1
00:00:02,440 --> 00:00:03,520
Speaker 1: As Media.

2
00:00:06,519 --> 00:00:08,959
Speaker 2: Hello and welcome to Better Offline. I'm your host ed

3
00:00:09,039 --> 00:00:23,919
Speaker 2: ze Tron. We're here in the beautiful iHeart Radio Studios

4
00:00:23,960 --> 00:00:25,880
Speaker 2: in New York City, and I've got a guest, of course,

5
00:00:25,960 --> 00:00:28,760
Speaker 2: Charlie Meyer, the esteem blogger and CEO of Picko. Charlie,

6
00:00:28,800 --> 00:00:32,480
Speaker 2: thank you for joining me. Yeah, thanks for having me. So, yeah,

7
00:00:32,560 --> 00:00:36,080
Speaker 2: you've you've gained some I would say, notoriety recently by

8
00:00:36,120 --> 00:00:38,960
Speaker 2: making blogs that go against the oinking of the hogs

9
00:00:39,000 --> 00:00:42,800
Speaker 2: of the Valley. And I think your scaling laws piece

10
00:00:42,920 --> 00:00:45,040
Speaker 2: was the one that really got going. Yeah.

11
00:00:45,080 --> 00:00:50,840
Speaker 3: So I my blog gets some love and mostly hate

12
00:00:50,880 --> 00:00:53,680
Speaker 3: on Hacker News. That's my distribution channel, right, and so

13
00:00:53,720 --> 00:00:55,520
Speaker 3: I'm trying to get off of that. We're gonna try

14
00:00:55,520 --> 00:00:57,840
Speaker 3: and build it like a newsletter type thing. But yeah,

15
00:00:57,840 --> 00:00:59,560
Speaker 3: I'll burst on Hacker News and every once in a

16
00:00:59,560 --> 00:01:01,880
Speaker 3: while I'll go something that blows up and I'll get

17
00:01:01,920 --> 00:01:03,200
Speaker 3: my get my haters in there.

18
00:01:03,280 --> 00:01:05,119
Speaker 2: So, so what is it that's pissing them off?

19
00:01:05,920 --> 00:01:08,720
Speaker 3: So, like I had to post a few weeks back

20
00:01:09,000 --> 00:01:11,760
Speaker 3: that was on I called it LLLMS or the Ultimate

21
00:01:11,800 --> 00:01:14,160
Speaker 3: demoware right, And so I had to find demoware as

22
00:01:14,240 --> 00:01:16,800
Speaker 3: software that you make the software, and it works well

23
00:01:16,840 --> 00:01:19,040
Speaker 3: in the thirty minutes that you're showing off to executives

24
00:01:19,080 --> 00:01:20,880
Speaker 3: or whoever's going to buy it, right, and then it

25
00:01:20,920 --> 00:01:23,560
Speaker 3: doesn't do the thing right, It doesn't do the thing

26
00:01:23,640 --> 00:01:25,920
Speaker 3: day to day. And so I listed some examples and

27
00:01:26,040 --> 00:01:28,360
Speaker 3: my startups in etech, right, so we do, you know?

28
00:01:28,440 --> 00:01:30,160
Speaker 3: And so like there's always something that I pick on.

29
00:01:30,440 --> 00:01:32,560
Speaker 3: Is I really hate AI tutors and we can get

30
00:01:32,560 --> 00:01:35,120
Speaker 3: into to that and how that all of courks. But

31
00:01:35,920 --> 00:01:38,200
Speaker 3: so I said, oh, I listed out a few things

32
00:01:38,200 --> 00:01:39,800
Speaker 3: that I thought were demoware. So it's like, oh, vibe

33
00:01:39,840 --> 00:01:41,680
Speaker 3: coading that makes dashboards, right now, that's an easy thing

34
00:01:41,720 --> 00:01:44,920
Speaker 3: to pick on. And I said, you know AI tutors,

35
00:01:45,040 --> 00:01:47,480
Speaker 3: And I said, well, maybe the kid won't want to

36
00:01:47,520 --> 00:01:49,880
Speaker 3: talk to an AI tutor. Yeah, that was the critique

37
00:01:49,920 --> 00:01:52,000
Speaker 3: I made, right, It is like maybe they just don't

38
00:01:52,000 --> 00:01:53,240
Speaker 3: want to talk to him, Yeah, won't want to talk

39
00:01:53,240 --> 00:01:55,200
Speaker 3: to a post they want to maybe they want to

40
00:01:55,200 --> 00:01:57,120
Speaker 3: have like a teacher who is like in the classroom.

41
00:01:57,240 --> 00:02:00,320
Speaker 2: Crazy idea maybe, yeah. But did people know like that?

42
00:02:00,640 --> 00:02:03,640
Speaker 3: Well, yeah, so some some people, uh, you know, we

43
00:02:03,640 --> 00:02:05,240
Speaker 3: can name names of for me too, But I actually

44
00:02:05,240 --> 00:02:08,200
Speaker 3: don't know how to pronounce them. So uh but anyway,

45
00:02:08,240 --> 00:02:10,160
Speaker 3: so people are in there and they're like, you have

46
00:02:10,200 --> 00:02:14,200
Speaker 3: no idea, Like if you think that AI can't tutor calculus,

47
00:02:14,240 --> 00:02:17,200
Speaker 3: like you have never even tried, It's like it's a classic,

48
00:02:17,280 --> 00:02:22,400
Speaker 3: Like you're missing out, Like you somehow are completely missing

49
00:02:22,400 --> 00:02:23,320
Speaker 3: the point.

50
00:02:23,120 --> 00:02:26,000
Speaker 2: And you know something's really good and the innovative when

51
00:02:26,040 --> 00:02:29,280
Speaker 2: the only defense of it is you're a moron. You

52
00:02:29,280 --> 00:02:31,760
Speaker 2: you well you have never tried it.

53
00:02:31,800 --> 00:02:33,840
Speaker 3: And it's like, well, what if I like have like

54
00:02:33,960 --> 00:02:36,000
Speaker 3: what if this is like actually the thing I spend

55
00:02:36,000 --> 00:02:37,400
Speaker 3: my time on is thinking about this.

56
00:02:37,680 --> 00:02:41,600
Speaker 2: You're a non AI I d E. Right, so coding.

57
00:02:41,400 --> 00:02:43,919
Speaker 3: Environment, yeah, yeah, well and that yeah, we got yeah,

58
00:02:44,000 --> 00:02:45,920
Speaker 3: lots of that, lots of that. But like so the

59
00:02:45,960 --> 00:02:49,000
Speaker 3: software that I make batting around a little bit here,

60
00:02:49,040 --> 00:02:53,040
Speaker 3: but Replet was a copy and so Replet was a

61
00:02:53,280 --> 00:02:57,839
Speaker 3: very loved by Teacher's id online and their whole thing

62
00:02:57,960 --> 00:03:01,040
Speaker 3: was like we teach like we held you teach coding

63
00:03:01,080 --> 00:03:03,720
Speaker 3: online because it's a way for you to run Python

64
00:03:03,760 --> 00:03:05,960
Speaker 3: and Java and all your code online and you can

65
00:03:06,000 --> 00:03:07,280
Speaker 3: do it and you can do it on crumb books

66
00:03:07,320 --> 00:03:09,919
Speaker 3: and you can collaborate and they had like teacher tools

67
00:03:09,960 --> 00:03:12,280
Speaker 3: and they sold the software to schools. I was a

68
00:03:12,280 --> 00:03:13,720
Speaker 3: teacher for a couple of years, so it's kind of

69
00:03:13,720 --> 00:03:15,720
Speaker 3: like my background as an engineer and then a teacher,

70
00:03:15,880 --> 00:03:18,519
Speaker 3: and I used replet and it's awesome. Well this was

71
00:03:18,560 --> 00:03:20,440
Speaker 3: this before they used A or this is before they

72
00:03:20,520 --> 00:03:23,839
Speaker 3: used AI. So so replate just for the listeners right now.

73
00:03:24,040 --> 00:03:26,799
Speaker 3: Replate is an AI powered coding environment that claims to

74
00:03:26,800 --> 00:03:28,640
Speaker 3: be able to vibe code software but doesn't really.

75
00:03:28,720 --> 00:03:29,760
Speaker 2: But what did it used to be?

76
00:03:30,040 --> 00:03:33,120
Speaker 3: So it used to be an excellent tool, just an

77
00:03:33,160 --> 00:03:35,960
Speaker 3: absolutely fantastic tool. It was it was just you go on,

78
00:03:36,120 --> 00:03:38,840
Speaker 3: you log on like Google Docs for coding, right, So,

79
00:03:39,080 --> 00:03:41,080
Speaker 3: like you think, okay, well you back in the day

80
00:03:41,120 --> 00:03:43,840
Speaker 3: you'd have to download Microsoft Word and whatever and that

81
00:03:44,040 --> 00:03:46,560
Speaker 3: socks and you know, it's great to bring that online

82
00:03:46,560 --> 00:03:48,520
Speaker 3: into the cloud. And they did that and they were

83
00:03:48,560 --> 00:03:51,360
Speaker 3: like very innovative. They were kind of the first to

84
00:03:51,440 --> 00:03:55,280
Speaker 3: market of having like a very fully featured online ide

85
00:03:56,000 --> 00:03:58,280
Speaker 3: and that is useful for exactly one thing, and that

86
00:03:58,360 --> 00:04:01,200
Speaker 3: is useful for teaching in schools, right because like you

87
00:04:01,240 --> 00:04:03,520
Speaker 3: have kids and they have two hundred dollars chrumebooks that

88
00:04:03,600 --> 00:04:06,600
Speaker 3: the school bought them, and so you get Replet and

89
00:04:06,680 --> 00:04:08,760
Speaker 3: like boom, I have a great way to teach computer science.

90
00:04:08,760 --> 00:04:11,320
Speaker 3: Now that is fantastic, and that's what it used to

91
00:04:11,360 --> 00:04:15,880
Speaker 3: be USI yeah, and now it's just and now it's yeah.

92
00:04:15,640 --> 00:04:18,440
Speaker 2: It's so listen this you probably heard me mentioned Replet

93
00:04:18,440 --> 00:04:20,200
Speaker 2: in the past. It's one of my least favorite, most

94
00:04:20,200 --> 00:04:22,840
Speaker 2: favorite companies. If you go on the replet reddit.

95
00:04:22,839 --> 00:04:25,640
Speaker 4: It's just the wallet inspect And so now that's kind

96
00:04:25,640 --> 00:04:28,240
Speaker 4: of like I've gotten rid of most of my like

97
00:04:28,279 --> 00:04:31,080
Speaker 4: doom scrolling places, but like this is I don't know

98
00:04:31,120 --> 00:04:33,080
Speaker 4: what type of scrolling it is, but like I go.

99
00:04:33,120 --> 00:04:36,160
Speaker 2: On our replet Yeah yeah, yeah, exactly, and it's just

100
00:04:36,279 --> 00:04:39,839
Speaker 2: so funny. It's just God being like, yeah, spend fifteen

101
00:04:39,920 --> 00:04:44,440
Speaker 2: hundred dollars. It doesn't really work. Yeah, but I think

102
00:04:44,480 --> 00:04:47,560
Speaker 2: if I spend five thousand more dollars it might. Well.

103
00:04:47,920 --> 00:04:50,000
Speaker 3: I mean people are like, okay, well should I spend

104
00:04:50,000 --> 00:04:52,479
Speaker 3: five thousand dollars? I mean we can we can be

105
00:04:52,520 --> 00:04:54,760
Speaker 3: reasonable that you can bring the numbers down to reality.

106
00:04:54,960 --> 00:04:58,280
Speaker 3: It's I spent fifty dollars, which for a person who

107
00:04:58,360 --> 00:05:01,600
Speaker 3: has got a bad suffare, Yeah, that's that's a big

108
00:05:01,640 --> 00:05:03,320
Speaker 3: waste of money. And they're like, okay, well it seems

109
00:05:03,320 --> 00:05:04,880
Speaker 3: like I might need to spend two hundred and fifty

110
00:05:04,920 --> 00:05:06,960
Speaker 3: dollars more or should I go on fiver her. Yeah,

111
00:05:07,000 --> 00:05:09,119
Speaker 3: and then it's just people and Reddit. I mean it's Reddit,

112
00:05:09,200 --> 00:05:10,480
Speaker 3: so they're just like skill issue.

113
00:05:11,080 --> 00:05:12,919
Speaker 2: Well that and also the people who were like it

114
00:05:13,200 --> 00:05:15,040
Speaker 2: am running into this problem.

115
00:05:14,920 --> 00:05:16,479
Speaker 3: And then a lot of that yeah, where it's like

116
00:05:16,760 --> 00:05:21,080
Speaker 3: and but replet just just for for anyone on there

117
00:05:21,200 --> 00:05:23,520
Speaker 3: on replet right now? Yeah, like what did they do

118
00:05:23,560 --> 00:05:26,400
Speaker 3: to teachers? So teachers? They had a product for teachers

119
00:05:26,440 --> 00:05:29,080
Speaker 3: that worked, that was great, that was well teachers. And

120
00:05:29,400 --> 00:05:33,480
Speaker 3: on this November somethink twenty twenty three, this is a

121
00:05:33,520 --> 00:05:35,160
Speaker 3: big day for my business because this is the only

122
00:05:35,200 --> 00:05:38,080
Speaker 3: reason I have my business is to replace replet. Because

123
00:05:38,680 --> 00:05:42,760
Speaker 3: they turned on AI autocomplete for kids, no way to

124
00:05:42,800 --> 00:05:43,320
Speaker 3: shut it off.

