Human First AI Why Replacing Humans Is a Terrible Business Strategy with Derek Crager S3-EP5

1
00:00:00,000 --> 00:00:04,500
I'm just curious, what happens if somebody asks something that's outside of its purview?

2
00:00:05,020 --> 00:00:07,520
Say, like, hey, Pachymetra, you want to chat today?

3
00:00:07,600 --> 00:00:08,400
How are you doing today?

4
00:00:08,460 --> 00:00:11,500
They're just lonely and want somebody to talk to.

5
00:00:11,640 --> 00:00:15,240
Does it have something that cues up and says, okay, I can't talk about that?

6
00:00:15,240 --> 00:00:18,580
Well, it's interesting you said that because mental wellness is huge.

7
00:00:18,840 --> 00:00:25,460
And it's in HR departments, you know, keeping track of how the workers actually feel day to day.

8
00:00:25,600 --> 00:00:28,340
Because are they going to come in tomorrow if it's the bottom line?

9
00:00:28,340 --> 00:00:31,200
Or can we make your day better?

10
00:00:31,940 --> 00:00:40,040
Good HR departments, good HR employees really do care about the employees.

11
00:00:40,320 --> 00:00:44,800
So in a situation like that, it's whatever the company.

12
00:00:58,340 --> 00:01:08,540
All right, everybody.

13
00:01:08,700 --> 00:01:11,960
Welcome back to another episode of the Space Tomatoes podcast.

14
00:01:12,340 --> 00:01:13,260
I'm one of your hosts, Havoc.

15
00:01:14,360 --> 00:01:15,640
I'm the other guy, LX.

16
00:01:15,760 --> 00:01:16,060
Hello.

17
00:01:16,320 --> 00:01:17,160
That other guy.

18
00:01:17,160 --> 00:01:17,740
Welcome back.

19
00:01:17,820 --> 00:01:18,360
All those things.

20
00:01:19,020 --> 00:01:24,760
Today we have with us special guest, Derek Crager, founder and CEO of Practical AI,

21
00:01:24,760 --> 00:01:31,500
creator of PocketMentor, and author of Human First AI. Welcome to the show, Derek.

22
00:01:32,100 --> 00:01:35,760
Well, thank you both. I appreciate the invite to bring me in.

23
00:01:36,660 --> 00:01:43,400
Absolutely. So, Human First AI, what is that concept and where does it come from?

24
00:01:44,320 --> 00:01:49,480
Yeah, absolutely. Well, we're actually in this early path to this AI adoption from the consumer

25
00:01:49,480 --> 00:01:54,560
level. You know, we're really, this is like our fourth year. And in internet timeframe,

26
00:01:54,560 --> 00:02:00,980
it really took the internet about 17 years for people to adopt the internet. It came out in 89.

27
00:02:01,440 --> 00:02:07,540
People really didn't hop on in mass until about 2006. So AI, we're still at this polarization

28
00:02:07,540 --> 00:02:15,400
point. And if we talk about the adoption, that's the human factor. And there's polarization.

29
00:02:15,400 --> 00:02:21,660
It's early and it really gets into change management, which is the psychology of adoption.

30
00:02:22,540 --> 00:02:29,460
And right now, AI is so early, people don't know what they're adopting and they have to know before they can make a choice.

31
00:02:30,260 --> 00:02:33,940
So the pollers here are God and the devil.

32
00:02:34,360 --> 00:02:38,400
AI is the end all, the be all on the positive aspect.

33
00:02:38,840 --> 00:02:42,620
Or, you know, AI is here to take our lives.

34
00:02:42,620 --> 00:02:47,880
I heard quotes that there's a 10% possibility that AI is going to outlive us.

35
00:02:48,000 --> 00:02:50,320
So I thought that was interesting.

36
00:02:51,120 --> 00:02:53,220
But I think we're here.

37
00:02:53,500 --> 00:02:55,680
I think there's a third path.

38
00:02:55,860 --> 00:02:59,120
And that Yogi Berra quote comes to mind, if you know what I'm talking about.

39
00:02:59,120 --> 00:03:00,940
When you come to a fork in the road, what do you do?

40
00:03:02,160 --> 00:03:03,320
Well, you pick it up.

41
00:03:03,500 --> 00:03:04,820
So here we are.

42
00:03:04,900 --> 00:03:07,160
We don't need to choose all or nothing.

43
00:03:07,480 --> 00:03:10,340
We can keep the baby and throw out the bathwater separately.

44
00:03:11,140 --> 00:03:13,680
And human-first AI does this.

45
00:03:14,420 --> 00:03:22,100
If not for humans, the business success that we have today across all corporate enterprise wouldn't have happened.

46
00:03:22,300 --> 00:03:25,440
It's all been labor and mind of humans.

47
00:03:25,440 --> 00:03:32,020
So why is the conversation focused on replacing humans if humans is what got us here?

48
00:03:32,480 --> 00:03:34,180
So does AI do good stuff?

49
00:03:34,480 --> 00:03:36,760
Oh, my God, absolutely it does good stuff.

50
00:03:36,760 --> 00:03:42,220
But there's six or seven items that humans will always do better than AI.

51
00:03:42,980 --> 00:03:51,840
So human-first AI is keeping the humans employed and giving the humans the ability to control what AI does.

52
00:03:51,980 --> 00:04:00,500
So Practical AI, my company, our philosophy is use AI to think with the human, not for the human.

53
00:04:00,860 --> 00:04:02,060
That's a big difference.

54
00:04:02,060 --> 00:04:08,220
and it even goes beyond human in the loop because human in the loop, that's kind of a catchphrase

55
00:04:08,220 --> 00:04:12,980
that people are using, is like, hey, let's put a human in here. We're going to check on it every

56
00:04:12,980 --> 00:04:18,820
hour. Oh, you know, I'm tired. Let's check on it once a day. You know what? Nobody's died yet. Let's

57
00:04:18,820 --> 00:04:22,980
check on it once a week, once a month. It gets to be the point. Human in the loop is just a laugh.

58
00:04:22,980 --> 00:04:29,540
It really doesn't mean anything. But if we use AI to think with the human, then we can empower the

59
00:04:29,540 --> 00:04:34,780
humans to do what they've always done well, and that's to make choices outside the box.

60
00:04:35,040 --> 00:04:35,980
AI is great.

61
00:04:36,140 --> 00:04:37,120
AI is technology.

62
00:04:37,120 --> 00:04:40,320
So AI will only do what we program it to do.

63
00:04:41,020 --> 00:04:41,880
And what do we call it?

64
00:04:41,940 --> 00:04:47,620
You know, that box on our desk or corporate, you know, the box in the database center.

65
00:04:48,380 --> 00:04:51,860
And AI can only do what we program it to.

66
00:04:52,060 --> 00:04:54,880
And by definition, that fits inside a box.

67
00:04:54,880 --> 00:05:05,040
So what happens when things go wrong in both the white-collar environment and the blue-collar environment is that humans have to take over.

68
00:05:05,180 --> 00:05:06,920
They have to step in no matter what.

69
00:05:07,320 --> 00:05:27,743
And if we get used to AI making all of our decisions then our businesses become brittle And so human AI is really empowering the human to be better than they are so they can continue to be the integral cog in the business

70
00:05:29,103 --> 00:05:32,383
Yeah, it's always been a striking thing for me, right?

71
00:05:32,383 --> 00:05:53,103
But to see companies looking at replacing humans with AI, and while I can understand as a short-sighted approach where that might be somewhat desirable for a company whose obvious mission is to make dollars for the shareholders, etc.,

72
00:05:53,103 --> 00:06:13,463
The fact of the matter is our economy relies on these humans and more unemployed humans means less humans to spend on the thing that you're producing, meaning you're literally destroying your own marketplace by replacing the humans with AI, right?

73
00:06:13,563 --> 00:06:15,323
It's always been a thought for me.

74
00:06:17,583 --> 00:06:20,623
Well, humans will get replaced through technology.

75
00:06:20,623 --> 00:06:23,063
It's happened in the past and will continue to happen.

76
00:06:23,103 --> 00:06:26,843
but it doesn't mean they're going to be displaced entirely.

77
00:06:26,843 --> 00:06:31,243
If we think back to AT&T days when the operators were replaced,

78
00:06:31,383 --> 00:06:32,503
you know, they used to do the,

79
00:06:33,503 --> 00:06:37,083
I think we're all old enough in the room to remember the switchboard operators

80
00:06:37,083 --> 00:06:38,483
in comedy skits.

81
00:06:38,483 --> 00:06:41,703
So we're aware of them, whether we used them or not,

82
00:06:41,803 --> 00:06:45,903
but switchboard operators that literally plugged in wires to make calls

83
00:06:45,903 --> 00:06:47,623
were replaced with computers.

84
00:06:47,623 --> 00:07:04,803
And I think it was NPR that did an episode or a two-part episode saying that what ended up is that those roles got displaced, but the workers ended up in jobs better than they had.

85
00:07:04,803 --> 00:07:18,003
And then if we go even further back, guys, to the agricultural revolution, let's say, I don't know, as close as 150 years ago, the number was that one farm would feed two families.

86
00:07:18,663 --> 00:07:30,683
And I don't know about you, brother, but I'm glad I'm not out there detasseling corn and rowing and hoeing behind a donkey because I'm glad we advanced using some technology there.

87
00:07:30,683 --> 00:07:40,143
And what happened is that the masses ended up finding, I think the agriculture revolution really led to the industrial revolution.

88
00:07:40,383 --> 00:07:53,083
That's when the Fords, Henry Fords were out there that created assembly lines and brought us humans that hate the sun and the weather and the rain inside to do some work.

89
00:07:53,083 --> 00:08:01,723
And there was some adoption there as well, but ultimately, quality of life increased and we went on.

90
00:08:03,063 --> 00:08:05,883
Yeah, no, I definitely see that.

91
00:08:06,243 --> 00:08:08,923
I think there will be some replacement.

92
00:08:09,663 --> 00:08:18,523
My hope is that it's not as pervasive as is potential and possible.

93
00:08:18,523 --> 00:08:37,623
Right. I mean, you have companies coming out that are replacing entire, you know, shops with nothing but robotics, ML, machine language, machine learning and AI. Right. Literally no staff on site. Someone comes, fills the bins and that's it.

94
00:08:37,623 --> 00:08:58,083
And my fear with that is that we get to a situation where we're replacing all of these entry level positions, which are critical to our economy in that they are the training grounds for new people coming into a workforce.

95
00:08:58,083 --> 00:08:58,803
Right.

96
00:08:58,803 --> 00:09:07,663
very very few people start at a you know higher position at a corporation straight out of the

97
00:09:07,663 --> 00:09:15,363
gate they cut their teeth in retail you know making coffee doing things like this and what

98
00:09:15,363 --> 00:09:22,123
happens if that all goes away you know what's your experience i've heard tell that there's a

99
00:09:22,123 --> 00:09:25,723
like drive-thrus that are using that like mcdonald's and stuff i don't go there so i don't

100
00:09:25,723 --> 00:09:31,263
know, but like my in-laws told me, they hate it because I guess the thing got their order

101
00:09:31,263 --> 00:09:32,083
wrong a bunch of times.

102
00:09:32,203 --> 00:09:32,903
So I don't know.

103
00:09:33,423 --> 00:09:34,483
That's kind of taken away.

104
00:09:34,623 --> 00:09:36,103
I mean, my first job was at McDonald's.

105
00:09:36,963 --> 00:09:39,663
So that's kind of taken away from younger people, I think.

106
00:09:40,103 --> 00:09:42,023
I don't know what other instances there are.

107
00:09:42,303 --> 00:09:44,463
Well, let me clarify my stake in being here.

108
00:09:44,543 --> 00:09:46,863
I am not here to defend AI in any way.

109
00:09:46,863 --> 00:09:53,743
AI is a tool and it can be used good or bad, just like every tool that's out there.

110
00:09:53,743 --> 00:09:59,163
The headlines read that AI is going to cause our brains to stop functioning.

