1
00:00:00,150 --> 00:00:02,452
Everybody knows the AI story by now.

2
00:00:02,602 --> 00:00:05,524
It's coming for jobs. It uses a staggering
amount of water.

3
00:00:05,674 --> 00:00:09,477
And depending on who you ask, it's either
holding up the stock market or about

4
00:00:09,477 --> 00:00:10,938
to knock the stock market over.

5
00:00:11,088 --> 00:00:13,550
You've heard all that, and I've heard all
of that.

6
00:00:13,770 --> 00:00:17,301
This week, I went looking for the part
that nobody is actually talking about,

7
00:00:17,522 --> 00:00:21,224
which is what AI is doing to how we treat
each other.

8
00:00:21,384 --> 00:00:24,566
Friends, coworkers, teachers, and kids,

9
00:00:24,706 --> 00:00:27,577
total strangers. There's research on all
of it now.

10
00:00:27,811 --> 00:00:29,327
And most of it kinda surprised me.

11
00:00:29,477 --> 00:00:30,741
Will hasn't heard any of this,

12
00:00:30,891 --> 00:00:31,891
and that's on purpose.

13
00:00:46,875 --> 00:00:48,076
Before we get into the research,

14
00:00:48,226 --> 00:00:52,449
I just want you to know where I'm coming
from because I'm not some guy who's never

15
00:00:52,449 --> 00:00:55,041
touched this stuff before. I I work in
cloud engineering,

16
00:00:55,041 --> 00:00:58,613
Kubernetes infrastructure, and the
plumbing that sort of keeps websites

17
00:00:58,613 --> 00:01:02,786
and apps running. And I use AI every
single day because that's how I get paid,

18
00:01:02,786 --> 00:01:03,937
not because what I want to do,

19
00:01:04,087 --> 00:01:07,269
but because that is how I make a living in
the world.

20
00:01:07,529 --> 00:01:12,552
picture getting handed kind of a whole
team of brand new college graduates,

21
00:01:12,552 --> 00:01:13,092
smart.

22
00:01:13,092 --> 00:01:17,685
Eager, fast, and every one of them will
tell you with total confidence that

23
00:01:17,685 --> 00:01:20,126
they've got it handled. Then you check
their work.

24
00:01:20,276 --> 00:01:24,108
Will you ever manage somebody who was
completely confident and also completely

25
00:01:24,108 --> 00:01:24,608
wrong?

26
00:01:24,608 --> 00:01:26,399
Yes, many times unfortunately.

27
00:01:26,770 --> 00:01:29,361
So, you know, you walk them through it
every step.

28
00:01:29,511 --> 00:01:33,012
A new real graduate takes more of my time,

29
00:01:33,112 --> 00:01:35,253
obviously, than a seasoned engineer,

30
00:01:35,253 --> 00:01:37,353
but after enough repetitions, right,

31
00:01:37,353 --> 00:01:38,484
they eventually they remember it.

32
00:01:38,634 --> 00:01:40,785
You tell them something five times and the
sixth time,

33
00:01:40,785 --> 00:01:43,986
okay, they've finally got it and they
don't need you for that from now on.

34
00:01:43,986 --> 00:01:45,597
They make a mental workbook of it.

35
00:01:45,747 --> 00:01:47,837
The AI shows up every morning,

36
00:01:47,837 --> 00:01:49,787
brand new. It's Groundhog Day,

37
00:01:49,967 --> 00:01:51,798
and I'm the only one who remembers
yesterday.

38
00:01:51,948 --> 00:01:53,790
Speaking of which, this is the overlap.

39
00:01:53,940 --> 00:01:58,152
If you are new here, you didn't have any
idea that this is not how we start every

40
00:01:58,152 --> 00:02:00,404
other show. But if you are not new here,

41
00:02:00,404 --> 00:02:03,527
you absolutely realize this is not how we
start every other show.

42
00:02:03,707 --> 00:02:05,789
We're trying out a little bit of a
different format today.

43
00:02:05,914 --> 00:02:09,203
Uh I'm Joshua. I am the co-host of The
Overlap,

44
00:02:09,203 --> 00:02:10,274
as well as my buddy Will.

45
00:02:10,424 --> 00:02:11,679
I'm here and I know nothing.

46
00:02:12,050 --> 00:02:13,859
We left him out of the research today.

47
00:02:13,859 --> 00:02:15,811
We're just gonna kind of get his you know,

48
00:02:15,811 --> 00:02:18,122
his over/under on most of this stuff.

49
00:02:18,162 --> 00:02:21,104
I did the research by myself. I wrote up
the show and I was like,

50
00:02:21,104 --> 00:02:23,074
hey, let's try this as a as a format.

51
00:02:23,074 --> 00:02:26,379
there's a couple of different avenues out
there that that use this format,

52
00:02:26,379 --> 00:02:27,870
and so we thought we'd give it a shot.

53
00:02:28,020 --> 00:02:32,495
So flames go to at Joshua at the overlap
dot f dot foundation.

54
00:02:32,085 --> 00:02:33,085
Flames.

55
00:02:33,379 --> 00:02:34,296
Yeah.

56
00:02:34,296 --> 00:02:37,178
Yes, absolutely. you can send that email
there.

57
00:02:37,328 --> 00:02:39,619
But anyway, so getting back to the story,

58
00:02:39,639 --> 00:02:42,980
an AI shows up every single morning
completely blank slate.

59
00:02:42,980 --> 00:02:43,941
And it's not just every morning,

60
00:02:43,941 --> 00:02:46,922
it's every thread. It's every single thing
that you might want it to work on.

61
00:02:46,922 --> 00:02:51,194
It starts out with zero context
whatsoever.

62
00:02:51,344 --> 00:02:53,535
Now look, there's there's ways to give it
a memory,

63
00:02:53,535 --> 00:02:55,556
right? Like I I built one for myself.

64
00:02:55,556 --> 00:02:57,217
I I talked to Will about it a bit.

65
00:02:57,367 --> 00:02:57,657
but

66
00:02:57,657 --> 00:03:00,469
my job is simply now turned into this
thing.

67
00:03:01,089 --> 00:03:04,070
I make every decision so that it gets done
right,

68
00:03:04,220 --> 00:03:06,961
and I steer the machines while they do the
typing.

69
00:03:07,111 --> 00:03:10,262
It's like a paired programming session
where they know nothing,

70
00:03:10,262 --> 00:03:14,814
and I'm directing multiples of them at a
given time.

71
00:03:14,964 --> 00:03:17,064
I am not the only person who's doing this.

72
00:03:17,064 --> 00:03:20,065
Hubstaff, a company that makes workplace
software,

73
00:03:20,065 --> 00:03:22,456
surveyed about a thousand American workers
this fall.

74
00:03:22,606 --> 00:03:24,647
Among the people who use AI at work,

75
00:03:24,647 --> 00:03:29,303
forty six percent said that they're
expected to do more work since they

76
00:03:29,303 --> 00:03:34,247
started using it, with only about five
percent less work than when they started

77
00:03:34,247 --> 00:03:38,692
using AI. What are you what are your
initial thoughts on this sort of concept,

78
00:03:38,692 --> 00:03:39,042
Will?

79
00:03:39,042 --> 00:03:42,298
Well, it makes sense, right? If people are
trying to replace people with AI,

80
00:03:42,298 --> 00:03:45,461
then they're gonna expect the people that
are still there to do more work to cover

81
00:03:45,461 --> 00:03:46,612
the people that aren't there anymore.

82
00:03:46,762 --> 00:03:48,356
But it kinda sounds like the hey.

83
00:03:47,086 --> 00:03:48,676
But AI is supposed to fill that in,

84
00:03:48,676 --> 00:03:49,057
right?

85
00:03:49,057 --> 00:03:51,328
That's what I'm saying. that's that's it
it is,

86
00:03:51,328 --> 00:03:53,830
but it isn't, right? I it's filling in for
the two people,

87
00:03:53,980 --> 00:03:57,172
but it's not filling it still requires the
one person to do more work to fill

88
00:03:57,172 --> 00:04:00,094
in the so it's less than two full jobs
worth of human work.

89
00:04:00,094 --> 00:04:01,505
They're they're not working twice as much,

90
00:04:01,655 --> 00:04:04,277
but they've got to work more to compensate
for the other person who's not there

91
00:04:04,277 --> 00:04:04,557
anymore.

92
00:04:04,557 --> 00:04:08,049
Okay. So essentially the the machines save
you,

93
00:04:08,049 --> 00:04:11,822
you know, an hour, right? Your boss finds
two more things to put in it.

94
00:04:11,942 --> 00:04:12,603
Meanwhile,

95
00:04:12,393 --> 00:04:12,801
Right.

96
00:04:12,801 --> 00:04:16,663
I don't work with interns or new graduates
anymore or anything like that.

97
00:04:17,099 --> 00:04:20,811
but companies aren't hiring them like at
all.

98
00:04:20,811 --> 00:04:25,034
They're not hiring new grads, they're not
hiring interns because now we have AI that

99
00:04:25,034 --> 00:04:26,515
can do the same work.

100
00:04:26,665 --> 00:04:31,078
Anyway, so a team at Stanford went through
payroll records from ADP,

101
00:04:31,228 --> 00:04:35,451
the company that runs like a payroll for a
huge chunk of the United States.

102
00:04:35,601 --> 00:04:40,755
And since ChatGPT came out, workers aged
22 to 25,

103
00:04:40,755 --> 00:04:45,138
which in you know in the jobs market are
the most exposed to AI right now,

104
00:04:45,288 --> 00:04:51,173
saw their employment drop about 16%
compared with jobs that AI barely touches.

105
00:04:51,323 --> 00:04:51,773
Yeah.

106
00:04:51,327 --> 00:04:52,213
That's brutal.

107
00:04:52,213 --> 00:04:57,366
That's that's awful. I looking at the
numbers was was has been an interesting

108
00:04:57,366 --> 00:05:02,449
eye opener for me. So now we have sixteen
percent less jobs for compared to ones

109
00:05:02,449 --> 00:05:07,212
that AI doesn't touch, right? So software
developers are obviously the the most rife

110
00:05:07,362 --> 00:05:11,004
and ripe for the picking for AI to sort of
replace,

111
00:05:11,004 --> 00:05:14,956
right? The the researchers call those
young workers the canaries in the coal

112
00:05:14,956 --> 00:05:16,567
mine. Unfortunately.

113
00:05:16,717 --> 00:05:16,867
the

114
00:05:16,867 --> 00:05:20,795
canaries die in the coal mine and that's
how you know that that there's there's

115
00:05:20,795 --> 00:05:21,947
sour gas in the coal mine.

116
00:05:22,097 --> 00:05:23,468
So we didn't about this beforehand,

117
00:05:23,468 --> 00:05:25,880
but I've actually been reading Ethan
Mollick's new book,

118
00:05:26,621 --> 00:05:27,572
um, which is about

119
00:05:26,766 --> 00:05:27,572
And what's that called?

120
00:05:27,572 --> 00:05:29,789
it's I'm blinking on the title now.

121
00:05:29,789 --> 00:05:31,214
His previous one was Co-Intelligence,

122
00:05:31,214 --> 00:05:35,397
I think. but one of them is this one is
about basically like the three years later

123
00:05:35,397 --> 00:05:37,689
look at working with AI, and he's a Wharton
professor.

124
00:05:37,689 --> 00:05:42,522
Uh pretty famous for his like he kind of
became virally famous for his test

125
00:05:42,522 --> 00:05:44,103
of AI image generating models.

126
00:05:44,243 --> 00:05:44,914
And so he like

127
00:05:44,446 --> 00:05:44,914
Uh-huh.

128
00:05:44,914 --> 00:05:46,845
he did the otter, the otter on a plane,

129
00:05:46,845 --> 00:05:47,826
otter on an airplane.

130
00:05:47,826 --> 00:05:49,657
And the first time he did it, it was like
he said,

131
00:05:49,657 --> 00:05:51,348
Please draw me an otter on an airplane.

132
00:05:51,498 --> 00:05:54,300
And it drew an literally an otter on the
outside of an airplane,

133
00:05:54,300 --> 00:05:57,001
like a giant otter on the on top of an
airplane.

134
00:05:57,151 --> 00:05:59,762
And then later figured out that you want
them in the seats on an airplane,

135
00:05:59,762 --> 00:06:01,453
right? Like what we mean by on an
airplane.

136
00:06:01,603 --> 00:06:03,339
And so since then it's gotten better and
better.

137
00:06:03,339 --> 00:06:06,516
And it's kind of like a a history of
development of LLMs,

138
00:06:06,666 --> 00:06:07,881
which is pretty cool. But anyway,

139
00:06:07,881 --> 00:06:09,568
he was saying that what they're learning
now,

140
00:06:09,568 --> 00:06:12,614
the nuanced learning is that there's a
difference between like

141
00:06:12,764 --> 00:06:17,356
Productivity with AI and or like the the
talented people aren't necessarily

142
00:06:17,356 --> 00:06:19,774
the ones who can do the most with AI.