125
00:05:43,360 --> 00:05:45,440
Speaker 2: Doesn't that defeat the purpose of learning? Yeah?

126
00:05:45,440 --> 00:05:48,159
Speaker 3: So I had a I have a small YouTube channel

127
00:05:48,200 --> 00:05:50,680
Speaker 3: with not you know, a million subscribers, but you know,

128
00:05:50,920 --> 00:05:52,479
Speaker 3: I talked to teachers on there, and you know, we

129
00:05:52,520 --> 00:05:55,200
Speaker 3: had a customer of mine on there and they were like, yeah,

130
00:05:55,240 --> 00:05:57,839
Speaker 3: you know that year, it just seemed like the guy

131
00:05:58,360 --> 00:06:00,680
Speaker 3: he just like he missed the fact the AI got

132
00:06:00,680 --> 00:06:03,080
Speaker 3: turned on, no one sentiment announcement or an email or

133
00:06:03,080 --> 00:06:03,480
Speaker 3: a warning.

134
00:06:03,560 --> 00:06:05,240
Speaker 2: All of his students were just amazing. They were.

135
00:06:05,480 --> 00:06:07,760
Speaker 3: He was like, yeah, everybody, everybody, everybody got an A

136
00:06:07,839 --> 00:06:10,359
Speaker 3: that semester, Like, I wonder, you know did that actually?

137
00:06:10,600 --> 00:06:13,440
Speaker 2: So did students actually end up getting great schools? Because

138
00:06:13,480 --> 00:06:14,320
Speaker 2: no one knows the AI go?

139
00:06:14,560 --> 00:06:17,080
Speaker 3: Yeah, I mean it took depending on who you If

140
00:06:17,080 --> 00:06:20,240
Speaker 3: you're wandering around the classroom looking at students and you

141
00:06:20,240 --> 00:06:24,200
Speaker 3: see them all tab completing, like because AI just.

142
00:06:23,800 --> 00:06:26,200
Speaker 2: Just for the listeners as well spelled out. So with

143
00:06:26,240 --> 00:06:28,919
Speaker 2: these AI platforms, you hit tab complete because it's basically

144
00:06:28,920 --> 00:06:30,200
Speaker 2: like auto correct decoding.

145
00:06:30,440 --> 00:06:33,800
Speaker 3: Yeah, and so so like AI for what it's worth,

146
00:06:33,880 --> 00:06:38,080
Speaker 3: you know, we can be really balanced podcast. But AI

147
00:06:38,400 --> 00:06:42,720
Speaker 3: can really well, it can solve intro to computer science

148
00:06:43,000 --> 00:06:46,400
Speaker 3: for ninth grader problems with the incredible accuracy.

149
00:06:46,480 --> 00:06:48,520
Speaker 2: Let's call Brown from the Internet bugs. He said. It

150
00:06:48,520 --> 00:06:50,560
Speaker 2: makes the easy things easier and the hard thing's hort.

151
00:06:50,760 --> 00:06:53,040
Speaker 3: Yes, and so yeah, so if you need to, if

152
00:06:53,080 --> 00:06:55,560
Speaker 3: you're in ninth grade and you're writing your first program,

153
00:06:55,760 --> 00:06:57,920
Speaker 3: yeah you can. You can tab complete the whole thing

154
00:06:58,120 --> 00:07:01,200
Speaker 3: in one go. It'll one shot at d that's incredible.

155
00:07:01,320 --> 00:07:03,480
Speaker 3: It'll one shot program.

156
00:07:03,560 --> 00:07:07,320
Speaker 2: It's this term that term just listeners is like it

157
00:07:07,360 --> 00:07:09,400
Speaker 2: means that you just give it a problem and it

158
00:07:09,440 --> 00:07:10,560
Speaker 2: solves it correctly. Yeah.

159
00:07:10,600 --> 00:07:13,400
Speaker 3: So it's like count to ten and the AI can

160
00:07:13,440 --> 00:07:15,280
Speaker 3: count to ten, which is incredible.

161
00:07:15,320 --> 00:07:17,640
Speaker 2: That's revolutionary. But fun fun fact if you try and

162
00:07:17,680 --> 00:07:20,320
Speaker 2: make chat GPT count to a million, it freaks out

163
00:07:20,360 --> 00:07:22,480
Speaker 2: if you do the voice mode. Adam Kinna told me

164
00:07:22,560 --> 00:07:25,120
Speaker 2: this one. If you go like counter a million, it

165
00:07:25,200 --> 00:07:28,080
Speaker 2: stops around nine or ten and then says should I continue,

166
00:07:28,480 --> 00:07:31,000
Speaker 2: and it just won't do it. It's very funny. I

167
00:07:31,040 --> 00:07:33,560
Speaker 2: love living in the future. Yeah.

168
00:07:33,600 --> 00:07:36,040
Speaker 3: So they they though they turn on the AI and

169
00:07:36,080 --> 00:07:39,200
Speaker 3: then they were like, we're not doing education and companies

170
00:07:39,240 --> 00:07:40,239
Speaker 3: have deprecated things.

171
00:07:40,440 --> 00:07:43,920
Speaker 2: Yeah, it happens actually not their main product whatever.

172
00:07:44,000 --> 00:07:46,920
Speaker 3: So that I mean, I'll tell you, you know, I'm

173
00:07:46,920 --> 00:07:49,360
Speaker 3: an indie developer whatever. Like our software does not make

174
00:07:49,360 --> 00:07:51,480
Speaker 3: a ton of money because there isn't that much money

175
00:07:51,480 --> 00:07:54,080
Speaker 3: selling an online ide to schools. There's money, but it's

176
00:07:54,080 --> 00:07:55,400
Speaker 3: not it's not a billion dollar business.

177
00:07:55,440 --> 00:07:58,240
Speaker 2: It's fine. I don't know this replet but yeah, but

178
00:07:59,520 --> 00:08:01,560
Speaker 2: you can't write billions of valuations.

179
00:08:01,680 --> 00:08:04,080
Speaker 3: No, no saying hey, we're you know, kids have chromebooks

180
00:08:04,120 --> 00:08:06,119
Speaker 3: and we're gonna you know, charge ten dollars a student

181
00:08:06,200 --> 00:08:07,920
Speaker 3: or whatever. Like you can't of course that you're not

182
00:08:07,960 --> 00:08:08,640
Speaker 3: going to raise a bill.

183
00:08:08,760 --> 00:08:09,000
Speaker 2: You're not.

184
00:08:09,360 --> 00:08:11,680
Speaker 3: That's not a billion dollar business, no, But the AI

185
00:08:11,840 --> 00:08:14,640
Speaker 3: thing seems magical. And then you know, the vibe coding

186
00:08:14,680 --> 00:08:17,320
Speaker 3: thing happened. And you know, as soon as the vibe

187
00:08:17,320 --> 00:08:19,720
Speaker 3: coding stuff started happening, they were like, we're all in

188
00:08:19,800 --> 00:08:23,640
Speaker 3: on this, and they deleted. They deleted every lead, so

189
00:08:23,760 --> 00:08:26,400
Speaker 3: not just deprecated. Right, So it's one thing to deprecate software.

190
00:08:26,440 --> 00:08:28,600
Speaker 2: So it's like and deprecate is when you stop supporting together.

191
00:08:28,560 --> 00:08:30,560
Speaker 3: You say it's it's no longer supported. When you put

192
00:08:30,600 --> 00:08:32,760
Speaker 3: up a big red scary banner on the top saying

193
00:08:32,880 --> 00:08:35,160
Speaker 3: your work is read only, you cannot create anymore. Right,

194
00:08:35,320 --> 00:08:37,280
Speaker 3: that is a really mean thing to do with. But

195
00:08:37,320 --> 00:08:40,280
Speaker 3: it happens. Software changes, you know, repl it for what

196
00:08:40,320 --> 00:08:42,400
Speaker 3: it's worth to be nice and fair to them, Like

197
00:08:42,600 --> 00:08:45,400
Speaker 3: they have investors and they're under the gun to provide

198
00:08:45,440 --> 00:08:48,720
Speaker 3: some returns and so whatever. The teacher thing isn't gonna

199
00:08:48,720 --> 00:08:51,160
Speaker 3: make them a ton of money. But they deleted the stuff.

200
00:08:51,720 --> 00:08:53,920
Speaker 2: Just why And when you say the stuff was this

201
00:08:54,080 --> 00:08:56,240
Speaker 2: like projects that schools have been working on.

202
00:08:56,400 --> 00:08:58,400
Speaker 3: It was so a teacher says, I'm going to spend

203
00:08:58,559 --> 00:09:01,200
Speaker 3: two three years putting in all my curriculum, all these

204
00:09:01,240 --> 00:09:03,720
Speaker 3: markdown files, all this stuff, all these tests. I'm going

205
00:09:03,800 --> 00:09:06,360
Speaker 3: to configure all this stuff. No deleted, gone.

206
00:09:06,200 --> 00:09:10,200
Speaker 2: Monsters deleted actual monsters and now but they sent the

207
00:09:10,240 --> 00:09:14,480
Speaker 2: warning email ed, Wow in July. When are teachers online

208
00:09:14,480 --> 00:09:17,160
Speaker 2: looking at their work email in July? Yeah? Classic, classic,

209
00:09:17,240 --> 00:09:18,640
Speaker 2: big month for teachers to.

210
00:09:18,640 --> 00:09:21,480
Speaker 3: Be on the well for American teachers, Yeah, it's not

211
00:09:21,559 --> 00:09:23,800
Speaker 3: a huge month. So yeah, in July we say we're

212
00:09:23,840 --> 00:09:25,640
Speaker 3: going to delete all your stuff and then it's gone.

213
00:09:25,960 --> 00:09:27,800
Speaker 2: Was there any way to back it up? Well there

214
00:09:27,920 --> 00:09:28,960
Speaker 2: wasn't until they deleted it.

215
00:09:29,000 --> 00:09:29,160
Speaker 4: All.

216
00:09:29,440 --> 00:09:31,440
Speaker 2: That's so cool. It's awesome.

217
00:09:31,520 --> 00:09:35,439
Speaker 3: So they're an awesome Like if you're a replet developer,

218
00:09:35,640 --> 00:09:38,280
Speaker 3: you know when when the next big thing comes up

219
00:09:38,320 --> 00:09:40,760
Speaker 3: and Replet may decide to delete all your stuff.

220
00:09:40,840 --> 00:09:44,640
Speaker 2: Well replet they launched Asient three. Oh yeah, that was

221
00:09:44,679 --> 00:09:47,679
Speaker 2: my favorite launch of product I've ever seen. Because I've

222
00:09:47,679 --> 00:09:50,079
Speaker 2: mentioned this on the show before, but they it's like, oh,

223
00:09:50,080 --> 00:09:53,079
Speaker 2: it's an autonomous coding thing and it's just the digital

224
00:09:53,120 --> 00:09:55,439
Speaker 2: mister bean, it's just like, why don't you go off

225
00:09:55,480 --> 00:09:57,920
Speaker 2: and build me a software thing? And it just fucking

226
00:09:57,960 --> 00:10:00,280
Speaker 2: spends one hundred dollars and goes, I am, I don't

227
00:10:00,280 --> 00:10:02,600
Speaker 2: know you like this. I don't fucking care. And then

228
00:10:02,640 --> 00:10:04,160
Speaker 2: they had to release the thing where you could make

229
00:10:04,200 --> 00:10:06,440
Speaker 2: it think less. They had to like add tweaks to

230
00:10:06,520 --> 00:10:09,840
Speaker 2: it because it was so bad. It's I actually feel like,

231
00:10:09,880 --> 00:10:11,400
Speaker 2: and I'm not putting words in your mouth there, I

232
00:10:11,440 --> 00:10:14,480
Speaker 2: feel like vibe coding maybe just fraud. I think it's

233
00:10:14,480 --> 00:10:17,000
Speaker 2: a fraud. I don't it should not be legal to

234
00:10:17,120 --> 00:10:18,880
Speaker 2: lielight because it is a fucking lie.

235
00:10:19,160 --> 00:10:23,319
Speaker 3: So so I will. I'll defend the vibe coder platforms.

236
00:10:24,679 --> 00:10:28,320
Speaker 3: But the defense now, I mean it's it's it is.

237
00:10:28,400 --> 00:10:31,280
Speaker 3: It's fraudulent, right, I mean like if you say, hey,

238
00:10:31,480 --> 00:10:34,880
Speaker 3: you don't know how to code at all, and uh yeah,

239
00:10:34,960 --> 00:10:37,400
Speaker 3: just sign on to this website and I mean look

240
00:10:37,400 --> 00:10:38,280
Speaker 3: at their marketing page.

241
00:10:38,559 --> 00:10:42,000
Speaker 2: That's exactly what I'm loading loading. It's with a nice

242
00:10:42,000 --> 00:10:44,520
Speaker 2: blue iPhone Air. Oh yeah, beautiful. I have the I

243
00:10:44,559 --> 00:10:47,840
Speaker 2: have the Space Black. Hell yeah, the iPhone app Rise up.