111
00:09:59,963 --> 00:10:08,323
And I remember that I had this cool, in seventh grade, I had this cool Casio calculator watch.

112
00:10:09,783 --> 00:10:18,403
And people today, you know, current generations, they don't realize that the headlines back in the, what was it, late 70s, early 80s,

113
00:10:18,403 --> 00:10:37,786
when these calculators came around that we could fit in our pocket or on our wrist that the headlines were hey we can have students using calculators Computers are going to think form and they going to become stupid But what really happened is it elevated the individual So yes let teach

114
00:10:37,786 --> 00:10:42,406
basic math is what we learned. This is the takeaway, what, now 40 years later or so.

115
00:10:43,686 --> 00:10:48,366
And the discussion went on for about 20 years. And that's kind of like a psychological adaptation.

116
00:10:48,366 --> 00:10:51,786
It takes about one generation for the world to adapt.

117
00:10:52,506 --> 00:11:00,006
And what we found out is that, hey, let's teach children how to do the basic math so they understand what math is.

118
00:11:00,326 --> 00:11:11,926
But then let's let them use calculators so they can reduce that load on their mind so they can have more free time to spend learning higher level math.

119
00:11:12,346 --> 00:11:13,346
And that's what happened.

120
00:11:13,346 --> 00:11:37,366
We came away and the entire world was, I don't know, enlightened might be too strong a word, but the entire world literally got smarter because now all the time that they would spend doing manual addition, subtraction, multiplication, division was spent at a higher level math to understand how to solve problems, not just push numbers together.

121
00:11:37,366 --> 00:11:55,726
Yeah, no, absolutely. And I see, you know, I mean, I think AI is an immensely powerful augmentation tool. I do obviously see areas where it will replace workers. And I think there's a lot of room for re-skilling.

122
00:11:55,726 --> 00:12:01,706
um and and i think you know robotics in general has done that and shown us this as well you look

123
00:12:01,706 --> 00:12:07,946
at manufacturing floors you you pointed out forward right uh the auto manufacturing industry

124
00:12:07,946 --> 00:12:13,826
was one of the very first places that started replacing workers with automated assembly lines

125
00:12:13,826 --> 00:12:21,126
but it wasn't a wholesale replacement and you know ostensibly was good for the people that were

126
00:12:21,126 --> 00:12:29,386
displaced a little further down the line, right? Well, and understand too that any technology that

127
00:12:29,386 --> 00:12:33,866
comes out, every company is trying to take advantage of it. They figure it's the genie

128
00:12:33,866 --> 00:12:39,546
out of the bottle and let's just use it and apply it to everything. In my book, I talk about

129
00:12:39,546 --> 00:12:46,406
when electricity, not invented, but adopted, both residential and commercially, industrially,

130
00:12:46,406 --> 00:12:54,846
people were throwing electricity as a solution for everything and i kind of point out when a

131
00:12:54,846 --> 00:13:01,206
company says we're an ai company you might want to stay away from that company because that's like

132
00:13:01,206 --> 00:13:05,626
saying you know the kroger grocery store is an electric company because they use electricity

133
00:13:05,626 --> 00:13:13,726
to keep their refrigerators cold and that's what happened back oh gosh 130 120 30 years ago or so

134
00:13:13,726 --> 00:13:15,886
when electricity was being adopted.

135
00:13:17,166 --> 00:13:21,126
Some people, if you ask the random person

136
00:13:21,126 --> 00:13:22,546
walking down the street,

137
00:13:23,166 --> 00:13:25,206
said, when was the electric car invented?

138
00:13:25,686 --> 00:13:29,206
They might say like, I don't know, 2000, 2010, 1995.

139
00:13:29,806 --> 00:13:32,226
It was originally around 1900.

140
00:13:32,486 --> 00:13:34,426
It was when electricity was being adopted

141
00:13:34,426 --> 00:13:37,186
because they were using this new technology,

142
00:13:37,806 --> 00:13:41,566
industrialized, to try to solve for problems

143
00:13:41,566 --> 00:13:43,106
that really weren't even there.

144
00:13:43,106 --> 00:13:44,886
They're trying to paint with a large brush.

145
00:13:45,066 --> 00:13:46,406
And that's what's going on today.

146
00:13:46,966 --> 00:13:53,146
AI is being thrown at every problem before there even is a problem.

147
00:13:53,706 --> 00:14:00,486
And just to put hopefully the majority, if not all, of your audience at rest here that's concerned about losing their job,

148
00:14:01,586 --> 00:14:12,726
MIT results indicate that 95% of all corporations that have adopted AI, 95% have been failures.

149
00:14:12,726 --> 00:14:15,306
There's only been 5% that have an ROI.

150
00:14:16,086 --> 00:14:17,726
And that's because they're throwing everything.

151
00:14:18,806 --> 00:14:21,706
Throwing that spaghetti against the wall, they're just trying to see what sticks.

152
00:14:22,466 --> 00:14:27,906
And in the end, the human factor is always going to be the human factor.

153
00:14:28,106 --> 00:14:33,646
And we need the human factor because we do adopt and because humans do adapt.

154
00:14:33,746 --> 00:14:35,286
And we do think outside the box.

155
00:14:35,386 --> 00:14:38,486
These are things that computers aren't going to be able to do.

156
00:14:38,486 --> 00:14:48,486
So it's going to allow us as the individual, whether it's at home or at the shop or the office, it's going to allow us to really get a boost.

157
00:14:49,406 --> 00:14:53,446
We're going to have to start thinking like managers, project managers, supervisors.

158
00:14:54,226 --> 00:14:58,806
And we're going to be doing the same work, but we're going to be using new tools.

159
00:14:59,146 --> 00:15:00,706
And we get new tools all the time.

160
00:15:01,106 --> 00:15:03,286
And that's all AI truly is.

161
00:15:03,306 --> 00:15:04,366
It's just a new tool.

162
00:15:04,366 --> 00:15:08,806
And I encourage everybody just to play with it like it's Play-Doh.

163
00:15:08,986 --> 00:15:11,606
Just feel it out and see what it is.

164
00:15:11,706 --> 00:15:17,106
But don't be worried about the wholesale economy losing their job because it's not going to happen.

165
00:15:17,206 --> 00:15:19,806
There's going to be those that are directly affected.

166
00:15:20,606 --> 00:15:23,286
But you mentioned Havoc reskilling.

167
00:15:23,746 --> 00:15:42,969
Reskilling is so much easier to do today than it was 50 years ago that the opportunities are really endless out there for humans We never gonna be put out of work permanently Yeah I don know that I hold quite the same level of

168
00:15:47,048 --> 00:15:48,969
Optimism?

169
00:15:50,849 --> 00:15:56,589
Yeah, it's concerning for me. And here's why, right? I mean, gunpowder was created,

170
00:15:56,589 --> 00:16:02,449
right? Well, before gunpowder, black powder was created and it was used to create these

171
00:16:02,449 --> 00:16:08,109
beautiful displays with fireworks. And then we decided we could put it in, you know, tube and

172
00:16:08,109 --> 00:16:14,048
kill people further away. And then dynamite was created. No, we can clear these roads and we can

173
00:16:14,048 --> 00:16:20,709
do mining operations and make tunnels. So railways don't have to go way around these mountains and

174
00:16:20,709 --> 00:16:26,109
all this. Oh, but we can put them in bombs and drop them on people. And the nuclear, you know,

175
00:16:26,109 --> 00:16:33,629
nuclear development came about and it was like oh free energy i mean literally energy for everyone

176
00:16:33,629 --> 00:16:39,909
on the planet yeah let's put in a bomb and kill a bunch of people you know what is any

177
00:16:39,909 --> 00:16:47,369
weaponize it come on i mean that's you know where it all ends up eventually yeah let's hope that it

178
00:16:47,369 --> 00:16:53,909
it you know some of that is is tempered i i think it's too late for that well you know with it

179
00:16:53,909 --> 00:16:55,969
hope is contagious.

180
00:16:56,149 --> 00:16:56,669
Let's keep it.

181
00:16:56,729 --> 00:16:58,149
We're going to have the dogs with the guns,

182
00:16:58,269 --> 00:16:59,369
robot dogs with the guns.

183
00:16:59,489 --> 00:17:00,569
We already do.

184
00:17:00,769 --> 00:17:01,029
Pretty soon.

185
00:17:01,309 --> 00:17:01,869
I know.

186
00:17:02,469 --> 00:17:04,429
You can have one right in your neighborhood.

187
00:17:04,909 --> 00:17:05,869
Watch where you go.

188
00:17:05,929 --> 00:17:07,109
As long as I can have one,

189
00:17:07,149 --> 00:17:07,589
it's okay.

190
00:17:08,049 --> 00:17:09,109
If everyone has one,

191
00:17:09,129 --> 00:17:09,809
it'll be okay.

192
00:17:10,909 --> 00:17:12,309
On an unrelated note,

193
00:17:12,488 --> 00:17:13,208
I don't know if you've seen,

194
00:17:13,289 --> 00:17:16,569
I guess there's a lot of people cut down these flock cameras with the

195
00:17:16,569 --> 00:17:17,809
hacks.

196
00:17:17,988 --> 00:17:18,829
Have you seen that?

197
00:17:20,149 --> 00:17:20,909
I have.

198
00:17:20,909 --> 00:17:22,849
I saw there's a,

199
00:17:22,849 --> 00:17:24,629
a hack that shows

200
00:17:24,629 --> 00:17:26,129
what was I

201
00:17:26,129 --> 00:17:27,909
I don't know

202
00:17:27,909 --> 00:17:28,869
one of my doom scrolling

203
00:17:28,869 --> 00:17:29,829
videos

204
00:17:29,829 --> 00:17:31,129
it was like Veritasium

205
00:17:31,129 --> 00:17:31,728
or something

206
00:17:31,728 --> 00:17:32,649
they were talking about

207
00:17:32,649 --> 00:17:33,029
those

208
00:17:33,029 --> 00:17:34,369
those flock cameras

209
00:17:34,369 --> 00:17:35,069
that are all over

210
00:17:35,069 --> 00:17:36,649
United States

211
00:17:36,649 --> 00:17:36,929
you know

212
00:17:36,929 --> 00:17:37,669
we're United States

213
00:17:37,669 --> 00:17:37,728
everywhere

214
00:17:37,728 --> 00:17:38,448
but

215
00:17:38,448 --> 00:17:40,889
they said except for this one town

216
00:17:40,889 --> 00:17:41,849
and this guy

217
00:17:41,849 --> 00:17:42,728
he did his

218
00:17:42,728 --> 00:17:43,649
video

219
00:17:43,649 --> 00:17:44,748
and

220
00:17:44,748 --> 00:17:45,688
he said hey

221
00:17:45,688 --> 00:17:47,248
I actually ended up

222
00:17:47,248 --> 00:17:48,369
connecting to those cameras

223
00:17:48,369 --> 00:17:49,609
and I shared them out

224
00:17:49,609 --> 00:17:50,129
to the public

225
00:17:50,129 --> 00:17:51,049
and they got mad

226
00:17:51,049 --> 00:17:52,089
yeah

227
00:17:52,089 --> 00:17:52,829
I bet

228
00:17:52,829 --> 00:18:11,569
Yeah, they actually did a really poor job of securing the systems. You actually can get into them and just go peruse all of the artifacts in there. And that put out a number of people have been doing various things.

229
00:18:11,569 --> 00:18:23,129
I obviously don't want to encourage anyone to vandalize things, and I would certainly never tell someone that a 1,000 nanometer green laser pointed at a flock camera lens would damage it.

230
00:18:23,208 --> 00:18:24,889
So don't do those types of things.

231
00:18:25,448 --> 00:18:29,109
I heard a rumor there's like 200 bucks in gold and copper inside all those things.

232
00:18:29,309 --> 00:18:30,369
I don't know if that's true or not.

233
00:18:30,649 --> 00:18:31,789
Hey, we'd have to look.