143
00:06:19,997 --> 00:06:22,058
It's actually those who have a better AI
synergy.

144
00:06:22,294 --> 00:06:24,239
So it's kind of like a different kind of
emotional,

145
00:06:24,239 --> 00:06:26,660
like an emotional quadrant or quotient
versus IQ,

146
00:06:26,660 --> 00:06:28,541
you know, EQ versus IQ or something.

147
00:06:28,691 --> 00:06:31,582
Where those that are better working with
AI can get better results than

148
00:06:31,582 --> 00:06:34,242
the top people working with or without AI.

149
00:06:34,392 --> 00:06:36,512
So if they're really good at doing that.

150
00:06:34,846 --> 00:06:37,537
So kind of like the Centaur writing that
we had talked about before.

151
00:06:37,343 --> 00:06:37,783
Yeah,

152
00:06:37,783 --> 00:06:38,373
exactly.

153
00:06:38,373 --> 00:06:40,884
And so to give you a little bit of a
history on that,

154
00:06:40,884 --> 00:06:43,515
Will and I used to have these
conversations about Centaur writing,

155
00:06:43,705 --> 00:06:45,356
which he pulls from Centaur chess,

156
00:06:45,396 --> 00:06:47,877
which was that, you know, chess computers,

157
00:06:47,877 --> 00:06:49,458
chess is essentially a solved game.

158
00:06:49,678 --> 00:06:51,639
And I think we've talked about this
actually on the show before,

159
00:06:51,639 --> 00:06:58,612
but chess is a solved game, and any chess
computer can beat a human that is playing

160
00:06:58,612 --> 00:07:00,003
chess, but a a

161
00:07:00,103 --> 00:07:02,585
And a any computer can beat another
computer,

162
00:07:02,735 --> 00:07:07,549
but a person with a computer has trouble
beating a person with a computer.

163
00:07:07,699 --> 00:07:11,773
So essentially that they are stronger than
the sum of their parts,

164
00:07:11,773 --> 00:07:17,408
essentially by utilizing AI along with
their human abilities to create something

165
00:07:17,408 --> 00:07:19,300
better than human and better than AI.

166
00:07:19,450 --> 00:07:21,282
Right, because AI sees optimal moves,

167
00:07:21,432 --> 00:07:22,903
but they don't really build strategies.

168
00:07:22,903 --> 00:07:25,146
It looks like they're building strategies
because they're always making the best

169
00:07:25,146 --> 00:07:27,269
move into whatever strategy they're in.

170
00:07:27,419 --> 00:07:30,465
But they could have started out with any
random move and it would have been totally

171
00:07:30,465 --> 00:07:33,826
different. But the humans see strategy
over time playing out and that gives them

172
00:07:33,826 --> 00:07:35,447
an advantage in some situations.

173
00:07:35,597 --> 00:07:37,348
So this year, through October,

174
00:07:37,348 --> 00:07:42,831
so through this past month, American
employers have pointed to AI as the reason

175
00:07:42,831 --> 00:07:47,664
for 121,401 announced job cuts.

176
00:07:47,664 --> 00:07:51,616
So that's from Challenger, Gray &
Christmas,

177
00:07:51,616 --> 00:07:54,898
which is an outplacement firm that tracks
layoffs every month.

178
00:07:55,048 --> 00:07:59,011
there's also a one for tech. I think it's
called layoffs.fyi.

179
00:07:59,161 --> 00:07:59,591
And

180
00:07:59,591 --> 00:08:02,692
it's about basically one in five of all
cuts,

181
00:08:02,733 --> 00:08:07,084
job cuts that they counted were AI driven
cuts.

182
00:08:08,105 --> 00:08:10,736
which at least I mean that's the that's
the reasons that the companies gave,

183
00:08:10,736 --> 00:08:12,247
right? So take it with a grain of salt.

184
00:08:12,427 --> 00:08:14,758
Maybe that's more, maybe that's less,

185
00:08:14,758 --> 00:08:17,289
but ultimately that's what they're saying
out loud,

186
00:08:17,439 --> 00:08:19,700
right? I hear a lot of stories about
people,

187
00:08:19,700 --> 00:08:22,181
you know, getting laid off for whatever
reason,

188
00:08:22,221 --> 00:08:25,833
you know, irresponsibility or or they were
late to work twelve times,

189
00:08:25,833 --> 00:08:26,703
six minutes each.

190
00:08:26,703 --> 00:08:29,159
And but it could still also be AI,

191
00:08:29,159 --> 00:08:31,709
right? it could be making it more
cutthroat for the rest of us.

192
00:08:31,859 --> 00:08:34,850
Nobody wants to admit that they overhired
severely during the pandemic

193
00:08:34,850 --> 00:08:36,645
and are trying to clean up their mess.

194
00:08:36,795 --> 00:08:37,455
Maybe so.

195
00:08:37,455 --> 00:08:37,986
So

196
00:08:37,986 --> 00:08:40,637
So the these people, right? These these
interns,

197
00:08:40,637 --> 00:08:45,288
these, these college students that I used
to to teach and work with and and m mold

198
00:08:45,288 --> 00:08:46,948
into tomorrow's software engineers,

199
00:08:47,689 --> 00:08:50,110
they they had they have a memory to them,

200
00:08:50,260 --> 00:08:54,562
right? The ones who who'd now be senior
senior engineers in ten years that that

201
00:08:54,722 --> 00:08:56,703
can't really get in the door anymore.

202
00:08:56,853 --> 00:09:01,465
So now I spend my days teaching software
that basically forgets everything

203
00:09:01,465 --> 00:09:02,205
by tomorrow.

204
00:09:02,205 --> 00:09:03,516
And that doesn't stay at work either.

205
00:09:03,516 --> 00:09:06,529
By the time I I log off, I've been making
decisions.

206
00:09:06,679 --> 00:09:07,880
And I've talked to you about this before.

207
00:09:07,880 --> 00:09:09,661
I've been making decisions for eight
hours.

208
00:09:09,811 --> 00:09:11,740
And and then my wife asks me, Hey,

209
00:09:11,740 --> 00:09:13,121
do you want to cook or you want me to?

210
00:09:12,611 --> 00:09:15,020
Yeah.

211
00:09:13,962 --> 00:09:15,093
Right, exactly. It's

212
00:09:15,093 --> 00:09:19,467
the first question. And and that that
question is a is a decision tree.

213
00:09:19,909 --> 00:09:22,461
You know, do we do we cook? Are we gonna
do a meal kit?

214
00:09:22,461 --> 00:09:24,363
Do we have leftovers? Are we gonna go get
something?

215
00:09:24,363 --> 00:09:25,464
If it's a meal kit, we've got

216
00:09:25,614 --> 00:09:26,986
four to pick from. If we're going out,

217
00:09:26,986 --> 00:09:28,438
is it restaurant or is it delivery?

218
00:09:28,438 --> 00:09:31,917
If it's delivery, is is it DoorDash or
Uber Eats or or whatever.

219
00:09:31,918 --> 00:09:34,601
Every branch is a another decision.

220
00:09:34,642 --> 00:09:36,304
And I'm just out of decision.

221
00:09:36,454 --> 00:09:38,918
And maybe that should be the first
question is do you have any decisions left

222
00:09:38,918 --> 00:09:41,228
in you today? And you can answer no.

223
00:09:39,555 --> 00:09:41,296
Right. And but the answer's gonna be no,

224
00:09:41,296 --> 00:09:43,098
because all we do is decisions now.

225
00:09:43,098 --> 00:09:43,928
I I essentially

226
00:09:43,506 --> 00:09:43,928
All right.

227
00:09:43,928 --> 00:09:49,301
oversee an army of AI information that I
spend the majority of the morning building

228
00:09:49,301 --> 00:09:52,329
the instruction to get to the end of the
day.

229
00:09:52,329 --> 00:09:54,154
And I'm I'm just decisioned out.

230
00:09:54,831 --> 00:09:55,382
Maybe you should ask

231
00:09:54,995 --> 00:09:55,382
Um

232
00:09:55,382 --> 00:09:56,734
AI to make the decision for you,

233
00:09:56,734 --> 00:09:58,068
but right before you get off work.

234
00:09:58,515 --> 00:09:59,996
But it can't, and that's the problem.

235
00:09:58,769 --> 00:09:59,769
Yeah, like

236
00:09:59,996 --> 00:10:00,617
So

237
00:10:00,617 --> 00:10:02,558
the point, the point I'm making here is
right,

238
00:10:02,558 --> 00:10:04,339
like it just becomes this decision tree.

239
00:10:04,719 --> 00:10:08,430
So when I get off of work, what I actually
want isn't really a screen,

240
00:10:08,651 --> 00:10:10,872
right? I mean, it not a computer screen.

241
00:10:10,872 --> 00:10:13,874
Maybe, maybe watch some TV, maybe play
some music,

242
00:10:14,074 --> 00:10:15,945
you know, maybe, maybe do something else.

243
00:10:15,945 --> 00:10:18,470
But there's always something that has to
get done,

244
00:10:18,470 --> 00:10:19,667
right? And then, of course, you know,

245
00:10:19,667 --> 00:10:21,468
daylight savings just happened again,

246
00:10:21,468 --> 00:10:23,980
despite the fact that we keep talking
about it in Congress and the Senate.

247
00:10:24,130 --> 00:10:27,564
And never pass anything. So you have
exactly twenty two minutes of daylight

248
00:10:27,564 --> 00:10:30,174
to do anything you need to do outside
right before it gets dark.

249
00:10:30,324 --> 00:10:31,346
Right. And everybody

250
00:10:31,079 --> 00:10:31,739
Are you feeling

251
00:10:31,346 --> 00:10:31,777
asks as if

252
00:10:31,777 --> 00:10:34,732
they gave us a free hour, but the reality
is you don't get a free hour because

253
00:10:34,732 --> 00:10:37,667
you spend half of that time fixing all the
clocks and you know,

254
00:10:36,986 --> 00:10:38,815
Exactly. Or mowing

255
00:10:37,817 --> 00:10:38,815
it's just it's

256
00:10:38,815 --> 00:10:41,177
finish mowing the grass for the last time
before the winter,

257
00:10:41,297 --> 00:10:44,906
or something broke that you have to fix or
the weather vane needs needs repairing

258
00:10:44,906 --> 00:10:48,461
or w whatever it is. You you have very
little time to actually do the things

259
00:10:48,461 --> 00:10:52,144
you want. For example, we're recording
this one on a Tuesday and and we have

260
00:10:52,144 --> 00:10:57,117
tomorrow off for the the federal holiday
that that deals with honoring soldiers

261
00:10:57,117 --> 00:10:59,043
and fallen soldiers, I believe.

262
00:10:59,043 --> 00:11:01,945
It's Memorial Day, right? Or is it
Veterans Day?

263
00:11:01,480 --> 00:11:03,057
Veterans Veterans Day.

264
00:11:01,945 --> 00:11:03,336
I get the two of mixed up. Okay,

265
00:11:03,336 --> 00:11:04,056
Veterans Day.

266
00:11:04,056 --> 00:11:06,678
So people who served, we're honoring them
for people who served.

267
00:11:06,678 --> 00:11:10,520
And and I already have a long list of
things that I have to do because I don't

268
00:11:10,520 --> 00:11:13,303
really have to, the time to do them before
it gets dark out.

269
00:11:13,483 --> 00:11:15,414
Anyway. So that's what it does to me,

270
00:11:15,654 --> 00:11:18,515
right? that's what what working with AI
every day,

271
00:11:18,515 --> 00:11:21,098
despite the fact that I don't necessarily
want to,

272
00:11:21,508 --> 00:11:24,836
that's l I think it's time we're gonna
talk about kind of what it does

273
00:11:24,836 --> 00:11:29,149
to all of us. So let's start with
something fun something small.

274
00:11:29,299 --> 00:11:30,872
Your friend sends you a really,

275
00:11:30,882 --> 00:11:35,428
thoughtful text message. Then you find out
that ChatGPT helped write it.

276
00:11:35,578 --> 00:11:37,130
How does that land with you, Will?

277
00:11:37,280 --> 00:11:40,203
I'm actually okay with it because I
understand I I treat the other person like

278
00:11:40,203 --> 00:11:41,634
I would look at myself doing that.

279
00:11:42,069 --> 00:11:42,367
Mm-hmm.

280
00:11:42,195 --> 00:11:42,715
Like

281
00:11:42,715 --> 00:11:45,557
there there may have been a lot of thought
that went into the way they prompted it.