244
00:10:47,920 --> 00:10:49,960
Speaker 2: It's a great phone. It's a great phone. I'm not

245
00:10:50,600 --> 00:10:52,560
Speaker 2: Apple made me pay for it. Yeah, turn your ideas

246
00:10:52,559 --> 00:10:55,079
Speaker 2: into apps? What will you create? The possibilities are endless.

247
00:10:55,240 --> 00:10:57,240
Speaker 2: And then it's a fake prompt that says, make me

248
00:10:57,280 --> 00:11:00,240
Speaker 2: a business tool for marketing teams that helps generate professal

249
00:11:00,360 --> 00:11:04,880
Speaker 2: business proposals and then add automated back up and recovery.

250
00:11:05,280 --> 00:11:07,560
Speaker 2: If I think if you asked Replet to do that,

251
00:11:07,640 --> 00:11:10,160
Speaker 2: it would cost three hundred dollars and nothing would happen.

252
00:11:10,440 --> 00:11:13,000
Speaker 2: I think it would just yeah on count a light

253
00:11:13,200 --> 00:11:14,640
Speaker 2: like barely functional code.

254
00:11:14,720 --> 00:11:17,559
Speaker 3: So I wrote a post on this and I was excited. Yeah,

255
00:11:17,600 --> 00:11:19,559
Speaker 3: and I was. I was excited to end the post

256
00:11:19,600 --> 00:11:21,920
Speaker 3: saying there has never been a successful thing.

257
00:11:21,960 --> 00:11:22,200
Speaker 2: Ever.

258
00:11:23,000 --> 00:11:26,760
Speaker 3: Unfortunately, Replet has added a set of case studies, and

259
00:11:27,000 --> 00:11:29,880
Speaker 3: I think that they use it. And so the case

260
00:11:29,920 --> 00:11:32,480
Speaker 3: studies are we sold to enterprises and we're going to

261
00:11:32,600 --> 00:11:38,640
Speaker 3: do prototypes of internal dev tools, not dev tools, internal

262
00:11:38,720 --> 00:11:42,199
Speaker 3: like you know, management software for inside your back office software.

263
00:11:42,240 --> 00:11:46,440
Speaker 2: So they haven't had a case study since what looks

264
00:11:46,480 --> 00:11:50,680
Speaker 2: like August, and one of them is how Zinus says

265
00:11:50,679 --> 00:11:53,720
Speaker 2: one hundred and forty thousand dollars with replet and it

266
00:11:53,760 --> 00:11:56,160
Speaker 2: but it also cuts development time by fifty percent.

267
00:11:56,280 --> 00:11:59,160
Speaker 3: But so then The question is did the person typing

268
00:11:59,200 --> 00:12:01,280
Speaker 3: stuff into uplet did they know how to code?

269
00:12:01,400 --> 00:12:04,280
Speaker 2: Exactly? See, that's if they know how to code whatever

270
00:12:04,520 --> 00:12:07,000
Speaker 2: the stuff, it's not vibe coding. It's just you could

271
00:12:07,000 --> 00:12:08,880
Speaker 2: have used custom you could use stuff. You could have

272
00:12:08,960 --> 00:12:12,320
Speaker 2: used any vs code or was it what's the free

273
00:12:12,320 --> 00:12:15,440
Speaker 2: one Amazon's doing now? Kiro? Maybe Kiro. And then there's

274
00:12:15,480 --> 00:12:17,920
Speaker 2: the I like the one that came out from China

275
00:12:18,080 --> 00:12:20,439
Speaker 2: and everyone was like, that's gonna send information to that

276
00:12:20,600 --> 00:12:24,840
Speaker 2: Chinese It's like will it? I don't know, but I

277
00:12:24,840 --> 00:12:27,160
Speaker 2: don't think they're going to Your clone of Flappy Bird

278
00:12:27,240 --> 00:12:28,800
Speaker 2: has got to be taken up. Yeah.

279
00:12:30,520 --> 00:12:33,720
Speaker 3: The way that the word vibe coding has the meaning

280
00:12:33,800 --> 00:12:35,920
Speaker 3: that it has today I believe is you do not

281
00:12:36,040 --> 00:12:38,200
Speaker 3: know how to code. You type prompt and you get

282
00:12:39,080 --> 00:12:41,760
Speaker 3: app out. And I'm not going to dock this person

283
00:12:41,760 --> 00:12:44,280
Speaker 3: because they were nice to me once, but like there's

284
00:12:44,280 --> 00:12:46,000
Speaker 3: a person online like they just do like, oh, here's

285
00:12:46,000 --> 00:12:47,760
Speaker 3: one hundred days of AI and I'm going to make

286
00:12:48,040 --> 00:12:52,480
Speaker 3: a fully functional software as a service company fully and

287
00:12:52,520 --> 00:12:54,280
Speaker 3: I don't know how to code. And then you look

288
00:12:54,280 --> 00:12:55,920
Speaker 3: at this person and they're typing in the prompts. It's

289
00:12:55,960 --> 00:12:59,440
Speaker 3: like they clearly have like a pretty strong technical background,

290
00:13:00,040 --> 00:13:01,560
Speaker 3: and then the thing still doesn't work. By the way,

291
00:13:01,679 --> 00:13:02,280
Speaker 3: that's like.

292
00:13:02,559 --> 00:13:05,320
Speaker 2: Cool, Like they know how things work and it's still broken.

293
00:13:05,400 --> 00:13:06,880
Speaker 3: Yeah, so I mean whatever, this person was like a

294
00:13:06,880 --> 00:13:09,520
Speaker 3: product manager or something, so like they know what an

295
00:13:09,559 --> 00:13:11,920
Speaker 3: API is, and they know what a web server is,

296
00:13:11,960 --> 00:13:13,720
Speaker 3: and they know the names of the different technologies, and

297
00:13:13,760 --> 00:13:16,080
Speaker 3: like that's going to get them part of the way there.

298
00:13:16,080 --> 00:13:18,160
Speaker 3: But the idea that you can end to end create

299
00:13:18,200 --> 00:13:20,400
Speaker 3: a software product that has some value.

300
00:13:20,320 --> 00:13:22,560
Speaker 2: Yeah, it's crazy. We would have heard about it. No,

301
00:13:22,720 --> 00:13:25,160
Speaker 2: that's kind of your demo. Workpost was really good about

302
00:13:25,160 --> 00:13:27,120
Speaker 2: this because it was kind of like, look, you can

303
00:13:27,160 --> 00:13:29,440
Speaker 2: do the proof of concept. You can do this, but

304
00:13:29,480 --> 00:13:31,440
Speaker 2: we've never seen the next stage. And someone else did

305
00:13:31,480 --> 00:13:33,320
Speaker 2: a really good one was like shovel weear. They said,

306
00:13:33,400 --> 00:13:36,480
Speaker 2: where's the shovelwear? Where's the crap software? They I remember

307
00:13:36,480 --> 00:13:38,199
Speaker 2: the first times I was on the Internet, the amount

308
00:13:38,200 --> 00:13:41,160
Speaker 2: of weird shareware. Shit, there was just like different forms

309
00:13:41,160 --> 00:13:44,800
Speaker 2: of IRC clients and shit, there were people making weird software.

310
00:13:44,800 --> 00:13:45,760
Speaker 2: Why isn't that happening?

311
00:13:46,320 --> 00:13:49,319
Speaker 3: Yeah, I mean so you'll see, Okay, I made flappy Bird,

312
00:13:49,440 --> 00:13:52,600
Speaker 3: I made a weird thing I made like you can

313
00:13:52,679 --> 00:13:57,000
Speaker 3: make little small pieces of software for yourself that maybe

314
00:13:57,080 --> 00:13:59,560
Speaker 3: have a little bit of value. It is fun, it

315
00:13:59,600 --> 00:14:02,440
Speaker 3: is an novel, yes if you know like.

316
00:14:02,600 --> 00:14:04,920
Speaker 2: But then it then it doesn't work like it's so,

317
00:14:05,000 --> 00:14:05,840
Speaker 2: I do not know.

318
00:14:05,920 --> 00:14:07,840
Speaker 3: I'm a web developer, so I know how to do

319
00:14:07,920 --> 00:14:09,920
Speaker 3: web apps. I call it code for shits though, yeah whatever,

320
00:14:10,000 --> 00:14:11,880
Speaker 3: but like it's that, that's what I know. That's that's

321
00:14:11,880 --> 00:14:14,360
Speaker 3: what I've been doing. I've been coding for whatever, and

322
00:14:14,559 --> 00:14:16,160
Speaker 3: I know how to do that. I do not know

323
00:14:16,200 --> 00:14:18,640
Speaker 3: how to make iPhone apps. So I was like, okay,

324
00:14:18,679 --> 00:14:20,880
Speaker 3: you know what I'm gonna do. They just announced Claude

325
00:14:20,920 --> 00:14:23,640
Speaker 3: whatever because I'm interested in this stuff. I'm an early adopter.

326
00:14:24,040 --> 00:14:26,880
Speaker 2: I don't And also, if it did, what however I

327
00:14:26,920 --> 00:14:29,880
Speaker 2: may feel about AI If it actually did well, if

328
00:14:29,960 --> 00:14:32,800
Speaker 2: Vibe coding was real, that would actually be a huge dal.

329
00:14:32,840 --> 00:14:35,480
Speaker 2: That would be a huge fucking dal. I wouldn't I

330
00:14:35,480 --> 00:14:37,960
Speaker 2: would have all my ethical concerns. If I could actually

331
00:14:38,000 --> 00:14:41,160
Speaker 2: build software without knowing anything, Wow, that would be great.

332
00:14:41,560 --> 00:14:42,280
Speaker 2: Never been the case.

333
00:14:42,400 --> 00:14:44,480
Speaker 3: But you tried, though, well, so I tried. So so

334
00:14:44,560 --> 00:14:46,600
Speaker 3: my idea was like, Okay, I use my phone too much.

335
00:14:46,840 --> 00:14:49,240
Speaker 3: I'm gonna make an app called app snooze. It takes.

336
00:14:49,400 --> 00:14:51,720
Speaker 3: So you say I want Gmail, I want it snooze

337
00:14:51,720 --> 00:14:53,600
Speaker 3: for a half hour. Got it so that when I

338
00:14:53,640 --> 00:14:55,800
Speaker 3: open up the Gmail app, it used this screen time

339
00:14:55,840 --> 00:14:58,040
Speaker 3: thing and it says blocked, and then thirty minutes is up.

340
00:14:58,080 --> 00:15:01,480
Speaker 3: I get it back, right. That is impossible to make

341
00:15:01,720 --> 00:15:05,720
Speaker 3: using iOS essentially without like a substantial amount of work.

342
00:15:05,760 --> 00:15:08,960
Speaker 3: It's it's based on like the limitations of how Apple

343
00:15:09,040 --> 00:15:11,480
Speaker 3: does their stuff with screen time. It just cannot be done.

344
00:15:11,640 --> 00:15:13,880
Speaker 3: So I type this stuff in claude is like sitting there.

345
00:15:13,920 --> 00:15:17,280
Speaker 3: It's like, oh, yeah, you're You're awesome. You are killing me, dude,

346
00:15:17,280 --> 00:15:19,840
Speaker 3: this is a great idea. You got this, you got this, yeah,

347
00:15:19,880 --> 00:15:22,080
Speaker 3: which it actually does say in one off there. I've

348
00:15:22,120 --> 00:15:24,440
Speaker 3: been watching like the World Series or whatever, and a

349
00:15:24,480 --> 00:15:27,160
Speaker 3: lot of NFL and like the chatchypt Ads. We can

350
00:15:27,200 --> 00:15:28,320
Speaker 3: hopefully talk about those.

351
00:15:28,440 --> 00:15:31,160
Speaker 2: I haven't told any of those, Okay, well, oh no, no,

352
00:15:31,240 --> 00:15:35,200
Speaker 2: I love to hear about this because I'm a Raiders fan. Yeah,

353
00:15:35,240 --> 00:15:37,920
Speaker 2: and I try not to watch. If I needed to

354
00:15:37,960 --> 00:15:42,480
Speaker 2: watch a poorly conceived product, I could just use my

355
00:15:42,520 --> 00:15:46,960
Speaker 2: season tickets, but I solved them, so wait, but keep going.

356
00:15:46,800 --> 00:15:49,320
Speaker 3: Though, okay, well but see so it's like you got this,

357
00:15:49,480 --> 00:15:53,560
Speaker 3: but then it's this is the whatever. So they have Haiku,

358
00:15:53,600 --> 00:15:55,840
Speaker 3: and they've sonn It, and they have Opus. Yes, so

359
00:15:55,920 --> 00:15:59,480
Speaker 3: whatever awesome names. But so they have son It, which

360
00:15:59,520 --> 00:16:02,600
Speaker 3: is the really good one. You know, it's very well respected.

361
00:16:02,680 --> 00:16:04,080
Speaker 3: It's supposed to be it's supposed to be the cloud

362
00:16:04,080 --> 00:16:05,320
Speaker 3: is supposed to be a good one for coding. And

363
00:16:05,360 --> 00:16:06,600
Speaker 3: so I was like, I'm gonna pay. I'm gonna pay

364
00:16:06,600 --> 00:16:08,640
Speaker 3: twenty dollars. I'm just gonna see what happens. If I

365
00:16:08,640 --> 00:16:10,200
Speaker 3: can get this thing on the app Store, that'll be great.