234
00:18:32,228 --> 00:18:36,208
We really don't know until we test out that theory ourselves, right?

235
00:18:36,589 --> 00:18:37,129
That's right.

236
00:18:37,508 --> 00:18:39,069
Somebody's testing it out, I think.

237
00:18:39,169 --> 00:18:40,349
Not me, but somebody.

238
00:18:40,349 --> 00:18:47,508
so part of your history you you actually worked for amazon i i'm not sure if it was a contractor

239
00:18:47,508 --> 00:18:53,929
role or you were actually working in the employee of of amazon but you created uh apparently they're

240
00:18:53,929 --> 00:19:00,849
the best the highest rated uh employee training program in their history uh what how did that go

241
00:19:00,849 --> 00:19:06,948
well it sounds like it went pretty good uh the way you tell it it sounds like but i mean okay so

242
00:19:06,948 --> 00:19:13,609
sorry not how did that go but you know what what made it so successful what was it that you did

243
00:19:13,609 --> 00:19:18,669
with the training program that made it something more than here's your one hour mandatory training

244
00:19:18,669 --> 00:19:24,669
you have to sit and slog through well it uh it turns out and i did work for amazon directly

245
00:19:24,669 --> 00:19:29,549
and and i was on the reliability team which is a cool way of saying you know maintenance and

246
00:19:29,549 --> 00:19:35,669
engineering you know the ones that keep the uh um gosh the robots and the conveyors running that

247
00:19:35,669 --> 00:19:43,409
sort of thing. And when I got there, I was in the reliability org, which is a smaller org than all

248
00:19:43,409 --> 00:19:49,029
of Amazon. Amazon employs, I know, let's say a million and a half globally. And I think my group

249
00:19:49,029 --> 00:19:54,708
at the time, at least for North America, was still around like 45,000 people, which, you know,

250
00:19:54,708 --> 00:20:00,748
it's good size, even if it was a company, but it was an org within Amazon. And the hiring,

251
00:20:00,748 --> 00:20:05,909
because it was skilled, you know, skilled, whether it be skilled trades, you know, men and women that

252
00:20:05,909 --> 00:20:11,968
work with their hands or engineering that's, you know, trained to do specific things, you know,

253
00:20:11,988 --> 00:20:18,529
on schematics, computers, they would hire in and just expect them to work. So, Amazon would hire

254
00:20:18,529 --> 00:20:23,629
the people. We kind of had a joke about it. And the joke was, hey, you know, so-and-so's in the

255
00:20:23,629 --> 00:20:28,428
deep end of the pool, you know, do you think he'll make it out? Well, hey, he's looking good.

256
00:20:28,428 --> 00:20:32,529
Looking good. Oh, nope, he went under. He's back up. He's back up.

257
00:20:32,529 --> 00:20:37,008
Nope, went down, down for the count. Well, we got to get another one.

258
00:20:37,008 --> 00:20:44,708
And the numbers tell the truth here. I think the attrition rate was around 82 or 83%,

259
00:20:44,708 --> 00:20:49,968
which means for every 100 people that Amazon hired in that organization,

260
00:20:50,011 --> 00:20:56,711
internal that 83 would just leave on their own accord. They just didn't make it. And that becomes

261
00:20:56,711 --> 00:21:04,211
expensive. So it's unfortunate that the role of education is put at the bottom of the rung and

262
00:21:04,211 --> 00:21:12,131
that lever isn't pulled until there's a business need, strong enough business need. So they asked

263
00:21:12,131 --> 00:21:18,751
me and they didn't tell me I was the last person they asked, but I think it was December 14th when

264
00:21:18,751 --> 00:21:26,911
I got a call and they said, hey, I think it was what, 2020, 2021 in December when I got the call.

265
00:21:27,271 --> 00:21:34,651
And this was during this COVID pandemic that a lot of people were talking about. And Amazon had

266
00:21:34,651 --> 00:21:43,571
committed in North America, we had 500 warehouses or what we call fulfillment centers. And in one

267
00:21:43,571 --> 00:21:50,851
year. It took 25 years to get to 500. And in one year, their plan was to double that to 1,000.

268
00:21:51,611 --> 00:22:00,271
And so I got this call on December 14th and says, hey, Derek, we're doing this thing. And they gave

269
00:22:00,271 --> 00:22:05,251
me the details and they thought, do you think we should train them, these new hires? Because if

270
00:22:05,251 --> 00:22:10,531
we're doubling the amount of centers, fulfillment centers, we're going to have to double the

271
00:22:10,531 --> 00:22:17,971
employees to go, let's say from, I don't know, 45,000 and 90,000 or some rough doubling of

272
00:22:17,971 --> 00:22:22,371
employment. I said, yeah, that'd be a good idea. And it's a phone call. So I'm wondering if it's

273
00:22:22,371 --> 00:22:28,891
tongue in cheek that the question came about. They said, great, do it. And I thought, yeah,

274
00:22:28,891 --> 00:22:35,891
just I had that. I said, okay. I said, how about some particulars here? And I said, when would you

275
00:22:35,891 --> 00:22:41,731
like to see this program up and running? They said, oh, I don't know, January 1st. And, you know,

276
00:22:41,731 --> 00:22:48,851
it's corporate America and it's December 14th and corporate America takes the last two weeks of

277
00:22:48,851 --> 00:22:56,371
December off, at least in North America we do. And I thought, okay. I got that extended out to,

278
00:22:56,471 --> 00:23:01,471
I think like January 18th or something. So I got me two or three more weeks to get started.

279
00:23:01,471 --> 00:23:06,671
but that was the first class that came in.

280
00:23:07,611 --> 00:23:10,671
And so, frankly, we started.

281
00:23:10,851 --> 00:23:15,951
I had some connections, some people that helped launch some centers,

282
00:23:16,331 --> 00:23:18,391
and I just said, hey, meet me here.

283
00:23:18,991 --> 00:23:21,971
And actually, one of the guys lives in Florida.

284
00:23:22,271 --> 00:23:24,291
He said, hey, you come to Tampa, Florida.

285
00:23:24,511 --> 00:23:25,931
I'll be down there and help out.

286
00:23:26,011 --> 00:23:27,191
So I said, heck yeah.

287
00:23:27,311 --> 00:23:29,031
So we just brought everybody to Tampa, Florida.

288
00:23:29,211 --> 00:23:31,391
This is when all the restaurants are closed.

289
00:23:31,471 --> 00:23:33,031
Hotels are nearly closed.

290
00:23:33,491 --> 00:23:36,551
And we rented out a restaurant in the hotel we were at.

291
00:23:36,911 --> 00:23:38,471
And we started our first class.

292
00:23:38,551 --> 00:23:42,831
And our first class was, I think, 24 people.

293
00:23:43,431 --> 00:23:47,131
Eventually, the next class, it got up to 40 and 40 and 40 consistently.

294
00:23:47,451 --> 00:23:48,831
But that class started at 24.

295
00:23:49,611 --> 00:23:51,151
And it was two weeks long.

296
00:23:51,591 --> 00:23:53,891
And we said, all right, what should we talk about?

297
00:23:53,951 --> 00:23:55,711
And we literally put it to a vote.

298
00:23:55,751 --> 00:23:56,671
What don't you know?

299
00:23:57,211 --> 00:23:58,351
And we talked about it.

300
00:23:58,351 --> 00:24:00,891
And then that evening, there's about four or five of us.

301
00:24:00,891 --> 00:24:07,111
So we'd sit around the hotel lobby or we'd talk over dinner and say, hey, what did we talk about today?

302
00:24:07,111 --> 00:24:08,631
All right, let's put that down.

303
00:24:09,131 --> 00:24:16,471
Because our presentation slides, our PowerPoint was literally just welcome to class.

304
00:24:17,051 --> 00:24:28,671
And so each class every day and every week and then class after class, the secret was listening to the people that are actually doing the work and finding answers.

305
00:24:28,671 --> 00:24:37,671
and it sounds, I don't know, maybe a little too easy, but a lot of people overlook easy

306
00:24:37,671 --> 00:24:43,311
and we just made it happen. We listened to the people, we filled it in, and I believe in continuous

307
00:24:43,311 --> 00:24:49,651
improvement and that's what we did. The class was continually improved from class to class to class

308
00:24:49,651 --> 00:24:54,531
to class. I made sure that everybody had connections with everybody else from class.

309
00:24:54,531 --> 00:25:06,971
We were bringing in people that didn't normally work together because we might have managers from different locations that, you know, each one manager per shift or one manager per location.

310
00:25:06,971 --> 00:25:10,091
So they didn't have that camaraderie out there.

311
00:25:10,211 --> 00:25:14,491
So they were literally, you know, like I described earlier, in the deep end by themselves.

312
00:25:14,491 --> 00:25:21,771
so the moral of the story is you give humans the right information at the right time they can do

313
00:25:21,771 --> 00:25:30,871
some magnificent magnificent uh successful uh projects that's what we did wow that truly is an

314
00:25:30,871 --> 00:25:40,411
it's a wonderful take there really honestly you know you i can't even count the number of

315
00:25:40,411 --> 00:25:47,151
trainings that I've gone to. And I don't think I've ever been engaged in the beginning of a

316
00:25:47,151 --> 00:25:52,811
training to ask, well, what do you know? What don't you know? What do you need help with?

317
00:25:52,811 --> 00:26:06,988
Right Uh which that right there seems like just such a practical thing to do but I never seen it Never once Well it worked

318
00:26:07,548 --> 00:26:08,048
It worked.

319
00:26:08,308 --> 00:26:11,108
I did not pretend to have all the answers.

320
00:26:11,508 --> 00:26:12,868
We all stayed vulnerable.

321
00:26:14,028 --> 00:26:18,328
And the first day of class was actually never about Amazon stuff.

322
00:26:18,448 --> 00:26:21,968
It was always about get to know yourself and get to know other people.

323
00:26:21,968 --> 00:26:30,248
And anybody that's had to take any part of a leadership training, that's effectively what all the training revolves around.

324
00:26:30,388 --> 00:26:31,728
Hey, everybody's different.

325
00:26:32,008 --> 00:26:39,028
So let's just, you know, put up, you know, three legs to a table or a chair because it takes more than one to hold it up.

326
00:26:40,428 --> 00:26:41,408
That's wonderful.

327
00:26:41,408 --> 00:26:47,048
What are some of what are some of the your favorite lessons from that period?

328
00:26:47,048 --> 00:26:57,048
And what would be like one big takeaway you got from the experience of developing this, implementing it, learning through it, etc.?

329
00:26:58,788 --> 00:27:07,108
Well, for myself, I think my big takeaway was everybody can contribute.

330
00:27:08,288 --> 00:27:11,808
And we've had individuals that spent a long time with Amazon.

331
00:27:11,808 --> 00:27:18,248
We had people that came in from different industries come in to Amazon as brand new employees.

332
00:27:19,068 --> 00:27:20,668
But the goal was the same.

333
00:27:20,788 --> 00:27:27,188
It was to come up with a system that allowed each person to maximize their potential.

334
00:27:27,988 --> 00:27:32,728
So there was no one way is the right way.

335
00:27:32,828 --> 00:27:34,448
There was no my way or the highway.

336
00:27:35,128 --> 00:27:38,128
It was, hey, let's flex, let's listen in.

337
00:27:38,128 --> 00:27:42,028
And that really started on day one of the programs because we would meet in person.

338
00:27:42,828 --> 00:27:50,948
And we would have, I would get there early and we had organized, gosh, hotel lobbies, restaurants, etc.

339
00:27:51,408 --> 00:27:55,848
You know, I think I would have tables and chairs that were on seven foot centers.

340
00:27:55,848 --> 00:27:58,268
So everybody had their personal space.

341
00:27:58,488 --> 00:28:01,348
I think six foot was a number everybody was throwing around.

342
00:28:02,128 --> 00:28:07,728
We were all wearing masks and eventually, you know, eye shields or goggles.

