is going to

100
00:05:43,520 --> 00:05:45,480
Speaker 2: be hotly contested and Google's going to be up there

101
00:05:45,520 --> 00:05:47,440
Speaker 2: with everyone. So have a play with that at the

102
00:05:47,440 --> 00:05:49,760
Speaker 2: moment that it's free, so definitely jump in and have

103
00:05:49,839 --> 00:05:55,000
Speaker 2: a dive. Copilot is excellent for the workforce. I personally

104
00:05:55,000 --> 00:05:57,640
Speaker 2: don't find it as user friendly as chat GPT, but

105
00:05:57,680 --> 00:06:01,320
Speaker 2: the way it locks down for your workforce is And

106
00:06:01,360 --> 00:06:03,000
Speaker 2: then do have a play with the others as well.

107
00:06:03,040 --> 00:06:03,880
Speaker 2: They're also good fun.

108
00:06:04,320 --> 00:06:09,359
Speaker 1: Okay, so can you walk me through a real world

109
00:06:09,360 --> 00:06:14,240
Speaker 1: example of using multiple genais to improve a piece of work?

110
00:06:15,120 --> 00:06:19,720
Speaker 1: And I'm thinking, let's say I do a lot of

111
00:06:19,960 --> 00:06:23,520
Speaker 1: business development work and part of my role, so creating

112
00:06:23,520 --> 00:06:27,160
Speaker 1: a sales proposal is something that I do a lot of.

113
00:06:27,279 --> 00:06:30,559
Speaker 1: So how would you recommend that I do that better?

114
00:06:31,120 --> 00:06:35,760
Speaker 2: Yeah? I do it in two chunks. One is the

115
00:06:35,760 --> 00:06:38,839
Speaker 2: structure and pitch and what you're trying to put together.

116
00:06:39,320 --> 00:06:42,960
Speaker 2: I'd work with say, chat YOURPT first, and then say

117
00:06:42,960 --> 00:06:45,640
Speaker 2: how should I structure it. Here's my goals, here's my customers,

118
00:06:45,680 --> 00:06:48,279
Speaker 2: ask questions and have a discussion with it, and then

119
00:06:48,520 --> 00:06:51,960
Speaker 2: totally separate to that. Do the same thing in claude

120
00:06:52,040 --> 00:06:54,240
Speaker 2: and then you'll get a structure. Once you've got that,

121
00:06:54,279 --> 00:06:56,680
Speaker 2: you've then got two opinions on what you should cover,

122
00:06:57,080 --> 00:06:59,480
Speaker 2: and then you've then built yourself out of the two

123
00:07:00,040 --> 00:07:02,680
Speaker 2: structure you want and the goals and kind of points

124
00:07:02,680 --> 00:07:04,919
Speaker 2: you want to cover. Once you've got that, do the

125
00:07:04,960 --> 00:07:07,000
Speaker 2: same thing with the actual writing. So say here's a

126
00:07:07,040 --> 00:07:09,400
Speaker 2: couple of paragraphs, let's flesh that out together. Do that

127
00:07:09,440 --> 00:07:11,640
Speaker 2: with chat GPT, and then do it with Claude and

128
00:07:11,640 --> 00:07:13,240
Speaker 2: you'll see you'll get two different results.

129
00:07:13,840 --> 00:07:16,280
Speaker 1: What are some other key factors that we might need

130
00:07:16,320 --> 00:07:20,160
Speaker 1: to consider when structuring a MULTIAI feedback loop.

131
00:07:20,760 --> 00:07:23,880
Speaker 2: Well, first off, don't overcope complicate it. Not every problem

132
00:07:24,000 --> 00:07:27,480
Speaker 2: needs to go to two different ais. But I would

133
00:07:27,600 --> 00:07:31,200
Speaker 2: use it less about the writing. It's more about the ideas.