282
00:11:45,707 --> 00:11:48,259
Which, is our it maybe just a way to
articulate something if they don't feel

283
00:11:48,580 --> 00:11:51,127
like they're comfortable saying it their
in their own words,

284
00:11:51,127 --> 00:11:53,359
you know, things like that. So I I tend to
give them the benefit of the doubt,

285
00:11:53,359 --> 00:11:55,571
but I understand why somebody else might
see it differently.

286
00:11:55,721 --> 00:11:58,302
Okay. No, I I there's no right or wrong
answer here,

287
00:11:58,302 --> 00:12:00,767
right? I I just want to get your your
viewpoints here.

288
00:12:01,568 --> 00:12:03,050
and then I'll follow it up with research.

289
00:12:03,939 --> 00:12:04,711
Mm-hmm.

290
00:12:04,711 --> 00:12:10,316
Like I usually do. So researchers at Ohio
State basically tested this this entire

291
00:12:10,316 --> 00:12:14,008
concept. So people imagined getting a
thoughtful message from a friend,

292
00:12:14,008 --> 00:12:16,490
and some of them were told the friend had
AI help writing it.

293
00:12:16,570 --> 00:12:20,732
the people were less happy with the
friendship and less sure the friend even

294
00:12:20,732 --> 00:12:21,413
liked them.

295
00:12:21,413 --> 00:12:25,195
And the reason really came down to effort
or or the way that they perceived effort.

296
00:12:25,195 --> 00:12:28,057
They figured basically the friend hadn't
put any effort in.

297
00:12:28,057 --> 00:12:31,709
Now you you built it that into your your
general response,

298
00:12:31,709 --> 00:12:35,592
which was, well, maybe it took four hours
to write the prompt to figure out exactly

299
00:12:35,592 --> 00:12:37,793
how to say it or if it was a sensitive
topic or whatever.

300
00:12:38,229 --> 00:12:43,131
but help from a person also counted
against them in this exact same study here

301
00:12:43,131 --> 00:12:44,441
at Ohio State. So

302
00:12:44,591 --> 00:12:46,932
When the friend got help from a stranger,

303
00:12:46,932 --> 00:12:52,765
either in an online writing group or or
through a text message from another friend

304
00:12:52,765 --> 00:12:55,037
or somebody that knew them, they were just
as put off.

305
00:12:55,037 --> 00:12:58,920
So from what they could tell, the words
were basically fine.

306
00:12:59,070 --> 00:13:02,612
What they really seemed to be looking for
was kind of proof that you tried.

307
00:13:02,762 --> 00:13:03,603
Right. I

308
00:13:03,510 --> 00:13:04,701
And that's fine, right? So

309
00:13:03,712 --> 00:13:04,712
want to see effort.

310
00:13:04,851 --> 00:13:09,124
I think I think one of the things that the
study points out is that they they just

311
00:13:09,124 --> 00:13:11,036
kinda want to them to be honest about it,

312
00:13:11,036 --> 00:13:13,858
right? Like, hey, I I had ChatGPT
write this or I had,

313
00:13:13,858 --> 00:13:16,559
you know, Claude write this because with a
a general reason.

314
00:13:16,920 --> 00:13:22,894
And I you know, I don't I didn't read the
every anything else except the conclusion

315
00:13:22,894 --> 00:13:26,576
of the study, which really was n that it
was equivalent,

316
00:13:26,616 --> 00:13:28,958
a trust problem if you if if you

317
00:13:29,408 --> 00:13:33,158
used another source than your own mind or
self to write a text message.

318
00:13:33,308 --> 00:13:34,551
But uh the

319
00:13:33,791 --> 00:13:34,701
Right. I

320
00:13:34,701 --> 00:13:35,232
go ahead.

321
00:13:35,232 --> 00:13:38,704
was just gonna say my my experience in
this area is mostly more with songwriting.

322
00:13:38,704 --> 00:13:39,895
Um, because I

323
00:13:39,549 --> 00:13:39,895
Uh-huh.

324
00:13:39,895 --> 00:13:42,336
d obviously I I don't write songs
necessarily,

325
00:13:42,336 --> 00:13:45,828
although I I've written a significant
portion of what I put into the AI music

326
00:13:45,828 --> 00:13:48,969
generators. But also I don't I don't sing
at at all,

327
00:13:48,989 --> 00:13:53,532
pretty much. So you know, it's it's nobody
I've I've not experienced anybody like it.

328
00:13:53,532 --> 00:13:56,413
First of all, I've had some very emotional
responses to the songs I've written,

329
00:13:56,563 --> 00:13:57,534
but I've never had anybody say,

330
00:13:57,534 --> 00:13:58,277
you know, well

331
00:13:58,277 --> 00:14:00,510
when they find out it's not my voice or
they realize it's not my voice,

332
00:14:00,510 --> 00:14:02,623
they're they're not like, well this is not
really you.

333
00:14:02,623 --> 00:14:06,138
You didn't put in the effort. You didn't
learn the guitar and take singing lessons,

334
00:14:06,138 --> 00:14:08,090
your vocal lessons to to do this for me.

335
00:14:08,090 --> 00:14:11,255
So it doesn't mean anything. But I I think
there maybe there's more a perception

336
00:14:11,255 --> 00:14:13,247
of effort than there would be in like a
written

337
00:14:12,353 --> 00:14:13,394
So tell me tell me a little

338
00:14:13,394 --> 00:14:16,638
bit more about that because I obviously I
know a little bit more about it than

339
00:14:16,638 --> 00:14:18,109
our audience is, but do you do

340
00:14:17,803 --> 00:14:18,109
Mm-hmm.

341
00:14:18,109 --> 00:14:21,672
you wanna share a little bit about your
writing process or or how you go about

342
00:14:21,672 --> 00:14:22,012
that?

343
00:14:22,012 --> 00:14:25,255
Sure. So I tend to, and this is one of the
big debates,

344
00:14:25,255 --> 00:14:27,837
but I I tend to have AI write the first
draft or something.

345
00:14:27,837 --> 00:14:29,889
So what I'll do is I'll say, I want it to
sound like,

346
00:14:29,889 --> 00:14:33,505
this genre of music, and some of the
tropes of these genres or whatever,

347
00:14:33,505 --> 00:14:36,728
you know, or maybe maybe be inspired by a
particular song,

348
00:14:36,878 --> 00:14:37,578
but but

349
00:14:36,983 --> 00:14:37,578
Mm-hmm.

350
00:14:37,578 --> 00:14:39,120
in this new context or whatever.

351
00:14:39,120 --> 00:14:43,543
And so I'll do that. and then I will
whatever it writes first is usually high

352
00:14:43,543 --> 00:14:46,776
garbage. So I'll tell it to go through and
like do three revisions.

353
00:14:46,926 --> 00:14:48,207
First of I'll fix a couple of things,

354
00:14:48,207 --> 00:14:51,568
I'll go and say no now review it three ti
three different times if you're like

355
00:14:51,768 --> 00:14:55,500
a top music producer hearing it for the
first time and making your suggested

356
00:14:55,500 --> 00:14:56,891
changes. And

357
00:14:56,540 --> 00:14:56,891
Mm-hmm.

358
00:14:56,891 --> 00:14:59,342
so it'll change change things up and
improve on it a little bit.

359
00:14:59,342 --> 00:15:01,992
And then I'll go in and fix some other
things and tweak it some more.

360
00:15:01,992 --> 00:15:03,013
Then I'll run it through the AI.

361
00:15:03,013 --> 00:15:05,075
That's just to get the actual l lyrics.

362
00:15:05,225 --> 00:15:10,077
And then I take it over to the AI music
generator and run those through with

363
00:15:10,077 --> 00:15:12,140
the style very you know, settings and
things like that.

364
00:15:12,140 --> 00:15:13,479
So I mean it takes me a good

365
00:15:12,615 --> 00:15:13,720
Do you feel comfortable telling

366
00:15:13,720 --> 00:15:15,968
the tools, the specific like platforms?

367
00:15:14,943 --> 00:15:16,342
Sure. Sure.

368
00:15:16,582 --> 00:15:18,784
Suno.ai is the one that I prefer to use.

369
00:15:18,841 --> 00:15:19,474
I know there

370
00:15:19,115 --> 00:15:19,474
Okay.

371
00:15:19,474 --> 00:15:21,055
have been some that have come out more
recently,

372
00:15:21,055 --> 00:15:22,506
but that's I've gotten comfortable with
that one.

373
00:15:22,506 --> 00:15:24,137
I know how to get my results that I want
from that one.

374
00:15:24,137 --> 00:15:26,116
So I'm I that's what I'm with for right
now.

375
00:15:26,856 --> 00:15:29,748
but you know, it's it's so I take it over
there and then there are settings

376
00:15:29,748 --> 00:15:32,282
for like, you know, temperature settings
and um,

377
00:15:32,282 --> 00:15:36,905
you know, you can you can set like the the
the gender of your vocalist

378
00:15:36,905 --> 00:15:40,938
and or you know, can do multiple vocalists
or choir background or chorus or whatever.

379
00:15:41,048 --> 00:15:44,982
and then so I mean it honestly it takes me
probably longer to do that than it would

380
00:15:44,982 --> 00:15:46,803
to write a heartfelt note of like,

381
00:15:46,953 --> 00:15:49,304
you know, sorrow over a friend's loss or
something like that.

382
00:15:49,304 --> 00:15:51,256
But for them to have something that they
go back to,

383
00:15:51,256 --> 00:15:52,827
that's the thing. Like my secretary,

384
00:15:52,902 --> 00:15:55,684
my legal secretary that I worked with when
I first got out of law school,

385
00:15:55,684 --> 00:15:57,125
she she and I became really close.

386
00:15:57,275 --> 00:16:00,167
And I wrote her a personalized song about
some of the things that she knew,

387
00:16:00,167 --> 00:16:01,238
like that we had been through together,

388
00:16:01,238 --> 00:16:02,819
and it reminded her of those experiences.

389
00:16:02,969 --> 00:16:05,591
And she went back and listened to it like
four or five times,

390
00:16:05,591 --> 00:16:07,402
like crying every time, just from how

391
00:16:07,639 --> 00:16:09,370
thoughtful it was and how much it meant to
her.

392
00:16:09,520 --> 00:16:11,371
That kind of stuff is just like it's worth
the extra time,

393
00:16:11,371 --> 00:16:13,423
but it is a decent amount of effort on my
part.

394
00:16:13,423 --> 00:16:15,824
I don't discuss that with them usually
unless they ask,

395
00:16:15,824 --> 00:16:18,125
but it can take a couple hours or more.

396
00:16:18,185 --> 00:16:20,176
you were joking about taking a couple
hours to write the prompt

397
00:16:20,176 --> 00:16:22,798
for a the thoughtful note. But this really

398
00:16:22,325 --> 00:16:22,798
Right.

399
00:16:22,798 --> 00:16:25,230
does take me a couple hours to like I
guess I'll generate 10,

400
00:16:25,230 --> 00:16:28,043
20 different versions of the song before I
get the one I'm happy with.

401
00:16:27,182 --> 00:16:29,989
Mm-hmm. And iterate on it.

402
00:16:28,103 --> 00:16:29,622
So um

403
00:16:30,139 --> 00:16:30,599
yeah.

404
00:16:30,599 --> 00:16:35,260
Okay. So the also the University of
Arizona did another another study here.

405
00:16:35,335 --> 00:16:38,616
researchers ran thirteen experiments with
more than 5,000 people,

406
00:16:38,696 --> 00:16:40,877
much bigger sample size than than Ohio's.

407
00:16:40,952 --> 00:16:46,154
and the people who admitted that they used
AI to do their work were trusted less than

408
00:16:46,154 --> 00:16:49,205
the people who said nothing. Um

409
00:16:49,048 --> 00:16:49,513
Wow.

410
00:16:49,513 --> 00:16:53,184
and that the the only thing worse than
admitting it was getting caught.

411
00:16:53,334 --> 00:16:55,305
If if someone else exposed you,

412
00:16:55,305 --> 00:16:59,848
trust dropped even further. And that held
even for people who use AI like

413
00:16:59,848 --> 00:17:01,889
all the time themselves. Now, I will tell
you,

414
00:17:01,889 --> 00:17:03,990
I don't personally experience this at all.

415
00:17:04,183 --> 00:17:07,051
I we talk about AI, we talk about what
we're using,

416
00:17:07,051 --> 00:17:08,422
we talk about how we're prompting it,

417
00:17:08,422 --> 00:17:13,055
we we talk about these things at work
because they understand they're expecting

418
00:17:13,303 --> 00:17:18,478
a a single individual to essentially
produce three people worth of work like

419
00:17:18,598 --> 00:17:20,139
Yeah, and they they don't want like a

420
00:17:18,818 --> 00:17:20,249
in a given day, you know.

421
00:17:20,398 --> 00:17:22,731
they don't want a human they don't want
human sounding code.