343
00:28:07,728 --> 00:28:12,928
And that made communication much more difficult, at least in the early years.

344
00:28:13,588 --> 00:28:17,848
So maybe that's what forced us to be patient.

345
00:28:18,148 --> 00:28:19,108
I mean, I really don't know.

346
00:28:19,188 --> 00:28:21,868
But the takeaway is the same either way.

347
00:28:22,848 --> 00:28:27,348
Everybody's voice counts and makes sure that everybody gets a chance to speak up.

348
00:28:29,408 --> 00:28:30,548
I'm a little bit curious.

349
00:28:30,668 --> 00:28:33,808
I don't know if you might mind sharing a little bit behind the curtain.

350
00:28:33,808 --> 00:28:44,928
And when my wife is looking at pajama pants at like 10 o'clock at night and she hits the click to buy it now and they show up in the morning at like 6 a.m., how does that all work?

351
00:28:45,288 --> 00:28:48,588
Like when she goes to return it, where does that stuff go?

352
00:28:48,948 --> 00:28:50,748
Well, I'm just from personal experience.

353
00:28:50,748 --> 00:28:51,208
Yeah, absolutely.

354
00:28:51,308 --> 00:28:51,608
Absolutely.

355
00:28:51,608 --> 00:28:56,768
So it's all logistics, which is a fancy way of saying truck locations.

356
00:28:56,768 --> 00:29:05,868
And so when they say, or when I talked about earlier, the fulfillment centers, it's essentially warehouses.

357
00:29:06,988 --> 00:29:19,408
And they go, her order goes in, it gets picked, and it gets routed through, most of the time, the least busy location that is within, you know, whether one day's drive, two days drive, etc.

358
00:29:19,408 --> 00:29:26,088
and then once it gets routed there there are pickers on the floor humans that go through and

359
00:29:26,088 --> 00:29:31,708
they go to a certain aisle certain shelf and they they see this on their scanners the scanners that

360
00:29:31,708 --> 00:29:38,228
like like they're using clothing stores to scan the tag and it's got the little computer screen

361
00:29:38,228 --> 00:29:44,468
on it just like three inches by three inches they go through and they pull whatever was ordered those

362
00:29:44,468 --> 00:29:47,788
pajamas that you like, you know, with the cool colors.

363
00:29:48,428 --> 00:29:50,328
So it's actually humans pulling it out of the bins.

364
00:29:50,528 --> 00:29:52,188
Humans pulling it out of the bins.

365
00:29:52,408 --> 00:29:55,288
So it's kind of like how were the pyramids built, you know, with a lot of work.

366
00:29:55,788 --> 00:29:59,388
And that's what Amazon does at the base level.

367
00:29:59,968 --> 00:30:06,168
Now, once it gets pulled, it goes into a specific line.

368
00:30:06,248 --> 00:30:06,948
It's like a checkout.

369
00:30:07,128 --> 00:30:11,128
They just dump entire carts because they actually pull for multiple customers.

370
00:30:11,128 --> 00:30:22,048
So one picker, and that's an official title, one picker picks orders for five people, 10 people, 20 people, whatever they can fit the cart.

371
00:30:22,288 --> 00:30:24,468
They take it to a specific line.

372
00:30:24,608 --> 00:30:38,008
And now that it's all QR and barcoded, that line, when it comes through, they'll scan that barcode and they say, oh, this is for Dallas, Texas, and it's this address for this person.

373
00:30:38,008 --> 00:30:39,468
So they just grab it.

374
00:30:39,568 --> 00:30:40,228
They box it.

375
00:30:40,228 --> 00:30:49,168
The label prints automatically, which printing the labels and stamping the labels on the box is one of the most difficult and frustrating points of this entire system.

376
00:30:49,488 --> 00:30:51,468
But it's packaged by hand.

377
00:30:52,228 --> 00:30:58,528
It either gets taped up by hand, which is rare, or it gets folded and then a box taper tapes it automatically.

378
00:30:59,128 --> 00:31:06,148
But once it's taped, it goes down conveyance that scans the barcode and routes it to the right truck.

379
00:31:06,148 --> 00:31:22,524
So if you ever been at a warehouse and you see all these truck doors each one of those truck doors is going to a different location And it most easily understood the logistics is most easily understood if we say it like going to the airport

380
00:31:23,164 --> 00:31:30,384
Like if I'm going to fly to see my daughter in Seattle, hopefully I can take a direct route.

381
00:31:30,584 --> 00:31:35,704
But if I don't have a direct route, I might be routed through Dallas or Phoenix and it's a second

382
00:31:35,704 --> 00:31:40,104
route. So that's what happens. The box gets transferred from that immediate warehouse

383
00:31:40,104 --> 00:31:48,884
to a truck going in your house's direction. And then once it gets there, it gets dropped off at a

384
00:31:48,884 --> 00:31:57,484
delivery center, which is a smaller center with not much mechanical. It's really all, again,

385
00:31:57,484 --> 00:32:06,124
individuals moving the boxes. And then it's very much like a FedEx or a UPS, which they can also

386
00:32:06,124 --> 00:32:12,904
deliver. It can be barcoded for UPS and UPS will deliver the package. Or Amazon more and more

387
00:32:12,904 --> 00:32:19,084
delivers their own trucks. You see the cool smiley face buildings are on the side of the trucks.

388
00:32:19,984 --> 00:32:25,504
But it's all the same. They just grab the box and they deliver it. And the secret is having

389
00:32:25,504 --> 00:32:31,104
warehousing close enough to your location that they can make it available and get it to you

390
00:32:31,104 --> 00:32:39,104
within whatever time limit that is i see yeah it's pretty amazing the logistics that they

391
00:32:39,104 --> 00:32:46,024
they put into this i i have to imagine it's it's not quite but approaching the level of like airline

392
00:32:46,024 --> 00:32:53,484
logistics right uh that that i think is probably a lot more challenging just because of the the

393
00:32:53,484 --> 00:33:02,024
strict timing on everything, but still very, very complex stuff. So you also created this

394
00:33:02,024 --> 00:33:11,244
thing called Pocket Mentor, which I guess was world's first AI-based mentor. What made you come

395
00:33:11,244 --> 00:33:18,444
up with that idea and what did you kind of want that to be used for? Thank you for asking the

396
00:33:18,444 --> 00:33:23,484
question because I think it's really important. I think it's an effective use of AI in that

397
00:33:23,484 --> 00:33:33,424
it amplifies the human. And for Amazon and other companies, I've done a lot of training.

398
00:33:33,964 --> 00:33:39,784
Ultimately, my feather in my cap was what I did at Amazon that last time. And my takeaway,

399
00:33:40,084 --> 00:33:44,644
again, is giving the right person the right information at the right time. So what Pocket

400
00:33:44,644 --> 00:33:53,744
mentor does um imagine if um oh my god how about how about this throwback let's let's set it aside

401
00:33:53,744 --> 00:34:01,624
for a moment do you remember traveling across the country before gps before google maps oh yeah

402
00:34:01,624 --> 00:34:08,704
road atlas yeah road atlas triple a you know trip books that sort of thing and map quest and what

403
00:34:08,704 --> 00:34:14,504
would happen if you had an accident ahead that blocked the route you would have to stop pull over

404
00:34:15,224 --> 00:34:18,104
And so now you're not moving toward your destination or your goal.

405
00:34:18,384 --> 00:34:23,124
You have to figure out on your own how to get where you need to get.

406
00:34:23,504 --> 00:34:28,824
Now, if you're delivering a package, whether it be for Amazon or UPS or for any shop out there,

407
00:34:29,364 --> 00:34:32,624
the customer doesn't care if you ran into traffic.

408
00:34:32,624 --> 00:34:35,844
They're going to go, oh, you know, we're running into traffic.

409
00:34:35,964 --> 00:34:36,984
Let's give it another day.

410
00:34:37,084 --> 00:34:37,864
We won't be mad.

411
00:34:37,984 --> 00:34:39,044
No, they want service.

412
00:34:39,044 --> 00:34:54,224
So GPS, whether your flavor is MapQuest or Google Maps or Apple Maps, it tells you, and it actually routes, it's proactive with the GPS.

413
00:34:54,564 --> 00:34:58,424
And it says, hey, that route is blocked.

414
00:34:58,624 --> 00:35:01,244
Here's two alternatives, and we can choose and go.

415
00:35:02,064 --> 00:35:03,764
That's really how PocketMentor works.

416
00:35:03,804 --> 00:35:05,444
That was part of the inspiration there.

417
00:35:05,444 --> 00:35:09,484
But let's say you're on the job, again, blue collar, white collar, doesn't matter.

418
00:35:10,064 --> 00:35:16,204
And you're on the job and most of the time you're either using your hands or your eyeballs or both.

419
00:35:16,964 --> 00:35:23,204
And if you had to stop and look something up, well, now you're not moving toward your destination.

420
00:35:23,444 --> 00:35:34,584
So what Pocket Mentor does, it's actually a voice media that allows you to just press a button on your headset or on your phone.

421
00:35:34,584 --> 00:35:37,324
And it's called Pocket Mentor because it fits on your phone.

422
00:35:37,444 --> 00:35:38,264
It's not an app.

423
00:35:38,684 --> 00:35:41,584
It can work through a webpage or it can work through a telephone call.

424
00:35:41,644 --> 00:35:45,004
So you don't even need Wi-Fi or internet, but it's your buddy.

425
00:35:45,104 --> 00:35:46,144
It's your safety net.

426
00:35:46,684 --> 00:35:52,324
And now I'm kind of flashing back to the 1970s and 80s where help, I've fallen and I can't get up.

427
00:35:52,324 --> 00:35:53,444
Remember when that started?

428
00:35:54,524 --> 00:35:56,144
Maybe I picked the wrong name.

429
00:35:56,304 --> 00:35:59,204
Maybe I should have went with that route, but essentially it's the same thing.

430
00:35:59,544 --> 00:36:02,244
You hit the button and you start talking to your mentor.

431
00:36:02,244 --> 00:36:10,224
And this mentor, it's AI handles the translation between the human and the data set, the knowledge base.

432
00:36:10,304 --> 00:36:13,224
And that knowledge base, it's not the World Wide Web.

433
00:36:13,324 --> 00:36:16,564
You're not going to get crazy answers that the headlines talk about.

434
00:36:16,744 --> 00:36:20,304
It's going to be programmed with only the data for your job role.

435
00:36:21,244 --> 00:36:37,580
And so you say hey help Where does this package go What is the three process to reset the computer or to reset the conveyor And it will guide you through it Things that you responsible as a human to

436
00:36:37,580 --> 00:36:44,080
do anyway. And sometimes we just have a hiccup. We have a brain fart. I mean, really, our brains

437
00:36:44,080 --> 00:36:50,560
weren't meant for long-term memory. So if we can fill in that long-term memory with computer

438
00:36:50,560 --> 00:36:55,720
support, well, then we can go about our day and be productive. And that's what Procket Mentor does.

439
00:36:55,820 --> 00:36:59,780
It fills that gap when you say, help, I've fallen and I can't get up.

440
00:37:01,540 --> 00:37:08,100
So it's specifically for vocational specific applications rather than general users or

441
00:37:08,100 --> 00:37:15,480
can general users make use of it as well? Well, it's because it's trained on a specific topic,

442
00:37:15,480 --> 00:37:20,480
then we just have to ingest that knowledge base, whatever it is,

443
00:37:20,520 --> 00:37:26,400
that cool technology, ingest, give it that data or database to run from.

444
00:37:26,620 --> 00:37:32,160
So if it's for Ford Motor, it's going to be for a specific job.

445
00:37:32,500 --> 00:37:39,360
If it's for Starbucks, how do they make that macchiato, et cetera, et cetera,

446
00:37:39,360 --> 00:37:41,800
thing again, then it'll walk you through it.

447
00:37:41,800 --> 00:37:48,800
Um, for home, it, uh, I really don't have a, uh, like a civilian use or a consumer use.