366
00:16:10,240 --> 00:16:12,000
Speaker 3: I'm gonna charge ninety nine cents. Let's see if I

367
00:16:12,000 --> 00:16:13,640
Speaker 3: make a hundred bucks. Sure, see if I make my

368
00:16:13,680 --> 00:16:16,560
Speaker 3: Apple Developer account back, Yeah, dump a hundred bucks into

369
00:16:16,560 --> 00:16:18,720
Speaker 3: the Apple Developer account. Awesome on Apple. By the way,

370
00:16:18,920 --> 00:16:21,600
Speaker 3: you can't actually do have the app coding that you

371
00:16:21,600 --> 00:16:23,640
Speaker 3: need to without paying on a hundred bucks. So good,

372
00:16:23,800 --> 00:16:26,680
Speaker 3: that's a business, that's that's Apple bang. But anyways, so

373
00:16:26,720 --> 00:16:28,760
Speaker 3: I do that. I want to make my hundred bucks back,

374
00:16:28,960 --> 00:16:30,560
Speaker 3: but it cannot be done.

375
00:16:30,360 --> 00:16:32,640
Speaker 2: Well, the app you couldn't build the Apple though it sounds.

376
00:16:32,560 --> 00:16:35,960
Speaker 3: But because because of like literal limitations in how iOS

377
00:16:36,120 --> 00:16:38,840
Speaker 3: works in terms of like you can't have a timer

378
00:16:38,880 --> 00:16:41,200
Speaker 3: that goes off and messes with screen time. That's just

379
00:16:41,240 --> 00:16:42,400
Speaker 3: not a thing that Apple was.

380
00:16:42,400 --> 00:16:45,200
Speaker 2: This thing called Brick, whether it's a physical device as well,

381
00:16:45,240 --> 00:16:48,240
Speaker 2: but that feels like a Bluetooth. Something's going on with Brick.

382
00:16:48,280 --> 00:16:50,360
Speaker 2: I don't know. But here's here's the thing as well.

383
00:16:50,400 --> 00:16:53,360
Speaker 2: With all of this, you just made me think it

384
00:16:53,480 --> 00:16:55,520
Speaker 2: is weird that the app doesn't just go yeah, I

385
00:16:55,560 --> 00:16:56,600
Speaker 2: can't build that NTE.

386
00:16:56,640 --> 00:16:58,400
Speaker 3: It'd be nice if it did say that, and it

387
00:16:58,440 --> 00:17:01,600
Speaker 3: was this was weird. I had never observe this behavior before.

388
00:17:01,640 --> 00:17:05,479
Speaker 3: And again I've posted online like okay, you know this,

389
00:17:05,760 --> 00:17:08,440
Speaker 3: you know three bees in Blueberry that thing?

390
00:17:08,920 --> 00:17:10,399
Speaker 2: Oh yeah, yeah.

391
00:17:09,960 --> 00:17:12,800
Speaker 3: Whatever, and people I posted that on LinkedIn and someone

392
00:17:12,880 --> 00:17:15,960
Speaker 3: was like, you are lying, and I post my link. Yeah,

393
00:17:16,240 --> 00:17:18,639
Speaker 3: I post the link to the chat and they're like,

394
00:17:18,680 --> 00:17:20,600
Speaker 3: you had a secret prompt that told it to be stupid.

395
00:17:20,720 --> 00:17:22,280
Speaker 2: Yeah, it's prompt injection. Yeah.

396
00:17:22,320 --> 00:17:24,119
Speaker 3: Yeah, like you have a system prompt that says like

397
00:17:24,240 --> 00:17:26,639
Speaker 3: it's like stupid as ship did.

398
00:17:26,720 --> 00:17:28,359
Speaker 2: Yeah a piece of ship. Dude.

399
00:17:28,359 --> 00:17:31,000
Speaker 3: I messed up. I put be stupid in my system.

400
00:17:31,040 --> 00:17:33,560
Speaker 3: I should have put I should have put be smart.

401
00:17:33,760 --> 00:17:35,200
Speaker 3: If I put smart, it would have worked.

402
00:17:45,359 --> 00:17:47,720
Speaker 2: So on this. You just reminded me when I was

403
00:17:47,760 --> 00:17:49,440
Speaker 2: dicking around with clode code. So I did the story

404
00:17:49,440 --> 00:17:51,280
Speaker 2: a few months ago about how you know, I don't

405
00:17:51,320 --> 00:17:53,040
Speaker 2: know if you've seen like vibrank where it's got like

406
00:17:53,040 --> 00:17:56,280
Speaker 2: people in cloth spend fifty thousand dollars. I love those people.

407
00:17:56,280 --> 00:17:58,520
Speaker 2: I think that's awesome. Well to try this myself, I

408
00:17:58,560 --> 00:18:00,320
Speaker 2: went on. I was like, what is the most token

409
00:18:00,359 --> 00:18:02,280
Speaker 2: intensive software you could build me. It's like, oh yeah,

410
00:18:02,280 --> 00:18:04,280
Speaker 2: an autonomous car and a bet of us. I'm like, cool,

411
00:18:04,280 --> 00:18:08,320
Speaker 2: build all of that. It just sat house just and

412
00:18:08,359 --> 00:18:10,080
Speaker 2: I don't even know what's sped out at the end.

413
00:18:10,560 --> 00:18:12,639
Speaker 3: Well, I mean it certainly, well you could have a

414
00:18:12,680 --> 00:18:14,520
Speaker 3: trillion dollars start up on your hands, but it's just

415
00:18:14,560 --> 00:18:15,720
Speaker 3: you should check out that code.

416
00:18:15,800 --> 00:18:18,800
Speaker 2: It's so sick that these things don't even go like, yeah,

417
00:18:18,840 --> 00:18:21,119
Speaker 2: we can't do that, Like I can't do an autonomous

418
00:18:21,119 --> 00:18:24,160
Speaker 2: car starts up. I don't have any training data. Very basic.

419
00:18:24,240 --> 00:18:25,840
Speaker 3: But if it was, if it was, if that thing

420
00:18:25,960 --> 00:18:28,560
Speaker 3: was smart or useful, right, it has the ability to

421
00:18:28,560 --> 00:18:30,440
Speaker 3: look things up online. Yeah, you should have looked through

422
00:18:30,440 --> 00:18:32,560
Speaker 3: the documentation, and it sort of said, well, what can

423
00:18:32,560 --> 00:18:34,240
Speaker 3: we do with timers? What can we do a screen time?

424
00:18:34,280 --> 00:18:36,159
Speaker 3: Can you hook up a timer to screen time. I

425
00:18:36,400 --> 00:18:37,840
Speaker 3: will let you do this in the background and get

426
00:18:37,880 --> 00:18:40,840
Speaker 3: the half hour time incorrectly, And it's like it's it's

427
00:18:40,880 --> 00:18:43,040
Speaker 3: it's demoware and it allows you to build demoware.

428
00:18:43,040 --> 00:18:44,479
Speaker 2: But it didn't even build a demo of this. Well,

429
00:18:44,840 --> 00:18:45,320
Speaker 2: so now it.

430
00:18:45,280 --> 00:18:47,040
Speaker 3: Built me something I was kind of excited about because

431
00:18:47,040 --> 00:18:49,280
Speaker 3: it it let me pick the app. So I picked Gmail,

432
00:18:49,280 --> 00:18:52,800
Speaker 3: and I picked thirty minutes and then it it it worked.

433
00:18:53,040 --> 00:18:55,640
Speaker 3: Gmail's turned off, right, thirty minutes elapse?

434
00:18:55,840 --> 00:18:58,879
Speaker 2: Right? Gmail is not back on? Oh so you just

435
00:18:58,960 --> 00:19:01,639
Speaker 2: cut Gmail? Well, I mean it's not been in your

436
00:19:01,640 --> 00:19:02,440
Speaker 2: email since.

437
00:19:02,359 --> 00:19:05,560
Speaker 3: No, exactly, Sorry customers, if you have been emailing me,

438
00:19:05,600 --> 00:19:10,000
Speaker 3: it's messed up. No, but it it was like it's

439
00:19:10,040 --> 00:19:13,040
Speaker 3: just lied, right, I mean, And so like that's imagine

440
00:19:13,040 --> 00:19:16,040
Speaker 3: being someone I'm a software developer. Okay, whatever, I'm gonna

441
00:19:16,080 --> 00:19:17,720
Speaker 3: I don't know iOS, but like I'm gonna go on

442
00:19:17,720 --> 00:19:19,400
Speaker 3: the Apple pages and see what's up, and I'm gonna

443
00:19:19,400 --> 00:19:22,560
Speaker 3: ask some meaningful follow ups and determine that this didn't work.

444
00:19:22,560 --> 00:19:24,320
Speaker 3: And Okay, I lost my hundred bucks in the developer

445
00:19:24,400 --> 00:19:26,680
Speaker 3: yeah account. But if you don't know how to code,

446
00:19:26,680 --> 00:19:28,960
Speaker 3: you're going to be like, what are you going to do?

447
00:19:29,359 --> 00:19:32,119
Speaker 2: Well, there's nothing you can do, because the reason I

448
00:19:32,440 --> 00:19:34,520
Speaker 2: read the replic pages and the cursor pages and the

449
00:19:34,560 --> 00:19:37,119
Speaker 2: Curse of One, it's people that can code a little

450
00:19:37,560 --> 00:19:40,399
Speaker 2: be at least a little bit. But replet is just

451
00:19:40,560 --> 00:19:43,200
Speaker 2: it's fifty percent and same with lovables. Read it as well,

452
00:19:43,200 --> 00:19:46,520
Speaker 2: Lovables and other listeners. It's another AI coding platform sold,

453
00:19:46,520 --> 00:19:49,760
Speaker 2: a survived coding thing, and it's all it's fifty percent

454
00:19:49,800 --> 00:19:53,040
Speaker 2: people being like I've spent three hundred dollars, and then

455
00:19:53,119 --> 00:19:55,639
Speaker 2: like ten percent people just lying. People be like I

456
00:19:55,720 --> 00:19:59,560
Speaker 2: just reached twelve thousand M or monthly recurring revenue. It's

457
00:19:59,560 --> 00:20:01,159
Speaker 2: all good for and everyone being like can I see it?

458
00:20:01,160 --> 00:20:04,240
Speaker 2: And they never respond. And then there's the there's the

459
00:20:04,280 --> 00:20:08,040
Speaker 2: people who are like looking for a REDPLT developer, and

460
00:20:08,080 --> 00:20:10,520
Speaker 2: it's like, so you're looking for someone that can write software,

461
00:20:10,840 --> 00:20:14,800
Speaker 2: because write and build software interesting, like a software developer

462
00:20:15,200 --> 00:20:17,880
Speaker 2: might say, and I don't know where to find one. Yeah.

463
00:20:18,280 --> 00:20:21,080
Speaker 3: It's almost like there are like hundreds of thousands, millions

464
00:20:21,119 --> 00:20:23,000
Speaker 3: of people trying to do that, but yeah, we don't

465
00:20:23,000 --> 00:20:26,520
Speaker 3: need them. We can just we talk into the thing

466
00:20:26,640 --> 00:20:29,119
Speaker 3: and turn your app into reality except you.

467
00:20:29,400 --> 00:20:33,040
Speaker 2: It's just it is really crazy how much vibe coding

468
00:20:33,080 --> 00:20:36,359
Speaker 2: is proliferate considering how fucking it's nothing.

469
00:20:37,000 --> 00:20:39,840
Speaker 3: Well so, but so if you need, like if you

470
00:20:39,840 --> 00:20:44,120
Speaker 3: need a prototype, So if like this whole thing boils

471
00:20:44,160 --> 00:20:46,720
Speaker 3: down to if if the expectations were real, if it

472
00:20:46,800 --> 00:20:49,920
Speaker 3: was like, yeah, turn your app, turn your sentence into

473
00:20:49,960 --> 00:20:53,960
Speaker 3: a prototype of an application in minutes. Okay, yeah, like

474
00:20:54,000 --> 00:20:57,720
Speaker 3: an MVP. MVP is like, oh yeah, used to work,

475
00:20:58,560 --> 00:21:00,800
Speaker 3: but well okay, oh sorry, I thought you were saying.

476
00:21:01,200 --> 00:21:02,639
Speaker 2: Hypothetical world where it works.

477
00:21:02,800 --> 00:21:05,040
Speaker 3: Well no, yeah, sorry, Well now in a hypothetical world

478
00:21:05,080 --> 00:21:07,360
Speaker 3: and where it does what it does today, you can

479
00:21:07,400 --> 00:21:10,800
Speaker 3: get like a mock up. If it's a build a

480
00:21:10,840 --> 00:21:14,040
Speaker 3: semi functional wireframe mockup of your application that you could

481
00:21:14,040 --> 00:21:16,240
Speaker 3: show to kind of validate your idea to your friends

482
00:21:16,280 --> 00:21:19,520
Speaker 3: in minutes. Yeah, for thirty dollars or whoever it's your

483
00:21:19,560 --> 00:21:22,720
Speaker 3: credits and up being. That's fine, But does that happen?