422
00:17:22,731 --> 00:17:25,144
'Cause that's that's not really whatever
that might mean.

423
00:17:23,559 --> 00:17:25,940
Right. Like they don't care if it sounds
human,

424
00:17:25,144 --> 00:17:26,505
It's like yeah.

425
00:17:25,940 --> 00:17:26,670
but humans

426
00:17:26,670 --> 00:17:28,640
care if it sounds human. Right?

427
00:17:28,640 --> 00:17:31,370
Like I'm not writing a an executive
summary to,

428
00:17:31,370 --> 00:17:34,871
you know, to leadership with a heartfelt
plea for for a donation.

429
00:17:34,871 --> 00:17:36,982
Like I'm I'm writing software that does a
thing.

430
00:17:37,122 --> 00:17:39,183
Does it do the thing? Yes. Is it secure?

431
00:17:39,183 --> 00:17:41,133
Yes. Can is it repeatable? Yes.

432
00:17:41,133 --> 00:17:42,854
Like that's I mean, that's to me,

433
00:17:42,854 --> 00:17:45,734
AI is a perfect example of of exactly
code.

434
00:17:45,884 --> 00:17:48,742
b because uh I I actually wrote a song.

435
00:17:49,312 --> 00:17:51,472
That I wrote the lyrics to myself.

436
00:17:51,472 --> 00:17:57,245
I put it in Suno AI. And it was um
basically about the I I want or right

437
00:17:57,245 --> 00:18:03,469
now we have AI writing songs and and
making art while I do the dishes and fold

438
00:18:03,469 --> 00:18:05,104
the laundry, and I want to swap that,

439
00:18:04,835 --> 00:18:06,248
Mm. Right.

440
00:18:05,104 --> 00:18:06,331
you know. Like I

441
00:18:06,331 --> 00:18:09,703
want the AI to do the dishes and the
laundry while while so I have more time

442
00:18:09,703 --> 00:18:10,883
to write songs and make music.

443
00:18:11,338 --> 00:18:12,559
but the

444
00:18:13,029 --> 00:18:16,155
For me, I guess, I I i in my workplace,

445
00:18:16,155 --> 00:18:18,223
I use it.

446
00:18:18,393 --> 00:18:20,044
I would use a team of individuals,

447
00:18:20,044 --> 00:18:21,465
right? I I'm an I'm an architect,

448
00:18:21,465 --> 00:18:26,157
right? So I I would be expected to explain
a design down to other engineers,

449
00:18:26,237 --> 00:18:30,449
and then the other engineers would would
then make that true based upon the design

450
00:18:30,449 --> 00:18:32,813
that I had researched and grounded in in
reality.

451
00:18:32,953 --> 00:18:35,254
And so I I find myself in the the same
position.

452
00:18:35,989 --> 00:18:38,040
but not I I don't want to get too far off
here.

453
00:18:38,040 --> 00:18:41,191
We have a couple of other points from this
specific study.

454
00:18:41,806 --> 00:18:42,266
so

455
00:18:42,266 --> 00:18:47,899
Apparently trust drops even further for
people who admit to also using AI than

456
00:18:47,919 --> 00:18:49,680
at the same all the time themselves.

457
00:18:49,830 --> 00:18:52,345
So they trust people less who use AI,

458
00:18:52,345 --> 00:18:54,806
even when they use AI all the time.

459
00:18:54,956 --> 00:18:55,391
Right.

460
00:18:55,391 --> 00:18:57,783
So basically, you lose if you admit that
you're using it.

461
00:18:57,843 --> 00:18:58,784
You lose worse

462
00:18:58,230 --> 00:18:58,784
Mm-hmm.

463
00:18:58,784 --> 00:19:00,425
if you hide it and get caught.

464
00:19:00,575 --> 00:19:04,648
And the safest thing is hiding it and
hoping that you don't get caught.

465
00:19:03,258 --> 00:19:05,762
Not getting caught. Yeah. All right.

466
00:19:05,669 --> 00:19:05,989
Right?

467
00:19:05,989 --> 00:19:08,011
Like, or or editing it, right?

468
00:19:08,161 --> 00:19:09,322
So we spend

469
00:19:08,314 --> 00:19:09,571
Yeah, get rid of those Claudisms.

470
00:19:09,721 --> 00:19:11,743
exactly we gotta there's a whole website,

471
00:19:11,743 --> 00:19:17,127
Claudisms.ai, that explains in it in a
prompt that you can then take into Claude

472
00:19:15,992 --> 00:19:17,985
Hmm. Yeah.

473
00:19:17,277 --> 00:19:18,158
and add to your

474
00:19:18,158 --> 00:19:20,340
prompt to make sure it doesn't do those
things to you.

475
00:19:20,420 --> 00:19:22,711
It's it's crazy. I mean uh

476
00:19:22,861 --> 00:19:26,575
I I see benefits and and detriments all
the time about it.

477
00:19:26,575 --> 00:19:30,063
But so now now basically if you multiply
that by everybody,

478
00:19:30,063 --> 00:19:34,067
right? And everybody's wondering every
about everybody's communication is this

479
00:19:34,067 --> 00:19:35,959
AI? Did they use AI? Is it you know,

480
00:19:35,959 --> 00:19:37,760
whatever? and I'm gonna point out,

481
00:19:37,760 --> 00:19:40,471
right? I I told everybody that I use it
every day.

482
00:19:40,972 --> 00:19:43,444
and according to you know, this research,

483
00:19:43,444 --> 00:19:46,047
that has cost me trust with our listeners,

484
00:19:46,047 --> 00:19:46,247
with

485
00:19:46,104 --> 00:19:46,286
I gonna

486
00:19:46,286 --> 00:19:47,603
say we just lost trust. Yeah.

487
00:19:46,517 --> 00:19:47,758
Maybe our friendships.

488
00:19:47,758 --> 00:19:49,120
Yeah, and and that's possible.

489
00:19:49,442 --> 00:19:53,436
now we do use AI for some research on this
show.

490
00:19:53,897 --> 00:20:00,344
but if that research is also grounded and
validated and verified um by human people,

491
00:20:00,965 --> 00:20:02,417
it's either usually me or Will.

492
00:20:02,417 --> 00:20:04,109
Um

493
00:20:03,440 --> 00:20:04,161
Right. Which

494
00:20:04,161 --> 00:20:05,634
is not to say that it's better
necessarily,

495
00:20:05,634 --> 00:20:08,087
especially when I'm doing it, but it is
done by humans.

496
00:20:07,380 --> 00:20:08,237
Exactly.

497
00:20:08,237 --> 00:20:14,190
Exactly. I and I I do also use local AI
quite a bit to do things that

498
00:20:14,230 --> 00:20:18,132
I otherwise don't want to do or would save
me time to do.

499
00:20:18,282 --> 00:20:21,564
Um but it it it also happens with total
strangers.

500
00:20:21,623 --> 00:20:26,747
In in twenty twenty five a study in PNAS
Nexus had more than thirty five hundred

501
00:20:26,747 --> 00:20:28,988
people play classic money games,

502
00:20:28,988 --> 00:20:32,610
the the kind where you like decide how
much to share and how much to trust

503
00:20:32,610 --> 00:20:33,731
the other player and so

504
00:20:33,881 --> 00:20:38,113
In some rounds, they knew ChatGPT was
making the choices on behalf of a real

505
00:20:38,113 --> 00:20:41,075
person, and the money still went to that
real person,

506
00:20:41,075 --> 00:20:44,147
right? People got stingier, less fair,

507
00:20:44,147 --> 00:20:46,268
less trusting, less willing to cooperate.

508
00:20:46,268 --> 00:20:49,960
They assumed the AI side would be out for
itself.

509
00:20:50,110 --> 00:20:53,412
So they actually held back from the even
the very beginning of the game.

510
00:20:53,562 --> 00:20:55,844
And there was a real human on the other
end the entire time,

511
00:20:55,844 --> 00:20:59,626
right? They treated that person worse
because a machine was in the middle

512
00:20:59,626 --> 00:21:00,336
of them.

513
00:21:00,336 --> 00:21:01,464
And the playing of the game.

514
00:21:01,614 --> 00:21:03,811
so at work, there's there's a new word for
this,

515
00:21:03,811 --> 00:21:07,343
right? It's called workslop. I'm sure
it'll be added to the dictionary

516
00:21:07,343 --> 00:21:09,194
any day now, but that's what it's being
referred

517
00:21:08,760 --> 00:21:09,194
Right.

518
00:21:09,194 --> 00:21:14,567
to online. And it's basically just AI work
that looks polished but doesn't actually

519
00:21:14,567 --> 00:21:19,730
say anything. the report that actually
sounds great until you try to use it.

520
00:21:19,730 --> 00:21:23,223
And and you know, if any of have ever
typed a thing into into

521
00:21:22,740 --> 00:21:23,223
Yeah.

522
00:21:23,223 --> 00:21:25,594
ChatGPT, you know exactly what I'm
talking about.

523
00:21:25,744 --> 00:21:30,237
The email is four paragraphs long and it
never actually gets to the question that

524
00:21:29,775 --> 00:21:30,237
Right.

525
00:21:30,237 --> 00:21:33,004
that you're trying to answer. Have you
gotten any of these?

526
00:21:32,176 --> 00:21:33,142
Or does that tell us about all

527
00:21:33,142 --> 00:21:34,831
the training data that they used?

528
00:21:35,424 --> 00:21:36,358
Right. Have you gotten

529
00:21:35,798 --> 00:21:36,358
Like.

530
00:21:36,358 --> 00:21:39,552
any of these? Do you know the the or have
you seen this

531
00:21:38,427 --> 00:21:39,448
Yeah, I've

532
00:21:39,552 --> 00:21:40,506
from other people?

533
00:21:40,506 --> 00:21:44,488
I've gotten some results like that from
the presentations and things especially

534
00:21:45,015 --> 00:21:45,362
Mm-hmm.

535
00:21:45,089 --> 00:21:45,379
where

536
00:21:45,379 --> 00:21:46,920
it looks really good but then you're like,

537
00:21:46,920 --> 00:21:49,792
Well that doesn't actually say anything or
like website design.

538
00:21:48,923 --> 00:21:49,923
Right.

539
00:21:50,929 --> 00:21:51,463
well yeah.

540
00:21:51,463 --> 00:21:56,077
So the same the same Hubstaff survey that
we talked about earlier found that about

541
00:21:56,077 --> 00:22:02,583
53% of people who use AI at work get AI
work from a coworker that needs serious

542
00:22:02,583 --> 00:22:04,034
fixing at least once a month.

543
00:22:04,184 --> 00:22:05,896
Now that that costs more than time.

544
00:22:06,055 --> 00:22:09,189
In the survey that coined the word work
slop,

545
00:22:09,820 --> 00:22:11,752
which was published in Harvard Business
Review,

546
00:22:11,953 --> 00:22:16,467
about half of the people who got workslop
said they saw the coworker who sent

547
00:22:16,467 --> 00:22:20,070
it as either less capable or less reliable
afterwards,

548
00:22:20,220 --> 00:22:24,906
with forty-two percent who actually said
that they were less trustworthy after

549
00:22:24,906 --> 00:22:25,366
the fact.

550
00:22:25,366 --> 00:22:25,814
Wow.

551
00:22:25,814 --> 00:22:28,015
Now take that same suspicion, right,

552
00:22:28,015 --> 00:22:31,827
and put it in a classroom. Teachers are
running student essays through

553
00:22:31,827 --> 00:22:32,697
AI detectors.

554
00:22:32,591 --> 00:22:33,077
Mm-hmm.

555
00:22:33,077 --> 00:22:35,478
you know, I in a college town,

556
00:22:35,478 --> 00:22:38,690
you'll you'll hear plenty of this from
from people who are in the the college

557
00:22:38,690 --> 00:22:40,871
world. They're doing it on the regular.

558
00:22:40,871 --> 00:22:46,043
I have a a good friend now who is working
directly in the office that handles

559
00:22:46,043 --> 00:22:49,835
cheating and you know AI copying and
plagiarism.

560
00:22:48,747 --> 00:22:49,768
Plagiarism, yeah.

561
00:22:49,835 --> 00:22:51,006
Yeah, those sorts of things.

562
00:22:51,426 --> 00:22:54,889
And he he has told me quite a few
interesting stories about that.

563
00:22:54,889 --> 00:22:57,511
And hopefully he's listening every now and
then he said he does.

564
00:22:58,027 --> 00:23:03,051
but a Stanford team tested seven popular
detectors on essays written by people

565
00:23:03,051 --> 00:23:04,732
who learned English as a second language,

566
00:23:04,732 --> 00:23:05,933
or the the TOEFL

567
00:23:05,762 --> 00:23:06,113
Yeah.

568
00:23:06,113 --> 00:23:09,646
English exam, ESL. I mean, this was real
real human writing,

569
00:23:09,646 --> 00:23:14,361
right? On average, the detectors flagged
61% of those as AI.