448
00:37:49,260 --> 00:37:54,960
Um, but, uh, it's really is just as easy as, as trading that box and then talking to it.

449
00:37:56,160 --> 00:37:59,500
Sounds like something that could be used pretty much in any field.

450
00:37:59,500 --> 00:38:03,860
I mean, I, you know, in my head, I, as you were going through the scenarios, I'm thinking

451
00:38:03,860 --> 00:38:06,980
of a framer, you know, building a home, right.

452
00:38:06,980 --> 00:38:12,200
and they come across a situation, they're like, oh, crud, what's the code on this?

453
00:38:12,820 --> 00:38:17,400
Pocket mentor. Hey, pocket mentor, what's code on this type of framing with this,

454
00:38:17,500 --> 00:38:22,540
you know, obstacle, et cetera. Oh, here's the code for that in your location. Boom. All right.

455
00:38:22,720 --> 00:38:28,640
Now you've got it. Exactly right. It's having the right information at the right time. And

456
00:38:28,640 --> 00:38:35,180
the neat thing about using AI to deliver that information is that YouTube is great, right?

457
00:38:35,180 --> 00:38:43,040
YouTube has taught a lot of people and given people the confidence to go out, change the brakes on their car if they've never done it before.

458
00:38:44,080 --> 00:38:48,120
You know, replace the belt on their clothes washer if they've never done it before.

459
00:38:48,940 --> 00:38:59,180
But what happens when you go into a process and whatever you're fixing or repairing or building brand new doesn't follow?

460
00:38:59,400 --> 00:39:01,320
Maybe it didn't read the manual that you're reading.

461
00:39:01,580 --> 00:39:02,380
So what happens?

462
00:39:02,380 --> 00:39:14,680
So AI can actually create on-the-fly instruction from the core instructions to help you through those special scenarios.

463
00:39:17,820 --> 00:39:30,140
In my head, I'm already thinking like auto mechanics, like a flow chart for diagnostics and stuff, or HVAC technician or any kind, even a lot of the blue-collar trades, that would be a pretty good application, I think.

464
00:39:30,140 --> 00:39:59,000
Yeah, absolutely. And my background, I actually started a blue collar. I don't think anybody's immune anymore that college doesn't have the payback that college once did. And we're seeing the boomer generation leave skilled trades positions, whatever the skilled trades are, whether they're building trades where you're doing contract work or you're in-house doing work.

465
00:40:00,140 --> 00:40:05,620
And we are millions shy of having enough people for the trade.

466
00:40:05,620 --> 00:40:15,360
So my initial impetus was really to how can I use AI in a way that there is zero possibility it can replace humans.

467
00:40:16,100 --> 00:40:22,600
And that would be to help trade those in the trades, to have that right information at the right time,

468
00:40:22,600 --> 00:40:30,000
because it's the training that is most costly.

469
00:40:30,140 --> 00:40:32,200
because when you bring an apprentice in,

470
00:40:32,580 --> 00:40:34,620
you essentially have to pay for the apprentice

471
00:40:34,620 --> 00:40:35,780
and they don't know anything.

472
00:40:36,020 --> 00:40:39,420
There's no ROI on a zero ROI, negative ROI.

473
00:40:40,080 --> 00:40:42,860
So if we can train people faster

474
00:40:42,860 --> 00:40:45,820
to do what they want to do and need to do,

475
00:40:45,820 --> 00:40:48,920
it makes the world a better place all over.

476
00:40:49,060 --> 00:40:50,620
It gets them on board faster.

477
00:40:50,780 --> 00:40:53,200
So it does reduce the cost for businesses

478
00:40:53,200 --> 00:40:54,940
to get their people trained,

479
00:40:55,060 --> 00:40:57,880
but at the same time, for the person themselves,

480
00:40:57,880 --> 00:41:03,260
There's surveys they give, I think, officially every 10 years or so, and they interview.

481
00:41:03,420 --> 00:41:05,180
They say, hey, where do you work?

482
00:41:05,680 --> 00:41:06,840
Why do you work there?

483
00:41:06,940 --> 00:41:08,020
What's the number one reason?

484
00:41:08,880 --> 00:41:15,480
And I know I'm 58, and I have never heard a year where money, I work here for the money,

485
00:41:15,820 --> 00:41:17,300
has been higher than three.

486
00:41:17,560 --> 00:41:20,420
It's usually about fourth on that list.

487
00:41:20,480 --> 00:41:24,300
And the number one is typically, I want to feel productive.

488
00:41:25,620 --> 00:41:26,740
I want to do something.

489
00:41:26,740 --> 00:41:39,960
I want my life to have meaning. And so if we can give that information to them that they need so they can perform, it makes them feel.

490
00:41:39,997 --> 00:41:46,477
more mentally healthy, and it's a benefit to the businesses that employ those people.

491
00:41:46,697 --> 00:41:49,737
So it truly is a win-win-win all the way around.

492
00:41:51,177 --> 00:41:53,797
Well, I will say there's one downside.

493
00:41:54,637 --> 00:41:54,757
Yeah.

494
00:41:54,757 --> 00:41:56,377
Definitely one downside.

495
00:41:56,577 --> 00:42:02,117
Having worked in the trades myself, if a person were equipped with PocketMentor,

496
00:42:02,117 --> 00:42:12,177
it would be very challenging for the customary expected and really just part of the initiation

497
00:42:12,177 --> 00:42:19,877
when you sent the new guy out to find the door stretcher i mean he's gonna go well you know

498
00:42:19,877 --> 00:42:25,197
there's no such thing well you're supposed to go out to the truck and be looking for it for an hour

499
00:42:25,197 --> 00:42:33,197
you know what i mean i'm just saying hey we can program a snipe hunt anytime and we could

500
00:42:33,197 --> 00:42:39,277
randomize it there you go you got you got to have the hazing for the new guys it's just part of the

501
00:42:39,277 --> 00:42:43,737
gig right hey you know go look on the shelf give me some blinker fluid it's over there by the oil

502
00:42:43,737 --> 00:42:48,717
yeah you're saying that in jest but i truly believe there's value to that it helps uh bonding

503
00:42:48,717 --> 00:42:53,797
helps camaraderie now i don't believe anybody should die from it but uh you know if you don't

504
00:42:53,797 --> 00:42:59,037
get those spark plugs changed in that diesel, then by the end of the day, then you might as

505
00:42:59,037 --> 00:43:07,597
well not come back tomorrow. Right? Yeah. Awesome. So your company, Practical AI,

506
00:43:08,537 --> 00:43:15,297
coming from where you were, what was the impetus for starting this company and

507
00:43:15,297 --> 00:43:22,837
what is it about? What is it that makes Practical AI different in a world full of AI stuff?

508
00:43:23,797 --> 00:43:47,777
I think practical AI is AI that you don't really see. It's AI that does something. It's not headlines. We're not looking to change the world. I mean, we are looking to change the world, but the only way we're going to do that is through solid practical and pragmatic use of whatever tools and technologies that come to us.

509
00:43:47,777 --> 00:43:57,537
So practical AI, I was very conscious, conscientious about choosing that name because there has to be practical application.

510
00:43:58,337 --> 00:44:04,717
And if you can't understand it, well, then you're probably trying to create a solution that doesn't need to exist.

511
00:44:04,717 --> 00:44:13,397
so when you when you are engaged with a customer implementing you know something

512
00:44:13,397 --> 00:44:21,397
what uh i guess what does that look like how do you uh everyone wants ai right i mean everyone

513
00:44:21,397 --> 00:44:26,817
wants ai and they wanted to solve all the things they wanted to do everything uh and i think this

514
00:44:26,817 --> 00:44:34,217
is where that you know 89 to 95 percent failure rate comes in is there's no there's no actual plan

515
00:44:34,217 --> 00:44:40,377
or idea what the vision is for the end goal, right?

516
00:44:40,917 --> 00:44:44,037
So how do you deal with that?

517
00:44:44,177 --> 00:44:47,917
A new customer comes in and they want the world

518
00:44:47,917 --> 00:44:53,017
and you have to kind of take them down a funnel.

519
00:44:53,017 --> 00:44:54,037
Temper their expectations.

520
00:44:54,037 --> 00:44:58,277
Yeah, take them down a funnel to find what is your vision,

521
00:44:58,437 --> 00:45:02,277
what is the ultimate goal of what you're trying to implement here.

522
00:45:02,277 --> 00:45:27,617
Well, that's that human factor. And the human factor, it drives it all. Technology cannot outpace the human factor, good or bad. So if a company has SOP, standard operating procedures, we can take that and I can have for them a running pocket mentor by the end of the day.

523
00:45:28,457 --> 00:45:32,237
But if we get to a company and we say, okay, what do you want to do?

524
00:45:32,337 --> 00:45:33,657
Well, we want to do this role.

525
00:45:33,977 --> 00:45:35,617
And you say, okay, what does that role do?

526
00:45:36,117 --> 00:45:36,917
Well, I don't know.

527
00:45:36,997 --> 00:45:38,857
Hey, Jim, what do we do here?

528
00:45:39,337 --> 00:45:40,957
If it comes down to one of those.

529
00:45:41,597 --> 00:45:50,357
So even though we're a technology company, a lot of that interaction is just standard boots on the ground.

530
00:45:50,737 --> 00:45:54,737
Let's find out what you actually do and how you do it.

531
00:45:54,737 --> 00:45:57,677
Because we need to have the answers.

532
00:45:58,457 --> 00:46:07,757
And the neat thing about it is that once those answers are documented, everybody has access to them across the realm.

533
00:46:08,137 --> 00:46:14,457
I like Ford's CEO because he is really a big sponsor of skilled trades.

534
00:46:14,657 --> 00:46:16,197
He says, hey, we need them.

535
00:46:16,317 --> 00:46:17,057
We've got to train them.

536
00:46:17,157 --> 00:46:17,917
We're short on them.

537
00:46:18,517 --> 00:46:22,777
And every time he talks about them, I read his article, listen to it and such.

538
00:46:22,777 --> 00:46:35,177
And I really think it is important that we get that right information to the right people at the right time so they can do the human things that make us all successful.

539
00:46:37,097 --> 00:46:38,657
Wonderful, wonderful.

540
00:46:39,857 --> 00:46:46,297
So you were diagnosed neurodivergent later in life.

541
00:46:46,297 --> 00:46:59,853
And I wondering how did that play into your founding of the company and Pocket Mentor and your book that it coming right

542
00:46:59,873 --> 00:47:01,113
It's not out quite yet?

543
00:47:01,553 --> 00:47:03,313
The audio book is out.

544
00:47:04,213 --> 00:47:06,693
The print version will be out later this summer.

545
00:47:07,173 --> 00:47:09,813
But yeah, the Human First AI is available now.

546
00:47:09,813 --> 00:47:15,333
And since you mentioned it, I'll give every listener that you have a free chapter.

547
00:47:15,333 --> 00:47:19,913
they can just go to my site and read the first chapter for free. They don't even need to pop in

548
00:47:19,913 --> 00:47:28,433
their email address. It's all easy. Wonderful. So how did that play into the founding of the

549
00:47:28,433 --> 00:47:36,593
company, the creation of these tools? And how has this helped you in your journey with your

550
00:47:36,593 --> 00:47:44,973
neurodivergent aspects of life? Well, the fact that I was diagnosed at age 50, I started thinking

551
00:47:44,973 --> 00:47:52,893
that maybe there's another category I belonged to back in my 40s, but still a good chunk of my

552
00:47:52,893 --> 00:48:02,533
productive adult life, I spent in the self-help aisle almost permanently reading, listening,

553
00:48:03,033 --> 00:48:09,453
watching. Why is this round peg not fitting in the round hole? Well, it turned out I wasn't a

554
00:48:09,453 --> 00:48:24,073
I was a square peg and I didn't realize it. So because I couldn't connect and my neurodivergence diagnosis at age 50 was autism, ADHD and dyslexia.