484
00:21:22,800 --> 00:21:25,200
Speaker 2: Doesn't it? You could kind of do that. You're lucky,

485
00:21:25,240 --> 00:21:27,160
Speaker 2: it's you roll the dice. That's the thing. It's always

486
00:21:27,200 --> 00:21:30,960
Speaker 2: if you're lucky. There's enough asterisks on this. It's just

487
00:21:31,040 --> 00:21:33,919
Speaker 2: insane that it's got this far because I've read a

488
00:21:34,000 --> 00:21:36,280
Speaker 2: lot of vibe coding articles and if you read like

489
00:21:36,600 --> 00:21:38,760
Speaker 2: Kevin Ruce of course in the Times and people like that,

490
00:21:39,040 --> 00:21:41,640
Speaker 2: you read these articles and you'd think, Wow, you can

491
00:21:41,720 --> 00:21:43,480
Speaker 2: just do this, you can just go and do this.

492
00:21:43,480 --> 00:21:46,679
Speaker 2: This is the future is today. But it's not really

493
00:21:47,720 --> 00:21:48,840
Speaker 2: not really the case at all.

494
00:21:49,040 --> 00:21:53,080
Speaker 3: But I think that the thing that's so like pernicious

495
00:21:53,119 --> 00:21:55,920
Speaker 3: about it is that it's so easy to just say

496
00:21:55,920 --> 00:21:56,960
Speaker 3: skill issue.

497
00:21:57,080 --> 00:22:00,000
Speaker 2: You just two words. You're prompt you're prompting it wrong,

498
00:22:00,160 --> 00:22:02,840
Speaker 2: David devait Yer old thing. But now it's yeah, you're

499
00:22:02,840 --> 00:22:03,520
Speaker 2: prompting it wrong.

500
00:22:03,640 --> 00:22:06,760
Speaker 3: You're prompting it wrong. And so and there's no real

501
00:22:06,960 --> 00:22:09,480
Speaker 3: way to disprove that because can we go back in

502
00:22:09,560 --> 00:22:12,920
Speaker 3: time and like, because it's all this probabilistic stuff. And so,

503
00:22:12,920 --> 00:22:15,679
Speaker 3: so I have a post that I've put up and

504
00:22:15,720 --> 00:22:17,720
Speaker 3: it's it's code doesn't happen to you, that's my thing

505
00:22:17,760 --> 00:22:21,160
Speaker 3: is So it's it's my because you know top programming

506
00:22:21,160 --> 00:22:24,000
Speaker 3: for a while, and so if you're teaching a new programmer,

507
00:22:25,040 --> 00:22:27,040
Speaker 3: sometimes if they have like if they've kind of gotten

508
00:22:27,119 --> 00:22:29,080
Speaker 3: unlucky and they have a bad attitude. They're you know,

509
00:22:29,160 --> 00:22:31,560
Speaker 3: and it's not their fault, but they might think, like

510
00:22:31,640 --> 00:22:34,199
Speaker 3: coding is really mysterious and it's really weird, and I

511
00:22:34,240 --> 00:22:36,480
Speaker 3: type code in and I press run, and it doesn't

512
00:22:36,520 --> 00:22:38,120
Speaker 3: always do what I want, and so I'm just gonna

513
00:22:38,119 --> 00:22:40,760
Speaker 3: like mess around, right and like vibe coding is like

514
00:22:40,800 --> 00:22:43,679
Speaker 3: a productionized version of code happens to you. It's like

515
00:22:43,760 --> 00:22:46,960
Speaker 3: you press button, code pops out, it does a mysterious thing,

516
00:22:48,000 --> 00:22:49,920
Speaker 3: and then like, you know, it's so it's like it's

517
00:22:49,920 --> 00:22:52,280
Speaker 3: like that idea, which was the wrong way to program,

518
00:22:52,520 --> 00:22:54,560
Speaker 3: but like that's the way we're doing it, like and

519
00:22:54,600 --> 00:22:56,280
Speaker 3: we're going to what is the right way, though the

520
00:22:56,320 --> 00:23:00,399
Speaker 3: right way would be. A computer is like you operate

521
00:23:00,440 --> 00:23:03,080
Speaker 3: a computer. You turn it on, You open the coding

522
00:23:03,080 --> 00:23:04,919
Speaker 3: software that you're going to use, whether it's an online

523
00:23:04,960 --> 00:23:07,879
Speaker 3: software like my Wonderful software or you know, something like

524
00:23:07,920 --> 00:23:10,760
Speaker 3: vs code, like something for professionals, whatever, and you type

525
00:23:10,760 --> 00:23:14,040
Speaker 3: in code and you run it, and then the computer

526
00:23:14,480 --> 00:23:18,240
Speaker 3: like executes it runs the code according to the.

527
00:23:17,760 --> 00:23:19,520
Speaker 2: Programming like code is instructions.

528
00:23:19,560 --> 00:23:22,160
Speaker 3: Code is instructions, and the code like happens de terministically

529
00:23:22,200 --> 00:23:24,520
Speaker 3: and maybe if you're, you know, developing a game, maybe

530
00:23:24,520 --> 00:23:27,080
Speaker 3: there's some random elements to the software that you're developing,

531
00:23:27,480 --> 00:23:30,520
Speaker 3: but there's no randomness, like the randomness is under your control.

532
00:23:30,880 --> 00:23:33,439
Speaker 2: Yeah, so it's true. It's the difference between treating it

533
00:23:33,480 --> 00:23:36,520
Speaker 2: as this mystic force that you pull together versus instructions.

534
00:23:36,560 --> 00:23:38,119
Speaker 3: It's instructions and so like if you're if you're a

535
00:23:38,119 --> 00:23:40,960
Speaker 3: really good programmer and maybe you're whatever, maybe you use

536
00:23:41,000 --> 00:23:43,679
Speaker 3: AI to save you some typing and you still have

537
00:23:43,760 --> 00:23:46,119
Speaker 3: that good attitude whatever, like you can use to say typing,

538
00:23:46,160 --> 00:23:46,560
Speaker 3: that's fine.

539
00:23:46,560 --> 00:23:49,719
Speaker 2: I mean, that's the only real vapp like that that

540
00:23:49,760 --> 00:23:51,879
Speaker 2: feels like the only consistent thing is just filling in

541
00:23:51,920 --> 00:23:54,160
Speaker 2: blanks that you know you code yourself. Like it feels

542
00:23:54,160 --> 00:23:56,719
Speaker 2: in it's also correct, and I'm like, I'm which may

543
00:23:56,720 --> 00:23:59,360
Speaker 2: be useful. I'm not gonna lie like I use it.

544
00:23:59,560 --> 00:24:01,840
Speaker 3: Yeah, yeah, like I but I used to have a

545
00:24:01,840 --> 00:24:04,679
Speaker 3: paid GPT account, but I don't trust it to do

546
00:24:04,840 --> 00:24:06,680
Speaker 3: like the models. And this is one of the things

547
00:24:06,680 --> 00:24:08,600
Speaker 3: that I brought up in my post is the models

548
00:24:08,640 --> 00:24:12,800
Speaker 3: like aren't better now right that Well, GPT three was

549
00:24:13,240 --> 00:24:16,880
Speaker 3: okay and GPT four was like much genera and then

550
00:24:16,920 --> 00:24:19,639
Speaker 3: GBT five is trash. Yeah, I mean right, relatively speaking,

551
00:24:19,640 --> 00:24:21,480
Speaker 3: maybe it's a little bit better, and maybe it costs

552
00:24:21,680 --> 00:24:23,400
Speaker 3: open a eye more, which is a big development.

553
00:24:23,440 --> 00:24:26,119
Speaker 2: It was very very great for them and whatever you

554
00:24:26,200 --> 00:24:28,640
Speaker 2: had to cost them less, but it costs more, costs.

555
00:24:28,320 --> 00:24:31,440
Speaker 3: More, but so stop getting better. So that there was

556
00:24:31,480 --> 00:24:32,960
Speaker 3: a time where I was like, I'm gonna I'm gonna

557
00:24:32,960 --> 00:24:34,240
Speaker 3: buy into this. IM an earlier after.

558
00:24:34,119 --> 00:24:35,560
Speaker 2: I'm kind of I'm kind of a booster, Like I

559
00:24:35,760 --> 00:24:39,359
Speaker 2: I was cured going to where's your head? Dot at?

560
00:24:40,240 --> 00:24:42,840
Speaker 2: That's is it? Dot com? It's just where's your adult at? Oh,

561
00:24:42,840 --> 00:24:45,760
Speaker 2: where's your Okay, well, where's your head? Dot At?

562
00:24:45,960 --> 00:24:47,720
Speaker 3: Cured me a little bit of this because I'm like,

563
00:24:48,119 --> 00:24:51,800
Speaker 3: and I've just had some situations where it's just failed

564
00:24:51,840 --> 00:24:54,359
Speaker 3: me so poorly. Like there was a confluence of events

565
00:24:54,400 --> 00:24:56,119
Speaker 3: this summer where I was just like, no, what happened?

566
00:24:56,240 --> 00:24:58,520
Speaker 3: I'm done with this. First of all, I got a

567
00:24:58,560 --> 00:25:03,680
Speaker 3: strong recommendation from uh GPT to buy a software called

568
00:25:03,680 --> 00:25:08,439
Speaker 3: a script, which is like editing, because I have I

569
00:25:08,440 --> 00:25:10,399
Speaker 3: have a YouTube channel and I want to like I

570
00:25:10,560 --> 00:25:12,199
Speaker 3: say a lot of ums and ohs, and maybe I'm

571
00:25:12,200 --> 00:25:15,160
Speaker 3: saying some ums and ohs right now whatever, who cares? Yeah,

572
00:25:15,440 --> 00:25:17,600
Speaker 3: And but I'm like, okay, I'm gonna make this YouTube channel.

573
00:25:17,640 --> 00:25:19,480
Speaker 3: I want the production quality to be decent. If there's

574
00:25:19,480 --> 00:25:22,000
Speaker 3: a shortcut for me, it seems like a I might

575
00:25:22,040 --> 00:25:25,000
Speaker 3: be able to do this. So I'm like to GPT,

576
00:25:25,080 --> 00:25:27,720
Speaker 3: I'm like, what is good AI get rid of ums

577
00:25:27,800 --> 00:25:29,880
Speaker 3: and OS software just.

578
00:25:29,840 --> 00:25:31,480
Speaker 2: Descript kind of like a Google search.

579
00:25:31,440 --> 00:25:34,040
Speaker 3: Yeah, yeah, yeah or whatever. It's an okay Google search

580
00:25:34,119 --> 00:25:36,000
Speaker 3: ye whatever. And so he said you got to use

581
00:25:36,040 --> 00:25:37,480
Speaker 3: the script, and I was like cool. And so I

582
00:25:37,520 --> 00:25:39,360
Speaker 3: put in my credit card twenty bucks or thirty bucks

583
00:25:39,400 --> 00:25:41,560
Speaker 3: or whatever whatever it was, just I put it in

584
00:25:41,600 --> 00:25:43,000
Speaker 3: a recording and just completely mangled it.

585
00:25:43,119 --> 00:25:43,359
Speaker 2: Yep.

586
00:25:43,520 --> 00:25:46,639
Speaker 3: And it's just like the audio was unusable. It was

587
00:25:46,680 --> 00:25:48,320
Speaker 3: it was off by an eighth of a second off

588
00:25:48,320 --> 00:25:50,400
Speaker 3: my voice, and it's just like there's no way there's

589
00:25:50,440 --> 00:25:52,840
Speaker 3: I have no recourse and I'm not an audio engineer,

590
00:25:52,920 --> 00:25:56,720
Speaker 3: and so I just okay, like I vibe, I vibe

591
00:25:56,840 --> 00:25:59,200
Speaker 3: edited my my video and just ruined it.

592
00:25:59,200 --> 00:26:01,480
Speaker 2: It's almost like every promise they make is it's going

593
00:26:01,520 --> 00:26:03,800
Speaker 2: to automate everything. It's like, ah, not really, as long

594
00:26:03,840 --> 00:26:04,800
Speaker 2: as you know what you're doing.

595
00:26:04,920 --> 00:26:06,840
Speaker 3: But this was like a meta level thing where the

596
00:26:06,880 --> 00:26:10,879
Speaker 3: AI recommended me either AI also screwed me over, and

597
00:26:10,920 --> 00:26:13,000
Speaker 3: so I'm like, okay, this is like this may this

598
00:26:13,040 --> 00:26:14,920
Speaker 3: is like because I was using it a search, right yeah,

599
00:26:14,960 --> 00:26:17,400
Speaker 3: and so but it failed me a search because it's

600
00:26:17,440 --> 00:26:21,719
Speaker 3: just emphatically And then you look, because then I was like, well,

601
00:26:21,720 --> 00:26:22,120
Speaker 3: what's wrong?

602
00:26:22,160 --> 00:26:24,040
Speaker 2: Am I? Is it a skill issue? Am I stupid?