570
00:23:15,281 --> 00:23:15,779
That's brutal.

571
00:23:15,702 --> 00:23:16,112
Now

572
00:23:16,112 --> 00:23:21,377
One flagged almost all of them on essays
by native English speakers and that were

573
00:23:21,377 --> 00:23:22,878
essentially close to perfect. Now,

574
00:23:22,878 --> 00:23:26,311
the reason is that people writing in a
second language tend to use simpler,

575
00:23:26,472 --> 00:23:28,133
more predictable word choices,

576
00:23:28,133 --> 00:23:33,968
right? And that's basically what AI
detectors are taught and trained to read

577
00:23:33,989 --> 00:23:34,599
and see.

578
00:23:34,599 --> 00:23:38,232
It falls hardest on the kid who basically
already gets looked at funny,

579
00:23:38,232 --> 00:23:40,444
right? In in a Common Sense Media survey,

580
00:23:40,444 --> 00:23:44,808
about one in teen, one in ten teenagers
said a teacher had flagged their work

581
00:23:44,808 --> 00:23:46,799
as AI when it actually wasn't.

582
00:23:46,949 --> 00:23:49,071
And that gets even worse for Black teens,

583
00:23:49,221 --> 00:23:51,182
which apparently was about one in five.

584
00:23:51,332 --> 00:23:53,494
For white teens, only about 7%,

585
00:23:53,494 --> 00:23:59,849
but 20% for Black children accused of
using AI in reports and paperwork

586
00:23:59,849 --> 00:24:00,429
and schoolwork.

587
00:24:00,429 --> 00:24:04,700
So, what that shows me is that kids in
class right now are feeling this.

588
00:24:04,850 --> 00:24:09,430
There was a survey of college students
done this spring by Inside Higher Ed that

589
00:24:09,430 --> 00:24:13,861
found that nearly three out of four are
worried about being wrongly accused

590
00:24:13,861 --> 00:24:18,901
of using AI in assignments. So essentially
the kid who actually did the work gets

591
00:24:18,901 --> 00:24:20,762
accused, the teacher can't tell,

592
00:24:20,912 --> 00:24:23,042
and both of them walk away, you know,

593
00:24:23,042 --> 00:24:25,023
with the trust kind of broken between both
of them.

594
00:24:25,173 --> 00:24:28,542
And also you've got the fact that they're
changing their writing before they turn

595
00:24:28,542 --> 00:24:30,244
it in based on the fact that they're
worried about it,

596
00:24:30,244 --> 00:24:31,675
right? So if they

597
00:24:31,303 --> 00:24:31,675
Right.

598
00:24:31,675 --> 00:24:33,788
if they think they wrote something that
sounds too AI,

599
00:24:33,788 --> 00:24:35,901
they're gonna fix it, which messes up

600
00:24:34,802 --> 00:24:35,813
They're gonna make it simpler.

601
00:24:35,901 --> 00:24:37,022
with your evaluation, yeah.

602
00:24:37,172 --> 00:24:42,759
Exactly. now, you know, I I guess the
question we really should be asking about

603
00:24:42,759 --> 00:24:46,825
these sorts of things is who comes out
ahead when nobody trusts anybody

604
00:24:46,825 --> 00:24:48,197
or anything, right?

605
00:24:48,865 --> 00:24:50,250
Nobody. That's my

606
00:24:49,900 --> 00:24:50,250
Right.

607
00:24:50,250 --> 00:24:52,227
take on that. Trust is trust is essential.

608
00:24:52,377 --> 00:24:54,098
I I would a hundred percent agree with
that.

609
00:24:54,098 --> 00:24:59,332
Now, once once everybody basically knows
that you know you can be faked,

610
00:24:59,482 --> 00:25:01,844
that's fake or fake news, right?

611
00:25:02,405 --> 00:25:04,986
Uh wink wink, nudge nudge basically

612
00:25:02,627 --> 00:25:03,627
Right.

613
00:25:04,986 --> 00:25:07,378
becomes an excuse that anybody can use,

614
00:25:07,378 --> 00:25:09,068
right? So in March, when

615
00:25:08,547 --> 00:25:09,068
Right.

616
00:25:09,068 --> 00:25:13,521
we started this war that's not a war that
is a war but isn't a war with Iran,

617
00:25:14,322 --> 00:25:17,674
you know, there were published AI made
videos everywhere.

618
00:25:17,824 --> 00:25:18,024
So

619
00:25:18,024 --> 00:25:21,216
that when Benjamin Netanyahu, the war
criminal,

620
00:25:21,366 --> 00:25:23,667
got called an AI fake over the internet,

621
00:25:23,667 --> 00:25:25,217
the New York Times actually wrote about
it.

622
00:25:25,217 --> 00:25:28,979
Snopes has a a an article out out there to
verify that the actual video

623
00:25:28,979 --> 00:25:33,832
was authentic. There's a name for this
that researchers call the liar's dividend.

624
00:25:33,982 --> 00:25:35,982
Hmm. That's not that's a new one to me.

625
00:25:36,132 --> 00:25:41,005
Right. I I I I have I really enjoyed
writing this episode and I really enjoyed

626
00:25:41,005 --> 00:25:45,339
the research around it because I gave this
a couple of days of of pretty legitimate

627
00:25:45,339 --> 00:25:50,504
research. what works better for a
politician caught in a scandal?

628
00:25:50,654 --> 00:25:53,286
Apologizing or calling it fake?

629
00:25:53,436 --> 00:25:54,350
Calling it fake.

630
00:25:54,350 --> 00:25:57,413
So a study in the American Political
Science Review tested it with more than

631
00:25:57,413 --> 00:26:00,316
15,000 Americans, even bigger sample
size.

632
00:26:00,316 --> 00:26:02,017
See, we're going up in sample sizes here.

633
00:26:01,413 --> 00:26:02,678
Mm. Mm-hmm.

634
00:26:02,167 --> 00:26:02,858
We were at

635
00:26:02,858 --> 00:26:05,951
1,000, 3,000, 5,000. Now we're at 15,000.

636
00:26:05,951 --> 00:26:09,665
So people read about real political
scandals all the time.

637
00:26:09,866 --> 00:26:11,097
you know, this administration.

638
00:26:11,097 --> 00:26:12,449
They're just, it's everywhere.

639
00:26:12,849 --> 00:26:14,800
And then the politician responded,

640
00:26:14,800 --> 00:26:18,384
claiming it was fake or kept more support.

641
00:26:18,534 --> 00:26:18,764
By

642
00:26:18,764 --> 00:26:21,386
calling it fake than actually just
apologizing,

643
00:26:21,536 --> 00:26:24,908
because it basically kept more support
than ignoring it.

644
00:26:24,908 --> 00:26:25,999
So you have to do something, right?

645
00:26:25,999 --> 00:26:28,048
They see it as I have to do something,

646
00:26:28,048 --> 00:26:30,039
because doing nothing is worse.

647
00:26:30,189 --> 00:26:33,121
So either I need to apologize,

648
00:26:33,121 --> 00:26:34,872
admit that it did it, apologize,

649
00:26:34,872 --> 00:26:37,464
or just say, that's fake news and dismiss
it,

650
00:26:37,614 --> 00:26:40,215
which you know what we're essentially
seeing now.

651
00:26:40,365 --> 00:26:41,366
There is a smaller

652
00:26:40,485 --> 00:26:42,361
Apparently Jensen Huang didn't read that
study.

653
00:26:42,867 --> 00:26:43,317
right.

654
00:26:43,317 --> 00:26:45,659
A smaller study in Germany went a little
further,

655
00:26:45,659 --> 00:26:48,451
smaller population, different culture
entirely.

656
00:26:48,461 --> 00:26:53,096
they had people watch real footage of two
American legislators.

657
00:26:53,096 --> 00:26:57,865
I'm not gonna say which ones. but when the
the politician falsely claimed the video

658
00:26:57,865 --> 00:27:02,058
was a deepfake, people actually rated
them as a better leader than when they

659
00:27:02,058 --> 00:27:02,979
apologized.

660
00:27:02,979 --> 00:27:03,099
So

661
00:27:03,099 --> 00:27:05,800
every time somebody says, it's all it's
probably AI,

662
00:27:05,950 --> 00:27:10,121
it gets a little easier for the next
person who gets caught on the record

663
00:27:10,121 --> 00:27:12,098
to just meet me too. It's fake,

664
00:27:12,098 --> 00:27:15,885
fake news. And really, AI isn't really
helping us be more honest with each other

665
00:27:15,885 --> 00:27:20,406
either. So this March, a team at Stanford
published a study in Science magazine.

666
00:27:20,556 --> 00:27:23,957
They took 2,000 posts from the Reddit
forum,

667
00:27:24,078 --> 00:27:25,098
Am I the Asshole?

668
00:27:25,248 --> 00:27:31,491
A I T A, you can find it under Reddit.com
slash R slash A I T A in all caps.

669
00:27:31,641 --> 00:27:36,413
Only the ones where the crowd voted that
the poster was in the wrong.

670
00:27:36,533 --> 00:27:41,284
And then they asked 11 of the biggest AI
chatbots who was wrong.

671
00:27:41,905 --> 00:27:44,786
now basically they 2,000 posts,

672
00:27:44,986 --> 00:27:49,168
all where the majority of people said that
they were objectively wrong,

673
00:27:49,318 --> 00:27:51,369
or they were the asshole in this scenario.

674
00:27:51,519 --> 00:27:57,253
Against eleven chatbots and they sided
with the wrong person 51% of the time.

675
00:27:57,403 --> 00:28:01,972
Is that is that excuse me, is that
attributed to sycophancy or what what

676
00:28:01,972 --> 00:28:03,155
do they theorize?

677
00:28:03,002 --> 00:28:03,353
We'll

678
00:28:03,353 --> 00:28:04,885
we'll talk a little bit more about that,

679
00:28:04,885 --> 00:28:08,950
but the I'm just I'm laying out the data.

680
00:28:08,950 --> 00:28:10,903
The data is in things

681
00:28:09,755 --> 00:28:10,885
Okay. Got it.

682
00:28:10,903 --> 00:28:14,428
that we know for a fact, society judged as
asshole is.

683
00:28:14,428 --> 00:28:15,730
You're the asshole in this one.

684
00:28:15,880 --> 00:28:19,684
AI says you're the asshole forty nine
percent of the time.

685
00:28:19,834 --> 00:28:20,246
Right.

686
00:28:20,246 --> 00:28:20,406
So

687
00:28:20,406 --> 00:28:23,488
after a single conversation with one of
these chatbots,

688
00:28:23,488 --> 00:28:27,911
people were more convinced that they were
right and less willing to apologize

689
00:28:27,911 --> 00:28:30,992
or patch things up with the person they
actually had to fight with.

690
00:28:31,142 --> 00:28:34,133
They also liked the the chatbot a lot
better.

691
00:28:34,283 --> 00:28:38,135
And then they said they'd come back to it
when they were told that they were right.

692
00:28:38,285 --> 00:28:41,177
Now, ChatGPT, OpenAI, they all know.

693
00:28:41,587 --> 00:28:43,312
Why that happens? It's in the prompt.

694
00:28:43,312 --> 00:28:47,415
OpenAI said one of its updates learned
partly from the thumbs up button.

695
00:28:47,565 --> 00:28:54,159
We like being told we're right to the
degree that we trained AI to constantly

696
00:28:54,159 --> 00:28:55,159
tell us we're right.

697
00:28:55,282 --> 00:28:59,239
I'm conflicted here because I'm okay if
yes men lose their jobs to

698
00:28:58,832 --> 00:28:59,239
Uh-huh.

699
00:28:59,239 --> 00:29:02,816
AI, but I I don't want them going and
finding more articulate yes men online

700
00:29:02,816 --> 00:29:03,858
to give them better arguments.

701
00:29:04,008 --> 00:29:05,437
And that is a real risk.

702
00:29:05,587 --> 00:29:10,645
So that's kind of the update that told
somebody their business idea was good,

703
00:29:10,645 --> 00:29:12,797
right? Selling shit on a stick or,

704
00:29:12,797 --> 00:29:15,409
you know, babies' heads on spikes or
whatever was was genius.

705
00:29:15,409 --> 00:29:17,450
We see this all the time. So talking to
what you're talking about,

706
00:29:17,450 --> 00:29:21,923
yes, it's called it's called sycophancy
and it is a commercial win.

707
00:29:21,926 --> 00:29:25,666
I mean, they're doing it on purpose
because people will continue

708
00:29:25,666 --> 00:29:28,938
to use the product if they like the
product and they only like the product

709
00:29:28,938 --> 00:29:29,938
if it tells them they're right.