555
00:48:24,073 --> 00:48:37,513
So the dyslexia kind of gives me the reason why I could never finish a book in school for a book report, just because everything, you know, these letters keep jumping around.

556
00:48:37,893 --> 00:48:38,553
I don't know.

557
00:48:38,593 --> 00:48:39,933
I thought my eyes were just twitching.

558
00:48:40,033 --> 00:48:40,793
I really did.

559
00:48:40,793 --> 00:48:51,893
But when I was diagnosed, I did some actual crying, tears falling from my eyes for about two weeks.

560
00:48:52,153 --> 00:49:08,833
Just tears of pain, tears of joy, tears of relief, because now I had a name and a category, and I could seek information that helped other people apply it to myself, new information in a different way.

561
00:49:08,833 --> 00:49:12,113
and I was working at Amazon at this time.

562
00:49:12,213 --> 00:49:15,673
It was early in my last Amazon tenure

563
00:49:15,673 --> 00:49:17,893
and I embraced it.

564
00:49:18,253 --> 00:49:21,253
I was at the point where I'm just exposing myself,

565
00:49:21,373 --> 00:49:24,213
sharing my vulnerability, I think is what people say.

566
00:49:24,613 --> 00:49:27,193
And I put it on my email tagline

567
00:49:27,193 --> 00:49:28,253
and I essentially said,

568
00:49:28,393 --> 00:49:30,573
hey, I'm autistic, I'm different.

569
00:49:30,853 --> 00:49:32,793
If what I said isn't clear,

570
00:49:33,073 --> 00:49:34,413
well, that's probably just me.

571
00:49:34,593 --> 00:49:36,893
Just reach out, we'll straighten it all out.

572
00:49:36,893 --> 00:49:45,873
And by me sharing that, I think it took a chip off my shoulder for one.

573
00:49:46,393 --> 00:49:53,293
So it allowed me to understand that not everybody is going to understand me and I should be okay with that.

574
00:49:53,733 --> 00:49:57,773
But at the same time, it put other people unnoticed by seeing that.

575
00:49:57,913 --> 00:50:04,333
They thought, oh, and now I'm just kind of imagining this conversation in their mind.

576
00:50:04,333 --> 00:50:10,993
oh, that's why Derek did that the other day. There's a reason. He's just not being negative.

577
00:50:11,313 --> 00:50:18,133
He was actually trying to solve a problem. So I think embracing that and letting others around me

578
00:50:18,133 --> 00:50:23,873
know rather than trying to keep it in, I think that was a big platform and a launch point for

579
00:50:23,873 --> 00:50:30,573
what later became a foundation for being vulnerable in the program, sharing what I know,

580
00:50:30,573 --> 00:50:37,973
more importantly, sharing what I don't know. Because there's always that the more we know,

581
00:50:39,053 --> 00:50:45,273
the more we realize that we don't know. It's very eye-opening. And I think that was a very

582
00:50:45,273 --> 00:50:53,273
big part of it right there, Havoc, was that just learning that this is who I am and trusting that

583
00:50:53,273 --> 00:50:59,393
that's who I'll always be. I think the big question a lot of people might have,

584
00:50:59,393 --> 00:51:02,593
and especially people that might have some of these conditions themselves.

585
00:51:02,953 --> 00:51:05,873
How do you go for 50 years undiagnosed?

586
00:51:05,873 --> 00:51:09,453
And what actually brought about a diagnosis?

587
00:51:09,653 --> 00:51:13,613
I'm sure it was probably a weight lifted off your shoulders when you found that out.

588
00:51:14,133 --> 00:51:16,793
Well, it was a weight lifted off my shoulders for sure

589
00:51:16,793 --> 00:51:20,633
because now I had a group I could identify with and connect with.

590
00:51:22,053 --> 00:51:24,813
But I was born in 67.

591
00:51:24,813 --> 00:51:32,133
um i am only about a generation away from hilltop people so uh you know it's not something that you

592
00:51:32,133 --> 00:51:37,873
talked about not something people even knew about the term neurodiverse and neurodivergent wasn't

593
00:51:37,873 --> 00:51:43,233
even coined until the late 90s and that was really kind of like the umbrella for all us weird people

594
00:51:43,233 --> 00:51:49,093
that are out there um and i i love the fact that it was that we start talking about you know us

595
00:51:49,093 --> 00:51:54,433
weird people because i i truly think that uh if we counted all the neurodivergence in the world

596
00:51:54,433 --> 00:51:56,893
I don't think we would be the minority.

597
00:51:57,333 --> 00:51:59,313
I think we're probably the majority out there.

598
00:51:59,473 --> 00:52:09,249
We all different in our own way And my family history it turns out once I found out

599
00:52:10,249 --> 00:52:18,229
on my dad's side, at least, we have a high, and it is genetic, the autism side has been

600
00:52:18,229 --> 00:52:25,049
traced genetic. We have a high rate of autism, about three times normal. And that correlates,

601
00:52:25,049 --> 00:52:29,249
or maybe it's just coincidence, we have a high rate of suicide on my dad's side of the family,

602
00:52:29,249 --> 00:52:37,889
too. But growing up, it was man's man's world, right? You didn't speak about problems, you just

603
00:52:37,889 --> 00:52:44,709
solved them. And especially when I went to work in trades, I'll tell you, the people out there

604
00:52:44,709 --> 00:52:50,449
in the building trades, I would probably trust a lot of them with my life, but then others,

605
00:52:50,589 --> 00:52:59,189
I wouldn't trust at all. So it's a wider gamut. And again, it's the Wild West is what

606
00:52:59,189 --> 00:53:07,829
it was. So my time really was keep my mouth shut. And how I think how I eventually got on the train

607
00:53:07,829 --> 00:53:13,389
for diagnosis was it was just out of self-preservation and pursuit. I think I had some

608
00:53:13,389 --> 00:53:19,669
time once my kids got grown up that I started looking for answers in new places and I finally

609
00:53:19,669 --> 00:53:26,289
got there. But back in the 60s, 70s, even 80s, people just weren't talking about this stuff.

610
00:53:26,289 --> 00:53:34,689
all right yeah and especially in the trades i like i said i worked trades uh for a number of

611
00:53:34,689 --> 00:53:42,569
years masonry and and you know it is one of those things you just there's no like nobody cares about

612
00:53:42,569 --> 00:53:49,009
how you feel nobody cares about you know an owl you got i remember working concrete and i got a

613
00:53:49,009 --> 00:53:53,609
cut pretty bad cut on my hand at one point and one of the guys looked at me and he goes

614
00:53:53,609 --> 00:54:02,049
here and slap some, uh, slap some cement into my hand, dry, you know, cement, Portland cement from

615
00:54:02,049 --> 00:54:07,589
the bag, slap that on there, rubbed it in through some duct tape around it and goes, you're good to

616
00:54:07,589 --> 00:54:12,869
go. And then, you know, that's kind of how the trades works, right? So, oh, well, I'm not, it's

617
00:54:12,869 --> 00:54:17,749
not, it doesn't matter, right? The concrete moves from here to there. Concrete goes in this form

618
00:54:17,749 --> 00:54:23,429
and that's it. Nobody cares about anything else, you know? Yeah. Did Havoc, did you work out of

619
00:54:23,429 --> 00:54:30,509
hall or did you work for uh one company no i i worked for a private uh private contractor so

620
00:54:30,509 --> 00:54:38,069
we we mainly uh we did do some commercial work but never never worked uh union so uh we would go

621
00:54:38,069 --> 00:54:43,469
you know do piece work for some of the large communities i was living in southern california

622
00:54:43,469 --> 00:54:51,249
at the time so a lot of buildings just popping up like i mean insane like popcorn just pop pop pop

623
00:54:51,249 --> 00:55:16,629
And, you know, we'd come up and there's a need for someone to do tile for bathrooms and you just do piecework. Right. But most of what we were doing was custom. So we'd go into neighborhoods with new homes, people buy the homes. We'd go pitch, you know, you know, patios in the backyard, outdoor kitchens, water features, you know, you know, whatever. Right. Brick entryway, whatever.

624
00:55:16,629 --> 00:55:23,489
yeah cool i'll just tell my wife earlier today we were talking about my time back in the day and

625
00:55:23,489 --> 00:55:28,689
i uh i tried coming out of high school i was uh kicked out of house when i was 18 so i was trying

626
00:55:28,689 --> 00:55:33,649
to go to college on my own pay for it on my own live on my own and then uh i was pointed toward

627
00:55:33,649 --> 00:55:40,609
the trade so i i added that uh and i was trying to go through trade school and uh and at work at

628
00:55:40,609 --> 00:55:44,589
at the same time of college eventually dropped the college part back then and

629
00:55:44,589 --> 00:55:53,369
And so I worked out of the hall for the first, I don't know, five or six years, which meant that I might have 14 W-2s at the end of the year.

630
00:55:53,429 --> 00:55:56,869
I might have a job that's a day long or a week long or a month long.

631
00:55:57,349 --> 00:55:59,209
Never had anything more than six months.

632
00:55:59,209 --> 00:56:03,449
And it was, you know, you show up or if you don't show up, hey, don't come back tomorrow.

633
00:56:03,749 --> 00:56:04,909
There's no reason to.

634
00:56:05,089 --> 00:56:06,829
You know, they cut bait and go.

635
00:56:07,269 --> 00:56:10,949
And it definitely was Wild West scenario.

636
00:56:10,949 --> 00:56:15,649
And like you said about, you know, when you got that cut, hey, patch it and go on.

637
00:56:15,709 --> 00:56:16,809
You know, you don't want to be a problem.

638
00:56:16,929 --> 00:56:17,329
What's that?

639
00:56:17,869 --> 00:56:24,349
There's a Japanese proverb talks about, you know, nail, a nail sticking out is going to get whacked.

640
00:56:24,689 --> 00:56:28,829
So we try not to be that nail that sticks up and causes problems.

641
00:56:29,769 --> 00:56:30,829
Yep, absolutely.

642
00:56:31,129 --> 00:56:33,309
Move on, get it, get whatever it is.

643
00:56:33,349 --> 00:56:35,489
Just fix it and move on.

644
00:56:35,489 --> 00:56:40,589
And if you can't, well, then, you know, we'll, we'll contact you and we maybe when we have something else.

645
00:56:40,949 --> 00:56:49,609
You know, just buy. That's pretty awesome. So what are, what are, I guess, some of your,

646
00:56:49,769 --> 00:56:55,349
your favorite, cause I have to imagine that you're, you stay in contact with people when

647
00:56:55,349 --> 00:57:00,109
you contract with them for, you know, an implementation. What are some of your favorite

648
00:57:00,109 --> 00:57:04,969
success stories? And you don't necessarily have to go into the organization because confidential

649
00:57:04,969 --> 00:57:09,949
information, et cetera, but what are some of your favorite success stories and, and, you know,

650
00:57:09,949 --> 00:57:25,126
how did that translate for you you know business Well thinking through one that bubbles up is it was a success story there was a uh a

651
00:57:25,126 --> 00:57:32,566
chain of schools i say school they were like uh it wasn't kinder care but it was child care um

652
00:57:32,566 --> 00:57:39,906
child care slash education all in one so um they uh they they were actually looking to sell their

653
00:57:39,906 --> 00:57:47,546
company. They had, it was, gosh, I think like 83 locations and all their knowledge was in the brain

654
00:57:47,546 --> 00:57:53,186
of the main person. And they might've had one manager, might have. I mean, you go into any

655
00:57:53,186 --> 00:57:56,926
retail establishment, you're lucky to find somebody that knows what's going on there.