603
00:26:24,560 --> 00:26:26,960
Speaker 2: And so I look and Reddit is just filled with

604
00:26:27,080 --> 00:26:28,880
Speaker 2: like this is the worst software, this is the worst

605
00:26:29,280 --> 00:26:33,440
Speaker 2: I used the script very briefly, and all I wanted

606
00:26:33,480 --> 00:26:35,800
Speaker 2: to do was take a bit of audio and turn

607
00:26:35,840 --> 00:26:39,040
Speaker 2: it into a video with the text happening. That's the

608
00:26:39,080 --> 00:26:42,199
Speaker 2: way you read their marketing material. You would look at

609
00:26:42,240 --> 00:26:43,720
Speaker 2: it and think it would take two seconds. Took me

610
00:26:43,720 --> 00:26:46,119
Speaker 2: about forty five minutes, and it was just by the

611
00:26:46,200 --> 00:26:47,679
Speaker 2: end of it, I'm like, I don't even want to

612
00:26:47,720 --> 00:26:50,960
Speaker 2: fucking that. I'm so angry because it's like, this should

613
00:26:50,960 --> 00:26:54,680
Speaker 2: be a button press. The whole point of AI bullshit

614
00:26:54,800 --> 00:26:57,000
Speaker 2: is meant it should be a button press, and it

615
00:26:57,040 --> 00:26:59,320
Speaker 2: never is. Bo wait, well there are other events though.

616
00:26:59,200 --> 00:27:01,959
Speaker 3: Well, so there's that, and so there's that, and then

617
00:27:02,000 --> 00:27:05,600
Speaker 3: it's like, so I'm also I'm a web developer, right,

618
00:27:05,640 --> 00:27:08,560
Speaker 3: and so I'm not very good. I can't program mobile apps.

619
00:27:08,640 --> 00:27:10,400
Speaker 3: That's a thing that I can't do, don't know how.

620
00:27:10,840 --> 00:27:14,200
Speaker 3: I'm also not a very good like infrastructure systems programmer

621
00:27:14,320 --> 00:27:16,880
Speaker 3: whatever that's you know, cloud stuff whatever. I'm not great

622
00:27:16,920 --> 00:27:18,680
Speaker 3: at that, but that is an aspect of my job

623
00:27:18,680 --> 00:27:20,800
Speaker 3: that I have to do. Our website requires some infrastructure

624
00:27:20,840 --> 00:27:23,720
Speaker 3: difficult stuff. Over the years, I've actually gotten quite a

625
00:27:23,720 --> 00:27:26,159
Speaker 3: bit better at that, and so that used to be

626
00:27:26,200 --> 00:27:28,320
Speaker 3: a use case for me for GPT was like, oh,

627
00:27:28,359 --> 00:27:30,840
Speaker 3: it'll I'll ask you some infrastructure related questions, like I

628
00:27:30,840 --> 00:27:32,720
Speaker 3: know how to code, I can piece of I can

629
00:27:32,720 --> 00:27:34,680
Speaker 3: put the puzzle together, and you know, this is actually

630
00:27:34,720 --> 00:27:36,359
Speaker 3: going to save me a little bit of time. But

631
00:27:36,440 --> 00:27:39,720
Speaker 3: I have outpaced GPT's ability and infrastructure developments, so it's okay,

632
00:27:39,720 --> 00:27:42,640
Speaker 3: well I'm doing this project. It's not helping. Yeah, it's

633
00:27:42,680 --> 00:27:44,879
Speaker 3: just wasting my time. Okay, no need for that. The

634
00:27:44,920 --> 00:27:47,439
Speaker 3: descript thing is BS And then I learned from at

635
00:27:47,520 --> 00:27:50,919
Speaker 3: Tetron that this stuff is horrendously expensive. So it's like

636
00:27:51,320 --> 00:27:53,800
Speaker 3: if this was just regular software's service and it costs

637
00:27:53,800 --> 00:27:57,680
Speaker 3: pennies to operate, and you know, it's like kind of helpful.

638
00:27:57,800 --> 00:28:00,280
Speaker 2: Whatever, yeah, be in offensive. It would be like.

639
00:28:00,200 --> 00:28:02,800
Speaker 3: It's fine whatever. There's a company they offered me this

640
00:28:02,880 --> 00:28:05,600
Speaker 3: eight this thing and it didn't work, and you know

641
00:28:05,640 --> 00:28:08,840
Speaker 3: it happens. Yeah, but it's like in the context of

642
00:28:08,840 --> 00:28:11,760
Speaker 3: a world in which this is the future, this is

643
00:28:11,840 --> 00:28:15,280
Speaker 3: magic in a click of a button you get perfect

644
00:28:15,320 --> 00:28:18,639
Speaker 3: audio out if that's the promise in the midst of

645
00:28:18,680 --> 00:28:20,639
Speaker 3: all this and an AI is recommending to it, this

646
00:28:20,760 --> 00:28:24,040
Speaker 3: is like meta leveled like off like dog shits a

647
00:28:24,160 --> 00:28:29,879
Speaker 3: situation and then it's and then it's the dressed disastrously expensive. Yeah,

648
00:28:29,920 --> 00:28:31,399
Speaker 3: like what is the point? What is the point of

649
00:28:31,440 --> 00:28:31,679
Speaker 3: all of this?

650
00:28:31,880 --> 00:28:35,280
Speaker 2: The point is we need to sell GPUs every and

651
00:28:35,440 --> 00:28:39,040
Speaker 2: literally in the car here they announced a seven years,

652
00:28:39,120 --> 00:28:41,400
Speaker 2: thirty eight billion dollar a day between open ai and

653
00:28:41,440 --> 00:28:44,479
Speaker 2: Amazon Web Services. It's just like, why so that we

654
00:28:44,520 --> 00:28:47,320
Speaker 2: can so that they can do sora to more so

655
00:28:47,360 --> 00:28:52,040
Speaker 2: they can generate more copyright infringement. It's and have do

656
00:28:52,120 --> 00:28:54,480
Speaker 2: you have you in the past used these coding models

657
00:28:54,520 --> 00:28:55,840
Speaker 2: a lot or is it just kind of like on

658
00:28:55,880 --> 00:28:56,600
Speaker 2: the side.

659
00:28:56,720 --> 00:29:00,440
Speaker 3: I have so even like I will say, even like

660
00:29:00,440 --> 00:29:03,120
Speaker 3: two weeks ago, I had a very discreete task where

661
00:29:03,120 --> 00:29:05,360
Speaker 3: it's like, in this one situation, I want to do

662
00:29:05,400 --> 00:29:08,719
Speaker 3: this one little thing. Yeah, and I knew exactly what

663
00:29:08,760 --> 00:29:11,240
Speaker 3: it was, and I was lazy, and so I said, right,

664
00:29:11,240 --> 00:29:13,600
Speaker 3: write the code, and so I put in and then

665
00:29:13,600 --> 00:29:15,360
Speaker 3: this is a joke comment and people should do this

666
00:29:15,400 --> 00:29:17,360
Speaker 3: more often. I put in a I paste it in

667
00:29:17,400 --> 00:29:19,480
Speaker 3: the code, and I said, this code comes courtesy of

668
00:29:19,560 --> 00:29:21,880
Speaker 3: chatch ebt. If you have any issues with this software,

669
00:29:21,880 --> 00:29:24,400
Speaker 3: please contact open AI. That's what I wrote in my code, and.

670
00:29:24,400 --> 00:29:27,880
Speaker 2: I should did and it worked great. Whatever. Okay, that's cool.

671
00:29:27,920 --> 00:29:31,040
Speaker 2: It's saved me twenty minutes. That's that's good. And that's

672
00:29:31,080 --> 00:29:33,040
Speaker 2: the thing that's the whole AI bubble. It's like, but

673
00:29:33,120 --> 00:29:36,480
Speaker 2: I'm not a paying subscriber anymore. Oh, that's even worse

674
00:29:36,520 --> 00:29:36,840
Speaker 2: for them.

675
00:29:37,080 --> 00:29:39,520
Speaker 3: Yeah, I know, I just you know, because it's it's

676
00:29:40,360 --> 00:29:42,440
Speaker 3: the stupidest model of theirs could have come up with

677
00:29:42,440 --> 00:29:45,640
Speaker 3: that code because it was so easy, you know, it

678
00:29:45,720 --> 00:29:47,920
Speaker 3: was it was finicky, it was annoying. There are bit

679
00:29:48,000 --> 00:29:50,280
Speaker 3: situations where I f saying, you know, I say, oh

680
00:29:50,280 --> 00:29:51,920
Speaker 3: there's a bug in this code. I paste it in

681
00:29:51,960 --> 00:29:55,400
Speaker 3: and it looks it over and it saved me in

682
00:29:55,480 --> 00:29:58,280
Speaker 3: aggregate tens of hours in the last three years.

683
00:29:58,720 --> 00:30:01,680
Speaker 2: That's fine. It's like, if it was regular sass, I'd

684
00:30:01,720 --> 00:30:02,840
Speaker 2: be like cool, yeah.

685
00:30:02,720 --> 00:30:04,880
Speaker 3: I if it was I think that's kind of part

686
00:30:04,880 --> 00:30:06,800
Speaker 3: of why I canceled the subscription, because if it was

687
00:30:07,560 --> 00:30:09,640
Speaker 3: you know, whatever you needed value your time, you know,

688
00:30:09,680 --> 00:30:11,760
Speaker 3: I could if it was twenty bucks a month and

689
00:30:11,800 --> 00:30:14,600
Speaker 3: it saved me an hour a month, hey, you know yeah,

690
00:30:14,600 --> 00:30:14,960
Speaker 3: it's like.

691
00:30:15,040 --> 00:30:17,760
Speaker 2: Like trip it or flighty, like it's a little bit

692
00:30:17,800 --> 00:30:20,120
Speaker 2: of software we pay for it does a thing. And

693
00:30:19,520 --> 00:30:22,040
Speaker 2: if it was it was ceiling from everyone and burn

694
00:30:22,280 --> 00:30:25,760
Speaker 2: down like it's just it. It only makes sense if

695
00:30:25,760 --> 00:30:28,200
Speaker 2: it was cheap, and it's the literal opposite. If this

696
00:30:28,320 --> 00:30:32,720
Speaker 2: was like cheap, like cheap CPU driven shit, and fine, sure,

697
00:30:32,760 --> 00:30:35,400
Speaker 2: but it's like I one day, I think we're going

698
00:30:35,440 --> 00:30:37,160
Speaker 2: to find out how expensive this is and it's going

699
00:30:37,200 --> 00:30:39,560
Speaker 2: to scare the shit out of people. But you know what,

700
00:30:39,600 --> 00:30:41,240
Speaker 2: that that actually makes me want to move to a

701
00:30:41,240 --> 00:30:43,880
Speaker 2: specific post. You made your scaling laws posts. Let's talk

702
00:30:43,880 --> 00:30:46,720
Speaker 2: about this. So you were a booster at one point

703
00:30:46,760 --> 00:30:49,160
Speaker 2: and you read the stuff short, but you wrote a

704
00:30:49,280 --> 00:30:52,040
Speaker 2: very eloquent piece about the scaling laws about how and

705
00:30:52,120 --> 00:30:54,280
Speaker 2: I've tried to work this into my work, but it's

706
00:30:55,440 --> 00:30:57,200
Speaker 2: we can have I don't know if i'd call it empathy,

707
00:30:57,240 --> 00:30:59,560
Speaker 2: but some understanding of how we got here with the

708
00:30:59,600 --> 00:31:03,040
Speaker 2: AI bubble, because when GPT four came out, it does

709
00:31:03,080 --> 00:31:06,080
Speaker 2: seem like tech people had a reason to be excited.

710
00:31:06,440 --> 00:31:10,720
Speaker 2: I was so excited. What was exciting? It was awesome.

711
00:31:10,960 --> 00:31:12,800
Speaker 3: You talked to it and it was just like this.

712
00:31:13,680 --> 00:31:15,760
Speaker 3: I would ask it coding problems that I found. So

713
00:31:15,840 --> 00:31:17,680
Speaker 3: I was still a teacher at the time, and I

714
00:31:17,720 --> 00:31:20,040
Speaker 3: was like, oh, man, like I have the APE Computer

715
00:31:20,080 --> 00:31:22,400
Speaker 3: Science exam coming up, and I need to like come

716
00:31:22,480 --> 00:31:24,920
Speaker 3: up with practice problems. And I was like generators out

717
00:31:24,960 --> 00:31:27,240
Speaker 3: of thirty practice problems, I obviously read them over it

718
00:31:27,280 --> 00:31:29,120
Speaker 3: to my due diligence, and I, you know, I tried

719
00:31:29,160 --> 00:31:31,120
Speaker 3: the I did a good job putting them together. But

720
00:31:31,160 --> 00:31:34,440
Speaker 3: it's like these are decent, Yeah, these are decent practice problems,

721
00:31:34,480 --> 00:31:36,680
Speaker 3: and like this is this is useful software. I did

722
00:31:36,760 --> 00:31:40,320
Speaker 3: not understand how expensive it was, but there was there's

723
00:31:40,320 --> 00:31:43,280
Speaker 3: a the number of things that would like the hell.