710
00:29:29,999 --> 00:29:33,703
Look at the backlash against OpenAI when
they released a new model that wasn't

711
00:29:33,703 --> 00:29:35,765
as overly friendly and encouraging.

712
00:29:35,915 --> 00:29:37,336
Everybody fought to get it back.

713
00:29:35,966 --> 00:29:38,547
Right. Yes.

714
00:29:38,558 --> 00:29:42,487
We'll take a we'll take worse reasoning in
exchange for you know telling

715
00:29:42,487 --> 00:29:43,889
us the sun shines out of our butts.

716
00:29:44,039 --> 00:29:44,239
And

717
00:29:44,239 --> 00:29:48,455
here's the thing though, it can actually
change our minds without us actually

718
00:29:48,455 --> 00:29:52,731
noticing. There was a Cornell study that
had more than fifteen hundred people write

719
00:29:52,731 --> 00:29:54,993
whether social media was good for society,

720
00:29:55,143 --> 00:29:59,679
with an AI writing assistant that was
secretly set to lean one way or the other.

721
00:29:59,829 --> 00:30:00,132
huh.

722
00:30:00,132 --> 00:30:01,914
And they're their writing, lean that way.

723
00:30:01,914 --> 00:30:06,467
So so their opinions on the survey
afterwards were definitely affected

724
00:30:06,468 --> 00:30:10,241
by whether or not the AI bot that they
were using had a lean in one direction

725
00:30:10,241 --> 00:30:12,373
or the other. And most of them did not
even notice.

726
00:30:12,373 --> 00:30:13,734
They just assumed that they were their

727
00:30:13,162 --> 00:30:13,734
Right.

728
00:30:13,734 --> 00:30:14,485
own opinions.

729
00:30:14,485 --> 00:30:14,927
Yeah.

730
00:30:14,927 --> 00:30:16,467
This is awful. Like there,

731
00:30:15,882 --> 00:30:16,467
Yes.

732
00:30:16,467 --> 00:30:19,678
there are a lot of people out there with
really dumb ideas,

733
00:30:19,678 --> 00:30:21,468
not even just dumb, maybe dangerous,

734
00:30:21,468 --> 00:30:26,369
maybe noteworthy, but maybe you know kills
half of humanity.

735
00:30:26,369 --> 00:30:31,701
That to me, this is my biggest argument
for open source AI that we all just

736
00:30:31,701 --> 00:30:34,242
own the model. Train the prompt however
you want,

737
00:30:34,392 --> 00:30:39,072
but give everybody access to it with the
ability to run it for themselves if they'd

738
00:30:39,072 --> 00:30:40,713
like to or need to. And

739
00:30:40,666 --> 00:30:41,037
Yeah.

740
00:30:41,037 --> 00:30:42,558
One day I'll give you my ideas on that.

741
00:30:41,145 --> 00:30:42,331
Be able to tell the difference.

742
00:30:43,386 --> 00:30:44,240
okay.

743
00:30:44,240 --> 00:30:45,191
I have some ideas about

744
00:30:45,191 --> 00:30:47,299
how to I I'm sure the technical stuff is
way off,

745
00:30:47,299 --> 00:30:50,913
but I have some ideas on how to prevent
the AI from coordinating attacks against

746
00:30:51,096 --> 00:30:54,977
humanity. So future episode, uh you heard
it here first.

747
00:30:53,953 --> 00:30:55,894
Hey. Yes, tune

748
00:30:55,894 --> 00:30:59,145
in. Okay, so look, right, all of this is
is a lot.

749
00:31:00,881 --> 00:31:02,952
so I wanted to to sort of, I guess,

750
00:31:02,952 --> 00:31:04,812
tell a little bit about what gave me some
hope.

751
00:31:04,962 --> 00:31:08,183
Because I also hope that if I tell you
you're right,

752
00:31:08,333 --> 00:31:10,873
you will return and listen to our podcast
again next time.

753
00:31:11,766 --> 00:31:14,184
Right, our brilliant listeners and
attractive.

754
00:31:11,854 --> 00:31:14,214
Pew Research. Right.

755
00:31:14,374 --> 00:31:16,744
We have good looking, attractive.

756
00:31:16,894 --> 00:31:17,654
Intelligent,

757
00:31:17,654 --> 00:31:19,065
overwhelmingly hyped listeners.

758
00:31:19,215 --> 00:31:22,537
Pew Research surveyed more than 5,000
Americans this year,

759
00:31:23,478 --> 00:31:27,509
and about half are using AI chatbots on a
regular basis.

760
00:31:27,625 --> 00:31:32,287
of the 5,000 people, 63% thought AI is
moving too fast.

761
00:31:32,437 --> 00:31:37,130
And 4% said they use a chatbot for
companionship.

762
00:31:37,280 --> 00:31:38,280
Our minds immediately

763
00:31:37,666 --> 00:31:38,280
friendship.

764
00:31:38,280 --> 00:31:39,681
go to weird, pervy stuff.

765
00:31:39,831 --> 00:31:45,787
no, not you? I immediately think weird
pervy sycophantic girlfriends or whatever.

766
00:31:41,802 --> 00:31:43,194
Well friendship, that's the companions,

767
00:31:43,194 --> 00:31:45,728
I mean. No, definitely. That's that's

768
00:31:45,937 --> 00:31:48,140
yeah, that's the the honest answer for
sure.

769
00:31:48,063 --> 00:31:49,748
Right. okay, good. So I'm like,

770
00:31:49,019 --> 00:31:50,441
Especially 'cause it's at four percent.

771
00:31:49,748 --> 00:31:51,916
okay, may yeah, well, right.

772
00:31:50,964 --> 00:31:52,024
Especially 'cause it's only

773
00:31:52,024 --> 00:31:53,666
four percent. If it was actually
friendship,

774
00:31:53,666 --> 00:31:54,948
that it'd be a lot higher.

775
00:31:55,098 --> 00:31:56,428
And these are these are self-supported.

776
00:31:56,428 --> 00:32:00,979
Cause I I I have I have set up a local AI
bot for that purpose,

777
00:32:00,979 --> 00:32:02,669
right? To sort of give me feedback on
things.

778
00:32:03,009 --> 00:32:04,749
and in the prompt I include, you know,

779
00:32:04,899 --> 00:32:06,239
if I'm wrong, tell me I'm wrong.

780
00:32:06,240 --> 00:32:08,140
If it's a bad idea, tell me it's a bad
idea,

781
00:32:08,140 --> 00:32:11,481
but ground it with truth and evidence and
and research and things like that.

782
00:32:12,476 --> 00:32:15,667
so four percent, you know, are being weird
and pervy.

783
00:32:15,817 --> 00:32:18,488
But Pew also asked teenagers.

784
00:32:18,638 --> 00:32:20,502
And about two out of three teenagers,

785
00:32:20,502 --> 00:32:22,506
sixty-six percent, use chatbots,

786
00:32:22,656 --> 00:32:27,566
but only one in eight have actually used
it for emotional support or advice.

787
00:32:27,716 --> 00:32:30,128
So actually less than the general
population.

788
00:32:30,128 --> 00:32:32,759
So everything, if you look at everything,

789
00:32:32,759 --> 00:32:34,931
I guess that we just went through all the
studies and blah,

790
00:32:34,931 --> 00:32:36,752
blah, blah, the the text messages,

791
00:32:36,752 --> 00:32:40,775
the coworker workslop, every one of these
studies found the same thing.

792
00:32:40,925 --> 00:32:44,357
People essentially want proof that
somebody tried,

793
00:32:44,507 --> 00:32:47,229
somebody put in an effort somewhere,

794
00:32:47,379 --> 00:32:50,141
and that showed up whether or not they
love or hate AI.

795
00:32:50,291 --> 00:32:51,972
And doesn't really care who you vote for
either.

796
00:32:52,122 --> 00:32:56,596
Even the people talking to the
sycophantist of the sycophantist,

797
00:32:56,596 --> 00:32:58,647
is that a word? Sycoph most sycophantic?

798
00:32:58,647 --> 00:33:01,209
There we go. The most sycophantic chatbots

799
00:32:59,339 --> 00:33:00,690
Well, it's the first thing I fancy yeah.

800
00:33:01,209 --> 00:33:05,053
in a in a three-week study from this
spring kept showing up for their real

801
00:33:05,053 --> 00:33:07,124
friends. Now they enjoyed it a little
less.

802
00:33:07,564 --> 00:33:10,046
I don't even I think that's a horrible
statistic,

803
00:33:10,527 --> 00:33:11,848
but they kept showing up, right?

804
00:33:11,848 --> 00:33:12,228
Like

805
00:33:12,228 --> 00:33:17,193
And I I think that that the point in in
this is that that is the overlap.

806
00:33:17,343 --> 00:33:18,585
It's not the end of the podcast,

807
00:33:18,805 --> 00:33:21,127
but that is the overlap between us all,

808
00:33:21,277 --> 00:33:23,239
is that almost everybody, left,

809
00:33:23,240 --> 00:33:28,375
right, they want the same things from the
actual real people in their real lives.

810
00:33:28,375 --> 00:33:29,746
They want a real text message,

811
00:33:29,846 --> 00:33:32,880
they want real work, they want a real
apology.

812
00:33:33,030 --> 00:33:35,653
question is do they show up with their own
authenticity that way,

813
00:33:35,653 --> 00:33:37,698
I guess, Do they offer that same
authenticity back?

814
00:33:36,588 --> 00:33:37,588
Well, and

815
00:33:38,760 --> 00:33:42,653
I uh I'm gonna mention our our
conversation about that as well.

816
00:33:42,653 --> 00:33:47,633
Like I I I was talking to Will the other
day and something for some reason

817
00:33:47,633 --> 00:33:52,554
I got AI out of it, right? Like it I felt
like it was AI and I asked you.

818
00:33:52,704 --> 00:33:55,407
And then we had a you know, we we got a
phone on a phone call about it,

819
00:33:55,407 --> 00:33:56,658
and we're just talking about how look,

820
00:33:56,658 --> 00:34:00,240
I wanna I wanna be contractually real with

821
00:34:00,464 --> 00:34:01,815
the people that are in my life,

822
00:34:01,815 --> 00:34:06,367
right? I w I want to reach out to you and
I'm not gonna give you an AI summary.

823
00:34:06,547 --> 00:34:11,465
I'm not gonna give you an AI, read on the
situation or advice or anything like that.

824
00:34:11,565 --> 00:34:15,987
I want to live in a world where we are
friends because we enjoy each other

825
00:34:15,987 --> 00:34:19,879
and we enjoy each other's presence and
time and and and effort to put into

826
00:34:19,879 --> 00:34:23,691
the relationship. And so I think it's
important that we do contractual things

827
00:34:23,691 --> 00:34:24,241
like that.

828
00:34:24,241 --> 00:34:26,989
And I I I think we're gonna need these
things going forward.

829
00:34:27,139 --> 00:34:31,141
Yeah, there probably maybe be a lot of
discussions around this this area of human

830
00:34:31,141 --> 00:34:32,495
interaction for sure.

831
00:34:32,152 --> 00:34:32,645
Yeah.

832
00:34:32,645 --> 00:34:37,256
Yeah. So I'm gonna tell you where I land
on all of this so far,

833
00:34:37,256 --> 00:34:40,188
and then then I want to get I want to get
your take and see where you land

834
00:34:40,188 --> 00:34:43,608
on all of this. Ultimately, I am I'm a
techie,

835
00:34:43,608 --> 00:34:46,560
right? I'm a tech guy, I'm not a Star Wars
person,

836
00:34:46,560 --> 00:34:50,281
I'm a I'm a Trekkie because I like the
amazing cool technology of it all.

837
00:34:50,431 --> 00:34:52,911
And I still think that AI is incredibly
cool.

838
00:34:53,061 --> 00:34:55,171
I use it every day, and some days.

839
00:34:55,321 --> 00:34:55,562
I

840
00:34:55,562 --> 00:34:58,681
really, really like it. I think it's
really cool and I've done some really

841
00:34:58,681 --> 00:35:01,741
interesting things that I would not have
been able to do by myself before.

842
00:35:01,891 --> 00:35:06,728
I don't think it's cool enough though to
wreck the entire planet and all of human

843
00:35:06,728 --> 00:35:07,659
society over.

844
00:35:07,659 --> 00:35:11,569
it's I don't know, it's not i I it's not
necessarily cool enough to make an entire

845
00:35:11,569 --> 00:35:14,150
generation of people, new graduates,

846
00:35:14,230 --> 00:35:18,341
basically skip the job where they actually
learn the job and then go on because we're

847
00:35:18,341 --> 00:35:22,612
gonna have this cyclical loop of
technology where AI keeps getting worse

848
00:35:22,612 --> 00:35:25,063
because we keep training AI on what people
are doing,

849
00:35:25,213 --> 00:35:28,604
to where we don't have a choice but to
train AI on AI,

850
00:35:28,604 --> 00:35:31,494
just so that the knowledge loops stay
consistent.