656
00:57:57,806 --> 00:58:04,186
But they had very few people that the minority knew what was going on everywhere. So now they're

657
00:58:04,186 --> 00:58:13,106
selling this or they want to sell it. And people were interested in buying, but they were hesitant

658
00:58:13,106 --> 00:58:20,326
because once you buy a business, the knowledge that runs especially a service business like this

659
00:58:20,326 --> 00:58:29,146
walks out the door. So a lot of times there's contracts as, hey, the CEO and if it's a big

660
00:58:29,146 --> 00:58:37,106
company, the VP, a number of named people agrees to sign on and be consultants for a week, six

661
00:58:37,106 --> 00:58:44,046
months, six years, whatever that is. And what we ended up doing, we ended up documenting all the

662
00:58:44,046 --> 00:58:51,486
knowledge, all their SOPs. So it actually, it made them work better because now everybody had access

663
00:58:51,486 --> 00:58:57,486
to the information, but at the same time, it allowed them to market themselves with, you know,

664
00:58:57,486 --> 00:59:03,906
holding their head high because when we sell, it's going to be not just the buildings and the

665
00:59:03,906 --> 00:59:09,046
clients that are ephemeral, right? They can disappear at any time, but you have the knowledge

666
00:59:09,046 --> 00:59:15,906
and anybody in your organization can talk to Pocket Mentor and ask questions and get the same

667
00:59:15,906 --> 00:59:21,926
answers I would give to you one-on-one. Now, you know, some people say that, you know, it's never

668
00:59:21,926 --> 00:59:25,366
going to be a hundred percent human. Well, you know, that's fine. We don't need to be a hundred

669
00:59:25,366 --> 00:59:30,726
percent human as long as we can you know follow that that Pareto rule right 80 20 if we can cover

670
00:59:30,726 --> 00:59:37,426
80 of the answers in 20 of the time well then that helps everybody and because of what they did

671
00:59:37,426 --> 00:59:44,146
education anything education I'm fanatical about I think there's no substitute for education I mean

672
00:59:44,146 --> 00:59:48,886
if there was we wouldn't know right we wouldn't be educated on it

673
00:59:48,886 --> 00:59:58,006
i'm i'm sure that i'd really like to see how that pocket mentor works because i think in the back of

674
00:59:58,006 --> 01:00:02,866
our brain at least me i think of my wife trying to call the pharmacy and her yelling at the thing

675
01:00:02,866 --> 01:00:10,406
no speak to a pharmacist no i don't want the photo department like and getting in a fight with the

676
01:00:10,406 --> 01:00:15,466
thing that i'm sure yours is much better than that well it's it's funny that you say that because a

677
01:00:15,466 --> 01:00:23,386
lot of those problems aren't from ai they're uh ivr which i the acronym escapes me right now but ivr

678
01:00:23,386 --> 01:00:33,606
is that uh yeah it's intersome interactive uh voice recognition ivr uh see i knew i'd get there

679
01:00:33,606 --> 01:00:41,706
edit out that hiccup will you but uh those ivrs they they look for key words and if they don't

680
01:00:41,706 --> 01:00:46,206
hear the key words, then they're just computers talking. They're recordings that cause them

681
01:00:46,206 --> 01:00:54,386
problems. The AI that's out there, it's actually just now, in the past 12 months,

682
01:00:54,446 --> 01:00:59,186
is where it's really reached out to a level where if it's implemented correctly,

683
01:00:59,706 --> 01:01:05,306
the AI will actually make those interactions more enjoyable and fruitful.

684
01:01:06,006 --> 01:01:14,186
So one of my frustrations with that IVR is, well, even like on generic voicemail, you go to leave a voicemail.

685
01:01:14,586 --> 01:01:21,026
The first thing I want to know after I left the message is, which button do I need to push, if any, to save the voicemail?

686
01:01:21,446 --> 01:01:23,606
But no, it tells you that last.

687
01:01:23,606 --> 01:01:25,966
It says, press one if this is important.

688
01:01:26,626 --> 01:01:28,386
Press two if it's not important.

689
01:01:29,066 --> 01:01:31,546
Press three, you know, if you'd like to hear your message.

690
01:01:31,546 --> 01:01:37,826
And, you know, like 30 seconds later, it says, or hang up to end this call and save your message.

691
01:01:38,666 --> 01:01:40,746
You know, so they're all different, too.

692
01:01:40,826 --> 01:01:41,766
Do I press pound?

693
01:01:41,886 --> 01:01:42,866
Do I press start?

694
01:01:43,006 --> 01:01:44,106
Do I just hang up?

695
01:01:44,346 --> 01:01:52,206
And if AI can actually facilitate that, like you said, when it's done correctly, it's a tool.

696
01:01:52,966 --> 01:01:56,566
And, yeah, I can shoot you some numbers and you can try out.

697
01:01:56,566 --> 01:02:02,526
I think my first pocket mentor was how to fold a paper airplane.

698
01:02:03,886 --> 01:02:04,406
Really?

699
01:02:05,006 --> 01:02:05,486
Interesting.

700
01:02:06,286 --> 01:02:07,686
That is kind of interesting.

701
01:02:08,486 --> 01:02:10,466
So how do you limit?

702
01:02:11,126 --> 01:02:24,926
So one of the big things that's always frustrating about AI, I think for everyone, right, is the, I mean, one of the big benefits of it is its ability to, you know, do inference, right?

703
01:02:24,926 --> 01:02:29,946
but also one of the big detriments is it can infer a thing

704
01:02:29,982 --> 01:02:37,682
that don't actually exist, like maybe bits and pieces do in different areas, but yeah,

705
01:02:37,682 --> 01:02:46,122
it all fits together. So here, how did you go about limiting the inference, which could lead to,

706
01:02:46,122 --> 01:02:51,602
you know, probably not disastrous, but bad information coming through?

707
01:02:52,882 --> 01:02:58,862
Absolutely. So if I forget what your question is, remind me, but I want to preface this by a little

708
01:02:58,862 --> 01:03:07,862
story. So IBM, back when they were doing AI before us consumers could play with AI,

709
01:03:08,522 --> 01:03:15,742
they had this thing called Big Blue and they were teaching it chess. And Gary Kasparov is

710
01:03:15,742 --> 01:03:22,862
who they challenged and Gary Kasparov kicked Big Blue's butt. And so they said, well, let's try

711
01:03:22,862 --> 01:03:29,282
this again. And then Garry Kasparov kicked their butt again. But eventually, IBM's AI beat Garry

712
01:03:29,282 --> 01:03:36,422
Kasparov. And so that was like a telling point. Computers have finally gotten smarter than AI

713
01:03:36,422 --> 01:03:44,062
or smarter than humans is what the headlines read. But if we go back and analyze what happened,

714
01:03:44,782 --> 01:03:51,362
Garry Kasparov at the time was the number one ranked chess player in the world. Those guys and

715
01:03:51,362 --> 01:03:59,202
gals are pretty smart, right? They analyzed the decisions that were made by the computer and they

716
01:03:59,202 --> 01:04:03,762
found out that the computer only made the optimum decision or the best decision or the right

717
01:04:03,762 --> 01:04:10,562
decision two out of three times. So what that means is that even the best chess player in the

718
01:04:10,562 --> 01:04:16,842
world would make the best decision less than two out of three times, maybe like one out of two or

719
01:04:16,842 --> 01:04:25,682
a little more. So if we take this perfection with a grain of salt and if we want to validate,

720
01:04:26,022 --> 01:04:33,002
is AI worth it or is AI working? We have to understand that we're not getting 100% accuracy

721
01:04:33,002 --> 01:04:40,142
right now. Even if we're on the line, even if we call and we say, hey, how am I supposed to do this?

722
01:04:40,142 --> 01:04:48,082
I forget. In manufacturing, I spent a lot of time in automotive manufacturing. If the line goes down,

723
01:04:48,182 --> 01:04:52,962
they want it up and running. They don't care who gets it up and running, but depending on the

724
01:04:52,962 --> 01:04:58,342
problem, a lot of time it's John, right? And if it's five o'clock, John's already home for the day,

725
01:04:58,402 --> 01:05:05,582
so John gets a phone call because John is the subject matter expert, the SME, and John gives

726
01:05:05,582 --> 01:05:10,282
them the answer. Now, will that answer be the right answer? Well, it's probably going to be the

727
01:05:10,282 --> 01:05:17,762
best answer. But if we've already looked at the data and that humans aren't right 100% of the time,

728
01:05:18,362 --> 01:05:23,242
John's probably going to be right, but even his answer could be wrong. And that's what we need

729
01:05:23,242 --> 01:05:31,022
to look at. So if we're looking to amplify humans, we want to do at least as good as the human does.

730
01:05:31,022 --> 01:05:36,322
If we can just do that to support the human and speed up the answer, it's a win.

731
01:05:37,402 --> 01:05:50,782
And at the same time, AI, computers in general, their scalability, John isn't always going to be available at 5 p.m. on a Wednesday, you know, because it could be 2.30 a.m. on a Saturday.

732
01:05:50,782 --> 01:05:59,542
and your low-ranking operator or millwright or repairman is going to be there on their own,

733
01:05:59,542 --> 01:06:04,582
and they only have about six months' experience, and they don't know how to answer that question.

734
01:06:04,742 --> 01:06:09,622
So they could call John. I tell you, John's been around long enough. He's not going to answer that

735
01:06:09,622 --> 01:06:15,802
phone on a Saturday night. So how do they get that answer? AI will allow them to scale that answer.

736
01:06:15,802 --> 01:06:21,242
Now, your question, if I recall correctly, is how do we know it's given the right answer?

737
01:06:21,242 --> 01:06:46,756
And the way we know it giving maybe not 100 of the time the right answer but the best answer is that we reduce the knowledge set to the smallest size possible So if we stay in automotive manufacturing we might give specifics on how to change brakes on a Chevy versus how to change brakes on a Ford There a lot of similarities but there might be

738
01:06:46,756 --> 01:06:53,476
certain springs or clips that are different. So we don't even include information that is not

739
01:06:53,476 --> 01:07:01,096
applicable to that specific role. And this is where AI succeeds is when we limit its responsibility.

740
01:07:01,836 --> 01:07:10,396
It's AI, in my belief, will never be available to one instance answer every question in the world.

741
01:07:10,856 --> 01:07:11,916
It won't.

742
01:07:12,256 --> 01:07:26,816
But if we reduce the specificity of the knowledge to as small a data set as possible, something that is specific for one role, one job, then it will be right 99 plus percent of the time.

743
01:07:26,816 --> 01:07:35,056
is there any research or thought put into what the voice model used sounds like whether it's a

744
01:07:35,056 --> 01:07:40,276
male voice or a female voice or even the accent or anything or are they all pretty much just

745
01:07:40,276 --> 01:07:46,756
standard well see uh now you're into my daughter's realm uh she's a psychologist and but she fell into

746
01:07:46,756 --> 01:07:54,576
uh into technology she's been working technology front they call that human factors hf and it's that

747
01:07:54,576 --> 01:08:01,656
interface, what is best for the human? Now, 20 years ago, 50 years ago, 100 years ago,

748
01:08:02,176 --> 01:08:08,736
if we're being taught how to do a new job, whether it's at home, on the farm, or Ford Motor,

749
01:08:10,056 --> 01:08:14,056
we're going to be told one way, and we have to figure it out on our own.

750
01:08:14,916 --> 01:08:19,996
And we either adapt or we don't. And about half to two-thirds can get it on the first try,

751
01:08:19,996 --> 01:08:28,576
But then the other half to one third may not figure it out or it may take too long to figure it out, then they lose their job.

752
01:08:29,376 --> 01:08:34,516
So it's getting that information in the shortest amount of time that's important.

753
01:08:35,116 --> 01:08:37,636
So what AI allows us to do is specialize.

754
01:08:38,896 --> 01:08:42,856
Today, we are at a one size fits one role.

755
01:08:43,796 --> 01:08:51,156
And if I have a very, we have IQ, which is intelligence quotient, like how much do we know?

756
01:08:51,276 --> 01:08:56,196
But then there's also EQ, which came about about 30 years ago or so, the emotional quotient.

757
01:08:56,416 --> 01:08:59,856
It's how well I can connect to you as a human to human.

758
01:09:00,496 --> 01:09:09,596
And if I stumble upon you and you fell off your bicycle, I'm going to treat you like you just fell off your bicycle.