724
00:31:43,280 --> 00:31:45,920
Speaker 2: Will also give you, so the read the listeners don't

725
00:31:45,960 --> 00:31:49,520
Speaker 2: get mad. To be clear, GPT four was twenty twenty three. Yeah,

726
00:31:49,560 --> 00:31:52,360
Speaker 2: we had we were very early in understanding. I mean

727
00:31:52,360 --> 00:31:54,760
Speaker 2: the environmental damage was there, really sure, but they were

728
00:31:54,760 --> 00:31:57,520
Speaker 2: also promisingly effects. But it took a full year until

729
00:31:57,560 --> 00:32:00,320
Speaker 2: June twenty twenty four when they it came out the

730
00:32:00,400 --> 00:32:02,800
Speaker 2: Open Eye would burn five billion dollars. So like early on,

731
00:32:02,880 --> 00:32:05,360
Speaker 2: we didn't really know the costs either. And if I'm

732
00:32:05,400 --> 00:32:08,200
Speaker 2: sure someone will find a fucking link anyway, keep going.

733
00:32:08,200 --> 00:32:08,360
Speaker 4: Well.

734
00:32:08,400 --> 00:32:11,280
Speaker 3: So I was I was pumped up because I saw

735
00:32:11,440 --> 00:32:15,240
Speaker 3: GPT three, was I I you know, I'm a tech person,

736
00:32:15,240 --> 00:32:17,320
Speaker 3: and so I remember seeing early demos a GPT three

737
00:32:17,400 --> 00:32:21,400
Speaker 3: and it was like interesting novelty. It would say stupid

738
00:32:21,400 --> 00:32:22,960
Speaker 3: things and it was kind of cool that it could

739
00:32:22,960 --> 00:32:25,960
Speaker 3: even generate sentences. That was awesome. Three point five came

740
00:32:25,960 --> 00:32:29,040
Speaker 3: out and GPT whatever chat GPT. It's like, that's pretty cool. Yeah,

741
00:32:29,080 --> 00:32:31,480
Speaker 3: I can use it as a search thing and it

742
00:32:31,520 --> 00:32:33,959
Speaker 3: says that I'm good, which I like when people say

743
00:32:34,000 --> 00:32:35,880
Speaker 3: I'm good? Do you like when people say you're good?

744
00:32:36,000 --> 00:32:38,400
Speaker 2: Ed doesn't happen very much, but I think, I, you

745
00:32:38,480 --> 00:32:41,120
Speaker 2: know what, I'll be honest, I knit that there's something

746
00:32:41,320 --> 00:32:45,160
Speaker 2: I think mentally about me where all of the anthwerpomorphization

747
00:32:45,280 --> 00:32:47,600
Speaker 2: pisses not even pissed milf. I'm just like, okay, shut up,

748
00:32:47,640 --> 00:32:49,959
Speaker 2: shut up, shut up. Yeah, it was bullied too much

749
00:32:49,960 --> 00:32:52,040
Speaker 2: as a kid that like, confidences don't work on me anymore.

750
00:32:52,400 --> 00:32:53,240
Speaker 2: I do want to write a.

751
00:32:53,320 --> 00:32:54,560
Speaker 3: I do want to write a thing at some point

752
00:32:54,560 --> 00:32:57,080
Speaker 3: about how if it wasn't chat GPT, if it was

753
00:32:57,120 --> 00:33:00,720
Speaker 3: like box get text yeah, and there was no inn promorphization,

754
00:33:00,960 --> 00:33:02,720
Speaker 3: if it was just like this is a thing that

755
00:33:02,760 --> 00:33:06,320
Speaker 3: can generate usable or interesting or like code for you,

756
00:33:06,360 --> 00:33:09,240
Speaker 3: but there's no chat element to it, that would actually

757
00:33:09,240 --> 00:33:11,440
Speaker 3: make it a lot better to me, like the answer

758
00:33:11,480 --> 00:33:14,120
Speaker 3: pomorphization of like, oh you're talking to a person that

759
00:33:14,280 --> 00:33:15,320
Speaker 3: really makes me mad.

760
00:33:15,480 --> 00:33:17,560
Speaker 2: Yes, And also I find every time it goes you

761
00:33:17,640 --> 00:33:19,400
Speaker 2: got it, you got to just shut up, shut up,

762
00:33:19,440 --> 00:33:20,000
Speaker 2: shut there.

763
00:33:20,120 --> 00:33:22,040
Speaker 3: So in one of these NFL ads, literally I don't

764
00:33:22,080 --> 00:33:24,080
Speaker 3: know if they're like doing a nod to the haters

765
00:33:24,160 --> 00:33:28,440
Speaker 3: or what, but they like, take, we're going, we're going a.

766
00:33:28,440 --> 00:33:30,600
Speaker 2: Couple, make sure the link is in the world.

767
00:33:30,640 --> 00:33:32,480
Speaker 3: Well yeah, no, there's like four of them. Hopefully they

768
00:33:32,480 --> 00:33:35,240
Speaker 3: aren't know you. But uh, it's like a guy doing

769
00:33:35,400 --> 00:33:38,920
Speaker 3: he's trying to do pull ups and it's like here's

770
00:33:38,920 --> 00:33:40,680
Speaker 3: your pull up plan, Like you need to do one

771
00:33:40,720 --> 00:33:42,560
Speaker 3: pull up and then you should do two pull ups,

772
00:33:42,640 --> 00:33:44,760
Speaker 3: and then you should do four or five pull ups,

773
00:33:44,800 --> 00:33:46,280
Speaker 3: and like, I'm actually you will be able to do

774
00:33:46,400 --> 00:33:48,960
Speaker 3: several pull ups and then at the end it's like

775
00:33:49,000 --> 00:33:49,440
Speaker 3: you got this.

776
00:33:50,800 --> 00:33:53,760
Speaker 2: Okay. If so the plan is you do more pull

777
00:33:53,840 --> 00:33:56,240
Speaker 2: ups of a time, you could probably just work that

778
00:33:56,280 --> 00:33:58,840
Speaker 2: out by doing doing pull ups with text in it. Mate,

779
00:33:59,120 --> 00:34:04,000
Speaker 2: I nobody said you got this, but yeah, exactly. Actually no,

780
00:34:04,160 --> 00:34:06,520
Speaker 2: my friend Mac, when I text him about pull ups,

781
00:34:06,520 --> 00:34:08,840
Speaker 2: he says much, He's like, you fucking got this. I

782
00:34:08,880 --> 00:34:12,320
Speaker 2: think he may have. Literally, it's just that's the commercial.

783
00:34:12,400 --> 00:34:14,960
Speaker 2: That's the commercial. That's the commercial. But it's the commercial.

784
00:34:16,040 --> 00:34:18,759
Speaker 2: People watching the NFL and they're like, oh shit, why

785
00:34:18,800 --> 00:34:21,120
Speaker 2: should I use chat GPT. Oh it's gonna tell me

786
00:34:21,120 --> 00:34:23,960
Speaker 2: a pull up plan where I increased from one to several,

787
00:34:24,120 --> 00:34:25,800
Speaker 2: one to seven. You got this? You got this? I

788
00:34:25,840 --> 00:34:26,720
Speaker 2: mean I didn't pause.

789
00:34:26,800 --> 00:34:29,160
Speaker 3: I mean maybe I did not pause the commercial, but

790
00:34:29,800 --> 00:34:32,520
Speaker 3: I it could it could have said some really interesting

791
00:34:32,520 --> 00:34:34,920
Speaker 3: stuff in the middle. I don't know, but the bullet,

792
00:34:34,920 --> 00:34:36,239
Speaker 3: because it has to have a bullet of the list,

793
00:34:36,840 --> 00:34:38,920
Speaker 3: I am pretty sure and I might be lying, and

794
00:34:38,960 --> 00:34:40,840
Speaker 3: so whatever, you know, send me some hate mail, but

795
00:34:40,920 --> 00:34:42,840
Speaker 3: like I'm pretty sure it said like do a couple

796
00:34:43,120 --> 00:34:45,640
Speaker 3: wait a week, you know, drink a protein shake and like,

797
00:34:45,880 --> 00:34:47,000
Speaker 3: you know, do a couple more.

798
00:34:47,600 --> 00:34:50,320
Speaker 2: It's just Google Search, except it makes up the results.

799
00:34:50,360 --> 00:34:52,399
Speaker 2: That's all three fucking years.

800
00:34:52,400 --> 00:34:54,719
Speaker 3: So I have a new idea, which is that it's

801
00:34:54,840 --> 00:34:58,000
Speaker 3: Yahoo Answers but the person has a lobotomy and was like,

802
00:34:58,160 --> 00:34:59,080
Speaker 3: it just did cocaine.

803
00:34:59,160 --> 00:35:02,760
Speaker 2: That's who answers. Yeah, I want it. That's just Yah

804
00:35:03,040 --> 00:35:06,000
Speaker 2: or Cora. But it's like but it's like light speed, yeah,

805
00:35:06,080 --> 00:35:08,960
Speaker 2: like the fast. Well Cora now is GPT like they

806
00:35:09,040 --> 00:35:11,879
Speaker 2: did because Adam DeAngelo is on the board of open Ai.

807
00:35:12,560 --> 00:35:15,560
Speaker 2: So it's just got GPT answers and GPT questions now.

808
00:35:15,840 --> 00:35:32,160
Speaker 2: So cool. But early on it was exciting and there

809
00:35:32,160 --> 00:35:35,120
Speaker 2: were these scaling laws walk me through through the listeners

810
00:35:35,120 --> 00:35:36,280
Speaker 2: who might not understand.

811
00:35:36,680 --> 00:35:39,640
Speaker 3: So yeah, so the post that I wrote, which very

812
00:35:39,760 --> 00:35:41,919
Speaker 3: nicely called eloquence. So if I could pay you twenty

813
00:35:41,920 --> 00:35:43,640
Speaker 3: hours a month to kind of just send me stuff

814
00:35:43,640 --> 00:35:44,480
Speaker 3: like you've got this.

815
00:35:45,160 --> 00:35:47,200
Speaker 2: I'll just I'll email those you got.

816
00:35:46,960 --> 00:35:49,279
Speaker 3: This, Okay, that's great, I'll put it on a schedule. Yeah,

817
00:35:49,280 --> 00:35:50,719
Speaker 3: if you could just do that, I'll pay you twenty

818
00:35:50,760 --> 00:35:54,640
Speaker 3: bucks a month. No, but so so there was an

819
00:35:54,680 --> 00:35:56,840
Speaker 3: idea that if you increase the size of the models.

820
00:35:56,840 --> 00:35:59,320
Speaker 3: I'm not an AI that's right, I'm an AI scientist,

821
00:35:59,320 --> 00:36:00,719
Speaker 3: and so in this post I said, I'm not an

822
00:36:00,719 --> 00:36:03,400
Speaker 3: AI expert or an economist. But like, you look at

823
00:36:03,400 --> 00:36:05,480
Speaker 3: this chart, this chart that they had and you can

824
00:36:05,520 --> 00:36:07,840
Speaker 3: link the thing. And I actually think I cited my sources,

825
00:36:08,440 --> 00:36:12,160
Speaker 3: the original like paper basically about the scaling laws. They

826
00:36:12,200 --> 00:36:15,359
Speaker 3: have this chart that is incredible. It is like, make

827
00:36:15,440 --> 00:36:19,400
Speaker 3: model ten times bigger, get the nice jump in performance.

828
00:36:19,480 --> 00:36:21,920
Speaker 3: Make model ten times bigger, get nice jump in performance.

829
00:36:22,760 --> 00:36:24,440
Speaker 3: And then the idea is like, okay, well if we

830
00:36:24,560 --> 00:36:26,560
Speaker 3: just keep making it ten times bigger, we will get

831
00:36:27,239 --> 00:36:29,040
Speaker 3: who knows how good they can get? Yeah, and it

832
00:36:29,120 --> 00:36:31,480
Speaker 3: kind and it did work, like it was working for

833
00:36:31,520 --> 00:36:33,960
Speaker 3: a minute. That's how they went, to the best of

834
00:36:33,960 --> 00:36:36,279
Speaker 3: my knowledge, that's how they went from three point five

835
00:36:36,320 --> 00:36:36,640
Speaker 3: to four.

836
00:36:36,680 --> 00:36:37,160
Speaker 2: I mean there's a.

837
00:36:37,200 --> 00:36:39,200
Speaker 3: Number of they have smart people over there, yeah, Like

838
00:36:39,239 --> 00:36:41,359
Speaker 3: I mean, we can be honest that like they're doing

839
00:36:41,360 --> 00:36:41,960
Speaker 3: clever stuff.

840
00:36:42,120 --> 00:36:45,319
Speaker 2: Small is also a very subjective. Sure, these are people

841
00:36:45,320 --> 00:36:47,160
Speaker 2: who want experts in mathematics.

842
00:36:47,360 --> 00:36:50,840
Speaker 3: Yeah, they're doing hard, real math and they're getting results.

843
00:36:50,880 --> 00:36:52,319
Speaker 3: Like the fact that it can do what it does

844
00:36:52,400 --> 00:36:52,880
Speaker 3: is incredible.

845
00:36:52,920 --> 00:36:54,640
Speaker 2: Yeah, it's kind of crazy that they can do it.