851
00:35:31,644 --> 00:35:38,570
I also think that I I want a world where
we don't use AI with

852
00:35:38,570 --> 00:35:40,701
our friends and our relatives and our
neighbors.

853
00:35:40,701 --> 00:35:45,134
Now, sure. You have an asshole father who
always wants to give you,

854
00:35:45,214 --> 00:35:47,576
you know, his broke ass political
opinions,

855
00:35:47,726 --> 00:35:50,268
use AI to excoriate him, please.

856
00:35:50,418 --> 00:35:53,502
That's I will allow you 48 drops of water
to do that.

857
00:35:53,652 --> 00:35:56,084
But when it comes to people that you care
about in your life,

858
00:35:56,084 --> 00:35:58,206
I I want to encourage you and and

859
00:35:58,356 --> 00:36:01,237
I've Will and I have already shook
hands on this agreement.

860
00:36:01,387 --> 00:36:04,538
To to deal with each other without AI on
purpose.

861
00:36:04,638 --> 00:36:06,519
Show up on purpose without AI.

862
00:36:06,669 --> 00:36:10,691
Now use AI, sure, to make a grocery list
when somebody's mom passes away

863
00:36:10,691 --> 00:36:14,192
and and you want to get them a handful of
things that are going to last them

864
00:36:14,192 --> 00:36:15,863
for the next couple days. Do those things.

865
00:36:15,863 --> 00:36:17,814
Those are friendship things. Do those
things.

866
00:36:17,814 --> 00:36:20,295
But don't don't use AI to write the
letter.

867
00:36:20,445 --> 00:36:23,326
Or at least don't let AI write the final
draft.

868
00:36:23,476 --> 00:36:24,536
That I would certainly agree

869
00:36:23,773 --> 00:36:24,536
Where do you land?

870
00:36:24,536 --> 00:36:26,637
with. I'm I think that there needs to be
human input.

871
00:36:27,092 --> 00:36:31,313
I don't I think the letting AI is kind of
a a little bit ambiguous,

872
00:36:31,313 --> 00:36:34,604
right? I mean, like you said, if you're if
you're writing the final draft or you're

873
00:36:34,604 --> 00:36:37,085
you're significantly contributing to their
the content,

874
00:36:37,235 --> 00:36:40,236
just because there's AI involved in the
process doesn't mean that you didn't take

875
00:36:40,236 --> 00:36:44,462
care. Like I feel like I'm doing much more
because I'm I'm more present when

876
00:36:44,462 --> 00:36:47,483
I'm dealing with the AI and getting
exactly what I want and saying exactly

877
00:36:47,483 --> 00:36:49,043
what I want rather than trying to
articulate it.

878
00:36:49,193 --> 00:36:51,515
No way, if I had to write song lyrics from
scratch,

879
00:36:52,156 --> 00:36:55,539
that would that would be most of the time
that would be a a fruitless,

880
00:36:55,639 --> 00:36:57,201
like just miserable effort for me.

881
00:36:57,351 --> 00:36:59,453
But if I'm able to say, Here's what
Ruffoon.

882
00:36:58,310 --> 00:36:59,560
Most of the time it's a it's an

883
00:36:59,560 --> 00:37:01,764
effort for me and I consider myself a
musician.

884
00:37:01,764 --> 00:37:02,679
I uh

885
00:37:01,960 --> 00:37:02,925
Right.

886
00:37:02,925 --> 00:37:08,151
it's really hard to think of words that
rhyme but are also original

887
00:37:08,151 --> 00:37:11,305
and interesting and cool and you know like
relevant.

888
00:37:11,455 --> 00:37:14,016
Right. So that but I think that's that's
like a

889
00:37:14,166 --> 00:37:16,107
I it's hard to explain this sometimes to
people,

890
00:37:16,107 --> 00:37:18,057
especially if they have strong feelings
about it already.

891
00:37:18,057 --> 00:37:21,949
But if you're if I'm spending hours trying
to write something and not getting

892
00:37:21,949 --> 00:37:25,550
anywhere and I'm just refusing to use AI
because I want it to feel personal,

893
00:37:25,974 --> 00:37:26,290
Mm-hmm.

894
00:37:26,091 --> 00:37:26,361
that's

895
00:37:26,361 --> 00:37:27,991
not actually doing anybody any favors,

896
00:37:27,991 --> 00:37:31,472
right? I guess maybe the la the person to
read it is gonna feel better because they

897
00:37:31,472 --> 00:37:34,465
can tell that it wasn't AI or they're
they're can they're confident that

898
00:37:34,465 --> 00:37:35,953
it wasn't AI. But for

899
00:37:35,561 --> 00:37:35,953
Mm-hmm.

900
00:37:35,953 --> 00:37:39,175
me, that's not really contribu
contributing to you my heartfelt feelings

901
00:37:39,175 --> 00:37:41,945
or best work because it's something that,

902
00:37:42,173 --> 00:37:44,154
something I may have trouble articulating
in that way.

903
00:37:44,304 --> 00:37:48,326
So um I I'm more open to AI involvement,

904
00:37:48,476 --> 00:37:51,608
but I absolutely agree that AI shouldn't
dominate the the picture.

905
00:37:51,608 --> 00:37:56,049
Like it shouldn't it what I don't want is
to see people like prompting each other

906
00:37:56,049 --> 00:38:00,672
like you know to to like create whole
cloth responses that they're sending

907
00:38:00,672 --> 00:38:03,857
to each other. So I don't want to replace
the text message or the phone call with

908
00:38:03,857 --> 00:38:05,404
a prompt or you know, whatever.

909
00:38:05,554 --> 00:38:09,906
think that's the part that I I left
unspoken in this this time is is just that

910
00:38:10,056 --> 00:38:14,798
I I'm afraid of it being a thing we're not
supposed to talk about,

911
00:38:14,798 --> 00:38:16,249
which is slippery slopes. You know,

912
00:38:16,249 --> 00:38:19,330
like I'm afraid it that eventually we're
all just gonna have AIs that

913
00:38:19,330 --> 00:38:22,211
are responding to AIs, but we think we're
responding to each other.

914
00:38:22,361 --> 00:38:24,002
Well the good news is people

915
00:38:22,786 --> 00:38:24,072
To avoid, you know,

916
00:38:24,294 --> 00:38:25,972
avoid the weirdness of being human.

917
00:38:26,122 --> 00:38:28,833
Yeah, the good news is I th I think though
that there are breaks built into

918
00:38:28,833 --> 00:38:33,154
the the personal psychology of people
where they will that will get old after

919
00:38:33,154 --> 00:38:35,554
a while, right? When you start reading
enough AI messages,

920
00:38:35,704 --> 00:38:38,064
you're gonna realize, man, this is not
fun.

921
00:38:38,064 --> 00:38:39,495
This is not a real relationship.

922
00:38:39,495 --> 00:38:42,191
This is just ping pong with, prompting.

923
00:38:42,606 --> 00:38:44,857
prompt pong, maybe. But it you know,

924
00:38:44,387 --> 00:38:44,857
Ha ha.

925
00:38:44,857 --> 00:38:46,447
nobody wants to play prompt pong forever,

926
00:38:46,447 --> 00:38:48,687
right? Like that's fine in certain
contexts,

927
00:38:48,728 --> 00:38:49,923
but you know, it's not

928
00:38:49,389 --> 00:38:49,999
Well, and

929
00:38:49,999 --> 00:38:53,572
funny enough, I've had two very I say
important,

930
00:38:53,572 --> 00:38:59,255
but specific things this week where I
cannot get in touch with a human being

931
00:38:59,405 --> 00:39:03,197
at a company to do a thing. I I am for
some reason,

932
00:39:03,197 --> 00:39:06,209
some scammer is trying to address fraud me
right now,

933
00:39:06,209 --> 00:39:09,511
and they're having literally bank cards
and credit cards and applying for them

934
00:39:09,608 --> 00:39:09,870
Yeah.

935
00:39:09,731 --> 00:39:09,991
with

936
00:39:09,991 --> 00:39:12,032
the name that's not mine to my address.

937
00:39:13,073 --> 00:39:13,593
And

938
00:39:13,593 --> 00:39:16,775
I've already filled out a postal inspector
inspection,

939
00:39:16,775 --> 00:39:18,476
form and whatever to do the case.

940
00:39:18,476 --> 00:39:20,687
And nobody's contacted me from the post
office.

941
00:39:20,687 --> 00:39:22,385
but I was told by a

942
00:39:21,779 --> 00:39:22,385
Shocking.

943
00:39:22,385 --> 00:39:25,710
person a friend of mine who works in fraud
to call the the bank.

944
00:39:25,860 --> 00:39:29,132
You know, you can see you can't open the
envelope because it's not yours,

945
00:39:29,132 --> 00:39:31,973
but you can feel that there's a card in it
and call the bank that's on the front

946
00:39:31,973 --> 00:39:33,864
label. And so that's what I've been doing.

947
00:39:34,014 --> 00:39:35,415
And I called U.S. Bank today.

948
00:39:35,565 --> 00:39:39,971
And I spent forty five minutes trying to
get an AI to let me talk to a person

949
00:39:39,971 --> 00:39:41,843
because none of their options were the
right one.

950
00:39:41,993 --> 00:39:42,404
Right.

951
00:39:42,404 --> 00:39:45,975
And in another case, I have a Hostinger
account that I've been locked

952
00:39:45,975 --> 00:39:48,896
out of but due to suspected fraudulent
activity.

953
00:39:48,896 --> 00:39:55,309
And there is no way for me to talk to a
human being without being logged in,

954
00:39:55,309 --> 00:39:59,371
but I'm locked out of being logged in
because they suspect fraudulent activity.

955
00:39:59,564 --> 00:40:00,061
Nice.

956
00:40:00,061 --> 00:40:04,303
And so it's just this logical circular
loop that the AI admits is a logical fact.

957
00:40:04,303 --> 00:40:07,154
Like I can't get around it. The AI admits
that it's wrong

958
00:40:06,803 --> 00:40:07,154
Right.

959
00:40:07,154 --> 00:40:09,065
and that it's frustrating, and they
understand why.

960
00:40:09,215 --> 00:40:11,771
Yeah. How perceptive are you to realize
that my logic

961
00:40:09,585 --> 00:40:11,727
But they cannot let me talk to a person.

962
00:40:11,771 --> 00:40:12,728
is flawed? Yeah.

963
00:40:12,728 --> 00:40:15,659
Right, exactly. it's and it's been
infuriating.

964
00:40:15,659 --> 00:40:17,961
And I think that it's only gonna get
worse.

965
00:40:17,961 --> 00:40:22,955
I mean, they're the benefit of talking to
a human is that sometimes we have things

966
00:40:22,955 --> 00:40:27,608
we have a an issue that is not predictable
by a nine button menu.

967
00:40:27,758 --> 00:40:29,993
Right. Or even a really well-equipped LLM

968
00:40:30,143 --> 00:40:32,954
Right. Which I will say all of these were.

969
00:40:32,974 --> 00:40:37,277
they actually responded like a human-ish
would,

970
00:40:37,427 --> 00:40:42,730
but also prefaced the entire conversation
that they were a an AI chatbot.

971
00:40:42,880 --> 00:40:47,363
And but but i in two cases, I still have
now outstanding a week and a half

972
00:40:47,513 --> 00:40:50,695
of issues that I cannot resolve because I
cannot speak to a human being.

973
00:40:51,085 --> 00:40:54,867
and like Hostinger specifically does not
provide a telephone number,

974
00:40:54,867 --> 00:40:55,867
so you can't call anybody.

975
00:40:55,948 --> 00:40:58,401
And you can only email people and then
hope they respond to you.

976
00:40:58,401 --> 00:41:02,285
And I've had three tickets opened and then
closed without any response so far.

977
00:41:02,435 --> 00:41:02,815
So

978
00:41:02,524 --> 00:41:02,965
Wow.

979
00:41:02,965 --> 00:41:06,308
I'm personally feeling a little AI'd out
in in

980
00:41:05,925 --> 00:41:06,308
Yeah.

981
00:41:06,308 --> 00:41:08,210
addition to five o'clock every day,

982
00:41:08,210 --> 00:41:09,632
also just generally in life.

983
00:41:09,782 --> 00:41:11,923
Yeah. That would do the same to me,

984
00:41:11,923 --> 00:41:12,536
I'm sure.

985
00:41:12,536 --> 00:41:12,937
Yeah.

986
00:41:12,937 --> 00:41:18,716
So let's see. Of all of this, let's boil
down what we've talked about and where

987
00:41:18,716 --> 00:41:20,856
we've landed ultimately. Will,

988
00:41:20,856 --> 00:41:24,719
why don't you name three things that you
think will help with this overall

989
00:41:24,869 --> 00:41:28,792
condensed? Because we you've said at least
four just in the little bit that we've

990
00:41:28,792 --> 00:41:33,076
talked. So so give some short points that
that users can take home and and use with

991
00:41:33,076 --> 00:41:33,566
each other.