134
00:07:31,640 --> 00:07:33,360
Speaker 2: So if you've got a complex thing you're not quite

135
00:07:33,400 --> 00:07:35,720
Speaker 2: sure about the ideas or the structure of a document

136
00:07:35,920 --> 00:07:39,080
Speaker 2: or some kind of analysis paper or a pitch deck

137
00:07:39,200 --> 00:07:41,320
Speaker 2: or something like that, these are the great things to

138
00:07:41,360 --> 00:07:45,840
Speaker 2: go do for two opinions. They're analysis or thinking deeply

139
00:07:45,880 --> 00:07:50,960
Speaker 2: through things, structure of documents, kind of what's right or wrong,

140
00:07:51,080 --> 00:07:53,440
Speaker 2: pros and cons. Those kind of things are brilliant to

141
00:07:53,480 --> 00:07:56,840
Speaker 2: do the two model approach. If you're doing just say writing,

142
00:07:56,960 --> 00:07:59,480
Speaker 2: it's probably not as appropriate or not as needed to

143
00:07:59,480 --> 00:08:03,120
Speaker 2: go to the two ais. But yeah, ideas constructs, those

144
00:08:03,160 --> 00:08:04,040
Speaker 2: kind of things brilliant.

145
00:08:04,440 --> 00:08:07,400
Speaker 1: Is there a point of diminishing returns, like when you

146
00:08:07,400 --> 00:08:11,080
Speaker 1: should just stop it with all the genais and just

147
00:08:11,240 --> 00:08:12,760
Speaker 1: go okay, we're done here.

148
00:08:13,280 --> 00:08:17,160
Speaker 2: Absolutely, And I'm guilty of trying them out because I

149
00:08:17,240 --> 00:08:20,400
Speaker 2: enjoy trying them out, But I think for most people, yeah,

150
00:08:20,440 --> 00:08:22,840
Speaker 2: I think just only a few times that you would

151
00:08:22,840 --> 00:08:26,560
Speaker 2: need to particularly on those deep problem solving things, and

152
00:08:26,880 --> 00:08:28,800
Speaker 2: don't need to do it all the time. But if

153
00:08:28,840 --> 00:08:31,400
Speaker 2: you've got those problems solving things, it's a really powerful tool.

154
00:08:32,240 --> 00:08:35,360
Speaker 1: Okay. So for someone listening today who wants to get

155
00:08:35,360 --> 00:08:38,400
Speaker 1: in and start trying this immediately, what is the easiest

156
00:08:38,400 --> 00:08:39,240
Speaker 1: way to get started?

157
00:08:39,520 --> 00:08:42,320
Speaker 2: First off, sign up to the different tools, sign up

158
00:08:42,360 --> 00:08:46,200
Speaker 2: to Claude, chat, EPG, Gemini, try it the others if

159
00:08:46,240 --> 00:08:49,920
Speaker 2: you'd like, and then just play with them and see

160
00:08:49,960 --> 00:08:52,000
Speaker 2: the different opinions. You get back on some of the

161
00:08:52,000 --> 00:08:55,240
Speaker 2: problems you're solving, and then you might find that certain

162
00:08:55,240 --> 00:08:57,040
Speaker 2: tools are better for certain problems.

163
00:08:57,600 --> 00:09:01,600
Speaker 1: Thank you so much, Neo. I'm feeling very inspire to

164
00:09:01,720 --> 00:09:06,520
Speaker 1: use this strategy today. Thank you so much. If you're

165
00:09:06,679 --> 00:09:10,120
Speaker 1: keen to learn even more Genai tricks and hacks, check

166
00:09:10,160 --> 00:09:14,559
Speaker 1: out Inventium's Jenai Productivity System. It's a self faced course

167
00:09:14,600 --> 00:09:19,200
Speaker 1: to transform you from a Jenai dabbler to jenaipro in

168
00:09:19,400 --> 00:09:22,880
Speaker 1: just a few hours. Head to inventium dot com dot

169
00:09:22,920 --> 00:09:26,400
Speaker 1: au forward slash Jenai to check it out and get

170
00:09:26,400 --> 00:09:29,880
Speaker 1: started today. If you like today's show, make sure you

171
00:09:30,000 --> 00:09:33,000
Speaker 1: hit follow on your podcast app to be alerted when

172
00:09:33,080 --> 00:09:36,400
Speaker 1: new episodes drop. How I Work was recorded on the

173
00:09:36,400 --> 00:09:39,600
Speaker 1: traditional land of the Warringery people, part of the Cooler Nation.

174
00:09:40,160 --> 00:09:42,160
Speaker 1: A big thank you to Martin Nimber for doing the

175
00:09:42,240 --> 00:09:42,839
Speaker 1: sound mix.