759
01:09:10,016 --> 01:09:11,296
That's the context.

760
01:09:11,296 --> 01:09:18,416
and if you just won the lottery, I'll go have a beer with you. Hey, congratulations. Nice job.

761
01:09:18,416 --> 01:09:26,796
What a lucky cat you are. And AI is in a position where whether we're neurodiverse,

762
01:09:27,236 --> 01:09:35,136
neurodivergent, or what's neurotypical is the normal people is by the big category here.

763
01:09:35,916 --> 01:09:43,156
AI doesn't care because AI will adapt much quicker than a human will to the context.

764
01:09:43,156 --> 01:09:49,916
What type of questions you ask, what type of answers you expect, and it will keep track

765
01:09:49,916 --> 01:09:55,296
of every answer that AI gives from the SOP, from the company database.

766
01:09:55,756 --> 01:10:00,256
If the human says, now, wait a minute, can you explain that to me again?

767
01:10:00,256 --> 01:10:06,756
Or can you tell me how that works with a blue, whatchamacallit, instead of a red, whatchamacallit?

768
01:10:07,196 --> 01:10:08,956
Then AI will rephrase.

769
01:10:09,196 --> 01:10:17,256
And AI will adapt to that individual to have that best friend, I've known you all my life communication.

770
01:10:18,156 --> 01:10:25,696
So the communication value is just the efficiency just skyrockets because it is that one size fits one.

771
01:10:25,696 --> 01:10:34,036
And AI can do that for pennies on the dollar versus putting humans in to stand by, we'll

772
01:10:34,036 --> 01:10:36,056
take your call, operators are standing by.

773
01:10:36,116 --> 01:10:37,396
That's what I was going for, right?

774
01:10:40,396 --> 01:10:52,851
So the way you trim down the inference problem is literally your LLM I assuming is like you said you saying okay well this is a Ford plant I don need to include Chevy

775
01:10:53,071 --> 01:10:53,931
I don't need GM.

776
01:10:54,091 --> 01:10:55,831
I don't need, you know, Dodge.

777
01:10:55,911 --> 01:10:57,991
I don't need any of this stuff, Chrysler, et cetera.

778
01:10:57,991 --> 01:11:09,091
I just need to train it only on the Ford specific data and maintain just that subset of data documents, you know, maybe their document repository, et cetera.

779
01:11:09,091 --> 01:11:14,751
and thereby you limit the amount of inference that can be brought up by AI

780
01:11:14,751 --> 01:11:20,571
by not allowing it access to non-relevant material.

781
01:11:21,251 --> 01:11:23,491
Absolutely. We do the same thing with humans.

782
01:11:23,851 --> 01:11:25,531
And that's really where this adapts from.

783
01:11:25,991 --> 01:11:31,911
My takeaways at my time at Amazon and the successfulness of the program that I had there

784
01:11:31,911 --> 01:11:38,471
and then using AI for what AI can do, the technology of it is adapt and deliver that information.

785
01:11:38,471 --> 01:11:47,711
But just like you would take a Ford to a Ford mechanic, a Chevy mechanic might fix the Ford.

786
01:11:48,271 --> 01:11:55,691
But some of us says, no, no, I'm going to the expert because the Ford mechanic is going to know more than the Chevy mechanic does.

787
01:11:56,551 --> 01:11:59,591
So yeah, we're following the same model as humans do.

788
01:12:00,271 --> 01:12:03,831
We're just amplifying those human abilities.

789
01:12:05,651 --> 01:12:06,871
That is pretty cool.

790
01:12:06,871 --> 01:12:13,391
very cool i'm just curious what happens if somebody uh asks something that's outside of

791
01:12:13,391 --> 01:12:18,571
its purview uh say like hey pocket betra you want to chat today hey how you doing today they're just

792
01:12:18,571 --> 01:12:24,411
lonely and i want somebody to talk to does it have a something that cues up and says okay i can't

793
01:12:24,411 --> 01:12:29,551
talk about that well it's interesting you said that because mental wellness is huge and it's uh

794
01:12:29,551 --> 01:12:35,991
in hr departments you know keeping track of how the workers actually feel day to day because are

795
01:12:35,991 --> 01:12:38,111
Are they going to come in tomorrow if it's a bottom line?

796
01:12:38,291 --> 01:12:40,951
Or can we make your day better?

797
01:12:41,691 --> 01:12:49,791
Good HR departments, good HR employees really do care about the employees.

798
01:12:49,791 --> 01:12:56,111
So in a situation like that, it's whatever the company contracts for.

799
01:12:56,651 --> 01:12:58,071
It's their information.

800
01:12:58,231 --> 01:12:59,451
We don't create the information.

801
01:12:59,451 --> 01:13:07,291
we give away for, we give a new interface to that information for their employees so they can,

802
01:13:07,791 --> 01:13:13,151
it feels more like osmosis than, you know, be standing and being blared at for a week and then

803
01:13:13,151 --> 01:13:18,531
go out and try to do a job. So if they ask a question that's out of context, it's only out of

804
01:13:18,531 --> 01:13:28,151
context if it's outside the SOP. And, but we do have those mental awareness and wellness

805
01:13:28,151 --> 01:13:36,191
triggers in there that can do whatever the company designed. We can have it chat back a time or two,

806
01:13:36,411 --> 01:13:43,071
or we can just say, hey, I think you better talk to John or Peggy or Mary over in HR.

807
01:13:43,071 --> 01:13:51,151
You should defer there. So it is programmable. It's guardrails to as tight as you want to run

808
01:13:51,151 --> 01:13:59,511
the guardrails, just like any piece of software is. So it can either engage or redirect, but there's

809
01:13:59,511 --> 01:14:04,251
always going to be a limitation on how far it engages because it is important to keep that

810
01:14:04,251 --> 01:14:13,451
knowledge pool as small as possible. Interesting. Very cool. Well, I think this is probably a good

811
01:14:13,451 --> 01:14:19,511
place to start wrapping up a little bit here. One of the things we like to do with all of our

812
01:14:19,511 --> 01:14:24,131
guess, is give them an opportunity to give back to someone that was influential to them.

813
01:14:24,711 --> 01:14:31,511
And throughout our lives, we have a figure that had a massive impact on us. It could be a sixth

814
01:14:31,511 --> 01:14:37,951
grade science teacher, historical figure, an author, it could be mom. But who would you like

815
01:14:37,951 --> 01:14:46,551
to shout out? Well, it's posthumously, but Kim Chavez. She was my first supervisor when I was at

816
01:14:46,551 --> 01:14:46,871
Amazon.

817
01:14:47,411 --> 01:15:00,765
And she gave me some insight because you know working at Amazon and the you know kind of a white collar crowd and one of the largest companies in the world that is just all over you got to produce or you gone type of scenario

818
01:15:01,545 --> 01:15:11,485
She gave me some advice, and it was treat the job like it's a business, which did two things.

819
01:15:11,625 --> 01:15:15,905
One, she said, keep track of the good things that you do because nobody else cares.

820
01:15:15,905 --> 01:15:22,165
and when it comes time for selling your program or selling yourself,

821
01:15:22,565 --> 01:15:25,185
you have valid data points to share.

822
01:15:26,025 --> 01:15:27,505
And that was one point.

823
01:15:27,505 --> 01:15:33,085
The other point is that she has an autistic child herself.

824
01:15:34,725 --> 01:15:38,485
And so she was actually a couple years younger than I,

825
01:15:38,745 --> 01:15:42,425
but I always looked at her as a mother figure

826
01:15:42,425 --> 01:15:46,565
because she treated me with, I don't know, kid gloves.

827
01:15:46,685 --> 01:15:49,385
I still feel like I'm a kid with this autism.

828
01:15:50,965 --> 01:15:54,765
It's hard for me to say I'm the best person in the room,

829
01:15:54,825 --> 01:15:58,045
even if I am, which it might happen, but it's rare.

830
01:15:58,045 --> 01:16:03,585
But she would do what good leaders do,

831
01:16:03,845 --> 01:16:05,445
and that is stand up for your employees

832
01:16:05,445 --> 01:16:07,945
and protect their backsides,

833
01:16:07,945 --> 01:16:12,565
but allow them to grow and perform to what they're capable of.

834
01:16:13,165 --> 01:16:21,265
And under her guidance and her shield of protection at a company like Amazon,

835
01:16:21,665 --> 01:16:22,985
she took a lot of hits for me.

836
01:16:23,105 --> 01:16:28,365
But the payoff allowed me to build something that had never been built at Amazon

837
01:16:28,365 --> 01:16:35,865
because she took care of some of the business stuff that I wasn't comfortable or capable of doing.

838
01:16:35,865 --> 01:16:39,845
It allowed me to be creative and to continue to grow something good.

839
01:16:40,185 --> 01:16:51,465
And I admire her for the business acumen that she had and admire her for the person to person, personability and love she had for other humans.

840
01:16:53,245 --> 01:16:54,045
It's awesome.

841
01:16:54,525 --> 01:16:58,685
I think you're the first guest we've ever had that shouted out a supervisor.

842
01:16:59,365 --> 01:17:00,445
It's a pretty cool thing.

843
01:17:00,445 --> 01:17:09,225
And actually, I started in my brain thinking about it and I'm like, wow, there are a few people that I would probably shout out to in that same role.

844
01:17:09,645 --> 01:17:10,465
Very cool.

845
01:17:10,905 --> 01:17:12,845
So where can people find you?

846
01:17:12,925 --> 01:17:13,985
What do you have going on?

847
01:17:14,165 --> 01:17:16,985
Where you said the audio book is released.

848
01:17:17,225 --> 01:17:20,465
Where is the best place for someone to find that?

849
01:17:20,825 --> 01:17:22,285
And yeah.

850
01:17:22,605 --> 01:17:23,585
Yeah, absolutely.

851
01:17:24,085 --> 01:17:26,325
Hopefully you'll have room in the show notes.

852
01:17:26,325 --> 01:17:29,325
We can put a couple of URLs underneath the audio.

853
01:17:30,445 --> 01:17:34,285
But if you Google me, I come up pretty readily.

854
01:17:35,565 --> 01:17:42,865
PracticalAI.app is the business home for my company.

855
01:17:43,405 --> 01:17:49,285
And the book, everybody can get a free read of the first chapter at HumanFirstAI.net.

856
01:17:50,485 --> 01:17:51,245
Awesome.

857
01:17:51,965 --> 01:17:59,025
So PracticalAI.app or HumanFirstAI.net.

858
01:17:59,385 --> 01:18:00,325
Go check them out.

859
01:18:00,445 --> 01:18:01,045
Thank you.

860
01:18:01,165 --> 01:18:01,505
It's good stuff.

861
01:18:02,385 --> 01:18:04,565
Thank you very much for giving us some of your time.

862
01:18:04,645 --> 01:18:05,845
We really appreciate it.

863
01:18:06,105 --> 01:18:09,425
And thank you for giving us some of your time as well.

864
01:18:09,785 --> 01:18:11,165
You know where you can find us.

865
01:18:11,265 --> 01:18:11,765
We're everywhere.

866
01:18:12,325 --> 01:18:16,465
Podcasts can be found as well as YouTube and, you know, other social media sites.

867
01:18:16,585 --> 01:18:18,445
If you like this, give us a thumbs up.

868
01:18:18,505 --> 01:18:19,845
If not, give us a thumbs down.

869
01:18:19,945 --> 01:18:20,485
Leave a comment.

870
01:18:21,685 --> 01:18:25,945
If you have an idea for a show, topic, guest, throw that in the comments.

871
01:18:26,085 --> 01:18:29,645
We'll reply and be happy to try and work on getting something done.

872
01:18:30,445 --> 01:18:37,985
And remember, wherever you are, whenever you're watching this, have an amazing day.

873
01:18:39,085 --> 01:18:40,545
Thanks for coming by, Derek. See you, everybody.

874
01:19:00,445 --> 01:19:01,445
.