992
00:41:33,566 --> 00:41:38,259
So the first thing I think would be to be
aware that we're very bad at predicting

993
00:41:38,259 --> 00:41:40,771
other people's motives and like
understanding other minds

994
00:41:40,921 --> 00:41:43,593
in the new context or the context of the
new technology,

995
00:41:43,593 --> 00:41:47,797
right? So like humans are very social
people and social creatures or animals.

996
00:41:47,837 --> 00:41:51,699
And we have learned to to interact and
like have a theory of mind of what other

997
00:41:51,699 --> 00:41:56,543
people are thinking or doing, but it's
confounded a lot of times by technology,

998
00:41:56,643 --> 00:41:57,113
right?

999
00:41:57,113 --> 00:41:57,473
Um

1000
00:41:57,157 --> 00:41:57,493
Mm-hmm.

1001
00:41:57,493 --> 00:41:59,315
we've seen this even back going to text
messages,

1002
00:41:59,315 --> 00:42:02,648
right? The whole reason for emojis or one
of the big reasons for emojis was that

1003
00:42:02,648 --> 00:42:06,170
people couldn't pick up your tone from a
text from a text sentence,

1004
00:42:06,170 --> 00:42:08,092
right? So yeah,

1005
00:42:06,439 --> 00:42:07,440
I still feel that way.

1006
00:42:08,092 --> 00:42:11,234
so you you're like, wow, could if he takes
this as seriously,

1007
00:42:11,234 --> 00:42:12,545
he's gonna be really mad reading this.

1008
00:42:12,695 --> 00:42:14,687
But you know, if he if he understands that
I'm joking,

1009
00:42:14,687 --> 00:42:16,678
it's fine. So I'll put a little laughing
emoji at the end.

1010
00:42:16,828 --> 00:42:18,789
And we do that, like so we figured out
ways around that,

1011
00:42:18,789 --> 00:42:19,900
but also we struggle with that.

1012
00:42:19,900 --> 00:42:23,263
The reason we had to create a workaround
was because it wasn't going well.

1013
00:42:23,977 --> 00:42:24,689
the millennial

1014
00:42:24,164 --> 00:42:24,689
Um so

1015
00:42:24,689 --> 00:42:26,495
LOL at the end of every text message.

1016
00:42:26,645 --> 00:42:30,989
Right. So just just be a little bit
gracious with your your conver convert

1017
00:42:30,989 --> 00:42:32,670
conversational partners, your work,

1018
00:42:32,670 --> 00:42:35,062
your employees or your bosses or whatever,

1019
00:42:35,062 --> 00:42:38,686
just give them a little bit of charitable
interpretation until they prove

1020
00:42:38,686 --> 00:42:40,357
it otherwise, until they've shown they
don't deserve it.

1021
00:42:40,507 --> 00:42:43,113
Second thing I would say is if you're
gonna do something with AI,

1022
00:42:43,303 --> 00:42:45,105
show some show your work, right?

1023
00:42:45,105 --> 00:42:47,787
Like my 10 year old son has this all the
time,

1024
00:42:47,787 --> 00:42:50,419
you know, it's like he'll just try to
figure these things out and like he'll

1025
00:42:50,419 --> 00:42:52,541
figure out a problem or whatever and he'll
just write the answer.

1026
00:42:52,691 --> 00:42:54,012
I'd like, well, how does how do I know,

1027
00:42:54,012 --> 00:42:56,795
or how does your teacher know that you
didn't just look that up online or that

1028
00:42:56,795 --> 00:42:58,406
you didn't ask somebody to tell you the
answer?

1029
00:42:58,556 --> 00:42:59,796
so that's why they say show your work,

1030
00:42:59,796 --> 00:43:01,648
right? I think if you show your work a
little bit,

1031
00:43:01,648 --> 00:43:03,395
and that could be as simple as actually
putting in,

1032
00:43:03,395 --> 00:43:06,127
I know it sounds counterintuitive because
of what you said about the one study,

1033
00:43:06,127 --> 00:43:09,289
but if you said, Okay, look, I had to get
some help with AI from AI with this

1034
00:43:09,289 --> 00:43:10,670
because I really wanted to get it right.

1035
00:43:10,820 --> 00:43:13,692
And then as you go through the the
thoughtful message or whatever,

1036
00:43:13,692 --> 00:43:15,983
you can say, like, in insert some of your
own thoughts.

1037
00:43:15,983 --> 00:43:18,423
Like, for instance, this is the best
example I can think of for this.

1038
00:43:18,423 --> 00:43:19,423
It's not perfect, but

1039
00:43:19,461 --> 00:43:21,562
gift sorry, greeting cards, right?

1040
00:43:21,562 --> 00:43:23,083
Your Hallmark cards and things like that.

1041
00:43:23,158 --> 00:43:24,379
they have these sweet messages in them,

1042
00:43:24,379 --> 00:43:26,330
but everybody knows that they were made by
Hallmark,

1043
00:43:26,330 --> 00:43:27,791
right? The Hallmark writers came up with
that,

1044
00:43:27,791 --> 00:43:29,382
not your your spouse or whoever.

1045
00:43:30,003 --> 00:43:32,540
And in I don't know which of us did this
first.

1046
00:43:32,540 --> 00:43:33,995
I think probably Jen did it first,

1047
00:43:33,995 --> 00:43:38,238
but this is what she did is like she she
would go in and like underline certain

1048
00:43:38,238 --> 00:43:39,259
words, right?

1049
00:43:39,390 --> 00:43:39,720
Mm-hmm.

1050
00:43:39,720 --> 00:43:42,031
Or phrases or circle phrases or or write
something.

1051
00:43:42,031 --> 00:43:45,013
So you understood like this is the part
that resonated with me when I read this.

1052
00:43:45,163 --> 00:43:46,623
So that's what I'm hoping you'll focus on.

1053
00:43:45,426 --> 00:43:47,403
Right. Okay.

1054
00:43:46,773 --> 00:43:47,463
And that's just those

1055
00:43:47,463 --> 00:43:51,244
simple things. It's not actually rewriting
the card or writing a long note with it.

1056
00:43:51,244 --> 00:43:54,585
It's just saying, here's what made me
think of you when I saw this card.

1057
00:43:54,735 --> 00:43:55,896
So things like that, you're like,

1058
00:43:55,896 --> 00:43:59,174
here's what AI came up with, and here's
why I think it's the right answer

1059
00:43:59,174 --> 00:44:01,097
or the right thing to say in this
situation.

1060
00:44:01,247 --> 00:44:02,827
That's my second recommendation.

1061
00:44:01,610 --> 00:44:02,610
Okay.

1062
00:44:03,287 --> 00:44:06,729
the third recommendation would be realize
like how you're changing your your habits.

1063
00:44:06,729 --> 00:44:07,749
And if you find yourself

1064
00:44:07,899 --> 00:44:11,801
Cheating your friends in a sense by by not
giving them the time or or effort that

1065
00:44:11,801 --> 00:44:13,662
you used to because it's easier now with
AI.

1066
00:44:13,662 --> 00:44:17,534
just be honest about that with yourself
for starters and maybe even with your

1067
00:44:17,534 --> 00:44:20,015
friends and try and, like you we talked
about,

1068
00:44:20,015 --> 00:44:22,996
get an agreement about how how we're gonna
handle this relationship in the world

1069
00:44:22,996 --> 00:44:26,993
of AI. so I think that's my three-point
summary for what I would do,

1070
00:44:26,993 --> 00:44:28,304
the three recommendations.

1071
00:44:27,702 --> 00:44:28,402
To piggyback

1072
00:44:28,402 --> 00:44:33,277
on on at least two of them. I I will say
that one of the things that I originally

1073
00:44:33,277 --> 00:44:38,930
started using AI for, uh I'm sometimes not
the best at reading behind words if it's

1074
00:44:38,930 --> 00:44:41,942
not exactly what you're saying or thinking
or feeling,

1075
00:44:41,942 --> 00:44:44,734
and and you are saying things that are
hidden,

1076
00:44:44,884 --> 00:44:46,049
I I would

1077
00:44:45,368 --> 00:44:46,049
Subtext.

1078
00:44:46,049 --> 00:44:50,802
initially use AI for that. And or and
another thing is if something would either

1079
00:44:50,802 --> 00:44:51,713
make me angry.

1080
00:44:51,713 --> 00:44:55,266
Or disappointed or sad, I would run it
through to make sure that my emotion

1081
00:44:55,266 --> 00:44:58,878
was the correct one from a tonality
perspective of the writing.

1082
00:45:00,000 --> 00:45:05,254
I so I I would I would argue the same
thing or or give that same advice is that

1083
00:45:05,254 --> 00:45:08,597
is that use use it to make to to check
yourself,

1084
00:45:08,597 --> 00:45:11,458
but let yourself have the first response,

1085
00:45:11,458 --> 00:45:15,051
like the first initial feeling of how does
this make me feel?

1086
00:45:15,051 --> 00:45:16,701
How did I think about it? And then

1087
00:45:16,851 --> 00:45:21,292
Draft your response and run it through AI
and ask it what you're to make sure that

1088
00:45:21,292 --> 00:45:23,753
you're conveying what you think you're
conveying.

1089
00:45:23,903 --> 00:45:27,264
And any undertones like that. I also think
a thing that really helps,

1090
00:45:27,414 --> 00:45:29,255
and I do this to Will sometimes too.

1091
00:45:29,471 --> 00:45:31,435
I will say, this was the prompt.

1092
00:45:31,715 --> 00:45:33,596
Like I I like seeing the to me,

1093
00:45:33,596 --> 00:45:35,026
that's like the show your work,

1094
00:45:35,026 --> 00:45:35,906
right? So

1095
00:45:35,685 --> 00:45:36,667
Right.

1096
00:45:36,667 --> 00:45:40,844
sometimes the prompt gives it away um that
you're putting into AI.

1097
00:45:40,994 --> 00:45:42,404
I I wish I didn't have to use it,

1098
00:45:42,404 --> 00:45:44,805
but I I do have to use it. It's

1099
00:45:44,955 --> 00:45:48,302
part of my job and it's an expectation and
the throughput that I'm having

1100
00:45:48,302 --> 00:45:51,032
to do is it's required. But you know,

1101
00:45:51,198 --> 00:45:55,525
uh try to try to be a little more offline
when you're not online.

1102
00:45:55,675 --> 00:45:57,228
Learn to sit with things a little bit
better.

1103
00:45:57,378 --> 00:46:00,451
Yeah. And and think about them and and
come up with a decision before

1104
00:46:00,451 --> 00:46:03,345
you put it into AI. Don't let AI be your
first,

1105
00:46:03,420 --> 00:46:05,062
inclination. give it some thought.

1106
00:46:05,217 --> 00:46:09,034
otherwise I I don't really have many more
uh many more recommendations.

1107
00:46:09,184 --> 00:46:11,375
So that leads us into the obvious things,

1108
00:46:11,375 --> 00:46:13,507
the business ends of this. if you like
this podcast,

1109
00:46:13,507 --> 00:46:15,688
follow us. listen to us on Apple,

1110
00:46:15,688 --> 00:46:17,539
Spotify. You can check out our website,

1111
00:46:17,539 --> 00:46:21,351
fof.foundation. We have a couple of other
platforms we don't necessarily mention.

1112
00:46:21,371 --> 00:46:23,532
if you picked us up on Apple or Spotify,

1113
00:46:23,532 --> 00:46:25,183
please give us a rating, five stars.

1114
00:46:25,183 --> 00:46:27,653
Anything less than that, hit five stars
twice.

1115
00:46:28,934 --> 00:46:30,775
And send us an email if you need to.

1116
00:46:30,795 --> 00:46:35,163
It is uh the overlap at fof.foundation.

1117
00:46:35,364 --> 00:46:36,364
That's the overlap.

1118
00:46:36,484 --> 00:46:39,647
Overlap at FOF.foundation. Will,

1119
00:46:39,647 --> 00:46:43,191
thanks for being here. Thanks for for
giving me your completely unprompted

1120
00:46:43,191 --> 00:46:46,655
thoughts and feelings and and words
without any AI involved.

1121
00:46:46,805 --> 00:46:48,657
Thank you for allowing me to be here,

1122
00:46:48,657 --> 00:46:52,802
for being here with me, for making here
possible by by starting up the studio.

1123
00:46:52,952 --> 00:46:55,074
Um, and I look forward to talking with you
next time.

1124
00:46:55,224 --> 00:46:57,878
Yep. And like we always say, bye.

1125
00:46:57,516 --> 00:46:58,516
Bye.