846
00:36:55,680 --> 00:36:57,440
Speaker 3: If that was all they were saying, If that was

847
00:36:57,440 --> 00:36:58,840
Speaker 3: all they were saying, if they were just like, we

848
00:36:59,000 --> 00:37:02,480
Speaker 3: did research, We've created this incredible piece of technology that

849
00:37:02,480 --> 00:37:04,640
Speaker 3: feels almost alien at this point, I mean, or at

850
00:37:04,640 --> 00:37:06,719
Speaker 3: the point when we discovered it. Now it feels like,

851
00:37:06,800 --> 00:37:08,239
Speaker 3: you know, just we take it for granted that it's

852
00:37:08,280 --> 00:37:10,959
Speaker 3: kind of this trash thing. But like at the moment

853
00:37:11,040 --> 00:37:12,680
Speaker 3: when it was released, it was like, oh my gosh,

854
00:37:12,680 --> 00:37:16,080
Speaker 3: like this is actually crazy. Yeah. And the idea was

855
00:37:16,080 --> 00:37:18,600
Speaker 3: we make it ten times bigger and we will get

856
00:37:18,640 --> 00:37:21,839
Speaker 3: a similar jump in performance, and that is GPT four

857
00:37:21,840 --> 00:37:24,400
Speaker 3: point five. It's just like a footnote in history.

858
00:37:24,680 --> 00:37:26,320
Speaker 2: Oh that was that was released.

859
00:37:26,400 --> 00:37:29,719
Speaker 3: And sam Altman was just like, uh, yeah, well we

860
00:37:29,800 --> 00:37:30,600
Speaker 3: made a big model.

861
00:37:30,640 --> 00:37:33,160
Speaker 2: It was the best announcement ever. I'm actually gonna put it,

862
00:37:33,160 --> 00:37:36,520
Speaker 2: but from what I remember, miss miss Clammy Sammy was like, yeah,

863
00:37:36,560 --> 00:37:38,240
Speaker 2: you know, it's I'm just gonna do it from memory.

864
00:37:38,239 --> 00:37:42,040
Speaker 2: I it's it was like, yeah, good news. It's really

865
00:37:42,080 --> 00:37:46,200
Speaker 2: good for writing bad news. It's really compute intensive. And

866
00:37:46,200 --> 00:37:48,000
Speaker 2: I was like yay.

867
00:37:48,080 --> 00:37:50,719
Speaker 3: And in that announcement I did quote this in the

868
00:37:50,760 --> 00:37:52,800
Speaker 3: post because I don't want to make stuff up and

869
00:37:52,920 --> 00:37:56,279
Speaker 3: like whatever, but they literally said, with each ten x,

870
00:37:56,360 --> 00:37:59,240
Speaker 3: with each order of magnitude, you know, ten x increase

871
00:37:59,280 --> 00:38:02,160
Speaker 3: in model size, we will get an improvement in performance.

872
00:38:02,360 --> 00:38:05,080
Speaker 2: Yes, but like, where's the like worth the big improvement?

873
00:38:05,320 --> 00:38:06,720
Speaker 2: Like it's gone? I don't.

874
00:38:07,000 --> 00:38:09,640
Speaker 3: I think that was them. I think that was the moment.

875
00:38:09,680 --> 00:38:11,759
Speaker 3: I don't know what day they announced four point five,

876
00:38:11,800 --> 00:38:14,799
Speaker 3: but I think every into twenty that was the that

877
00:38:14,960 --> 00:38:17,040
Speaker 3: was a game over. Yeah, and then they did the

878
00:38:17,040 --> 00:38:19,279
Speaker 3: reasoning stuff, and the reasoning stuff was the reasoning.

879
00:38:19,040 --> 00:38:21,920
Speaker 2: Thing was September twenty twenty four. And my favorite thing

880
00:38:21,960 --> 00:38:26,120
Speaker 2: about that was reading all the tech press writing about

881
00:38:26,120 --> 00:38:27,719
Speaker 2: it and being like, can any of you tell me

882
00:38:27,760 --> 00:38:29,640
Speaker 2: what this does? Can any of you tell me why

883
00:38:29,640 --> 00:38:31,719
Speaker 2: this matters to this day? And I'm, by the way,

884
00:38:32,040 --> 00:38:34,520
Speaker 2: I'm not. Actually, it took a minute for me to

885
00:38:34,560 --> 00:38:36,120
Speaker 2: work out what the fuck? And it's just a hat

886
00:38:36,160 --> 00:38:37,799
Speaker 2: on a hat thing. It's like, instead of spitting out

887
00:38:37,800 --> 00:38:40,080
Speaker 2: an output, it goes what would the output be? Oh,

888
00:38:40,160 --> 00:38:42,920
Speaker 2: I will I will go through it and choose these steps,

889
00:38:42,960 --> 00:38:45,760
Speaker 2: which is it's test time compute and it's.

890
00:38:45,600 --> 00:38:48,280
Speaker 3: Mantle and I could have had a moment of reflection

891
00:38:48,400 --> 00:38:51,200
Speaker 3: when the reasoning models came out where because I was.

892
00:38:51,120 --> 00:38:53,880
Speaker 2: Like, okay, it was still the height of the fever, though.

893
00:38:53,760 --> 00:38:56,160
Speaker 3: I know, but I so I asked it a hard question.

894
00:38:56,440 --> 00:38:58,719
Speaker 3: So I did a maths degree and a computer science

895
00:38:58,760 --> 00:39:01,840
Speaker 3: degree in right, So I was like, take this topic

896
00:39:02,000 --> 00:39:06,719
Speaker 3: from sophomore year abstract algebra and do this like visualization

897
00:39:07,000 --> 00:39:08,719
Speaker 3: of the thing, right, And I took one of the

898
00:39:08,760 --> 00:39:10,879
Speaker 3: early reasoning models, which everyone is like, oh, oh one,

899
00:39:10,960 --> 00:39:13,319
Speaker 3: so sure, because it's like you have a PhD level

900
00:39:13,320 --> 00:39:16,560
Speaker 3: thing in the pocket. Okay, so PhD level thing should

901
00:39:16,560 --> 00:39:19,840
Speaker 3: be able to take real sophomore math sophomore and college

902
00:39:19,840 --> 00:39:21,319
Speaker 3: math concept and visualize it.

903
00:39:21,400 --> 00:39:23,000
Speaker 2: Yeah, you should be able to do that. And then

904
00:39:23,040 --> 00:39:26,759
Speaker 2: it didn't, and then I kind of just did it.

905
00:39:26,840 --> 00:39:29,759
Speaker 2: I was kind of just like, oh, I guess I hmm.

906
00:39:30,480 --> 00:39:32,680
Speaker 2: And then I just didn't think about it.

907
00:39:32,719 --> 00:39:35,319
Speaker 3: And then I just kept on kind of hoping that

908
00:39:35,400 --> 00:39:36,560
Speaker 3: something exciting would happen.

909
00:39:36,719 --> 00:39:39,640
Speaker 2: Yeah, and I can a bit of empathy here, I

910
00:39:39,719 --> 00:39:41,759
Speaker 2: get if you're and at that time, so it was

911
00:39:41,800 --> 00:39:44,719
Speaker 2: September twenty twenty four. A month later, they'd raise six

912
00:39:44,719 --> 00:39:46,960
Speaker 2: point six billion dollars to get a credit facility of

913
00:39:47,000 --> 00:39:50,040
Speaker 2: four billion dollars. Like they it looked like open Aye

914
00:39:50,080 --> 00:39:53,319
Speaker 2: was going places. Unless you're like me and you've read

915
00:39:53,360 --> 00:39:56,680
Speaker 2: every single possible financial thing you can get hand on,

916
00:39:56,760 --> 00:39:59,399
Speaker 2: and you've obsessed with the numbers. But I can get

917
00:39:59,440 --> 00:40:02,080
Speaker 2: why someone it was stick within the booster ring might

918
00:40:02,160 --> 00:40:06,000
Speaker 2: not immediately be like fuck because yeah, I don't know,

919
00:40:06,080 --> 00:40:08,080
Speaker 2: people hadn't built things of reasoning, and it did actually

920
00:40:08,160 --> 00:40:10,760
Speaker 2: take a few months for people to work out products

921
00:40:10,760 --> 00:40:11,320
Speaker 2: with reasoning.

922
00:40:11,800 --> 00:40:14,160
Speaker 3: Yeah, I mean, and whatever. I mean, I don't know

923
00:40:14,200 --> 00:40:16,280
Speaker 3: what the improvements were. And they improved on the benchmarks.

924
00:40:16,280 --> 00:40:16,719
Speaker 2: That's fine.

925
00:40:16,760 --> 00:40:20,080
Speaker 3: It's kind of like and I'm sure that the coding

926
00:40:20,080 --> 00:40:21,279
Speaker 3: results are marginally better.

927
00:40:21,760 --> 00:40:23,000
Speaker 2: That's the thing. Though marginally.

928
00:40:23,120 --> 00:40:25,839
Speaker 3: Yeah, it's always margin, but now it's marginally. But that's

929
00:40:25,880 --> 00:40:28,840
Speaker 3: the thing. Three to four was sick, huge jump.

930
00:40:29,000 --> 00:40:29,600
Speaker 2: That was sick.

931
00:40:29,719 --> 00:40:31,960
Speaker 3: That was not marginal. If you had, if you had,

932
00:40:32,040 --> 00:40:33,800
Speaker 3: if your lights were on, if you were paying attention

933
00:40:33,840 --> 00:40:35,319
Speaker 3: and you typed a thing into three and you typed

934
00:40:35,360 --> 00:40:37,120
Speaker 3: a thing into four, you should be impressed.

935
00:40:37,120 --> 00:40:39,359
Speaker 2: Oh, I remember the jump. I wasn't doing better off

936
00:40:39,400 --> 00:40:41,160
Speaker 2: alone at the time, didn't do that until February twenty

937
00:40:41,200 --> 00:40:43,640
Speaker 2: twenty four. But I remember being like, oh that'sk but

938
00:40:43,719 --> 00:40:46,960
Speaker 2: I remember just being like okay, now I was I

939
00:40:47,040 --> 00:40:48,840
Speaker 2: was it was like, wow, we made a computer do

940
00:40:48,880 --> 00:40:49,839
Speaker 2: this and this cool?

941
00:40:49,960 --> 00:40:52,720
Speaker 3: Okay now yeah, And so like I was vaguely aware

942
00:40:52,760 --> 00:40:54,560
Speaker 3: of the of the line chart that I mentioned in

943
00:40:54,560 --> 00:40:57,040
Speaker 3: that post, and so I was like, oh, like all

944
00:40:57,080 --> 00:40:59,120
Speaker 3: they had to do it was like it is a

945
00:40:59,160 --> 00:41:02,600
Speaker 3: freight train toward like actual really cool thing because it's like,

946
00:41:02,800 --> 00:41:05,600
Speaker 3: just make it bigger. And therefore if we just need

947
00:41:05,640 --> 00:41:07,400
Speaker 3: to make it bigger, then we do need more compute

948
00:41:07,440 --> 00:41:08,520
Speaker 3: and the reasoning models.

949
00:41:08,680 --> 00:41:10,960
Speaker 2: Would you finally got another way to throw compute because

950
00:41:10,960 --> 00:41:13,759
Speaker 2: it's the training compute and well compute to generate and

951
00:41:13,800 --> 00:41:15,640
Speaker 2: on so test time compute. Wow.

952
00:41:15,920 --> 00:41:21,200
Speaker 3: Yeah, so like that's that's where the freight train's over. Yeah,

953
00:41:21,200 --> 00:41:23,319
Speaker 3: and I and I just four point five came out,

954
00:41:23,400 --> 00:41:25,600
Speaker 3: didn't really think about it that hard. They started doing

955
00:41:25,640 --> 00:41:27,080
Speaker 3: the reasoning stuff and it's like, okay, well they have

956
00:41:27,080 --> 00:41:29,200
Speaker 3: marginal improvements and they say it did really good on

957
00:41:29,239 --> 00:41:31,759
Speaker 3: a math Olympiad or whatever, and like that's that's interesting.

958
00:41:32,600 --> 00:41:35,640
Speaker 3: But but then it's like another whole year goes by

959
00:41:36,280 --> 00:41:39,120
Speaker 3: and then and then GPT five comes out and like

960
00:41:39,200 --> 00:41:40,319
Speaker 3: what what was that?

961
00:41:40,520 --> 00:41:43,200
Speaker 2: It was nothing and it was so strange.

962
00:41:42,840 --> 00:41:45,080
Speaker 3: And so that that was the final So when I'm

963
00:41:45,080 --> 00:41:46,840
Speaker 3: talking about my confidence and the guns, that cured me

964
00:41:46,880 --> 00:41:49,919
Speaker 3: of my boosterism. Like I started reading your stuff about

965
00:41:49,960 --> 00:41:51,840
Speaker 3: it being expensive. Yeah, but then I was like, this

966
00:41:51,960 --> 00:41:55,480
Speaker 3: is interesting. I've started reading this guy Ed's posts. GPT

967
00:41:55,640 --> 00:41:58,000
Speaker 3: five is coming out next week. I wonder if this

968
00:41:58,000 --> 00:41:59,680
Speaker 3: guy is going to have an extraordinary amount of egg

969
00:41:59,719 --> 00:42:02,320
Speaker 3: on his face. Yeah, like you might have been scared.

970
00:42:02,640 --> 00:42:05,760
Speaker 2: I wasn't because I have the stonewell of the Buddha.

971
00:42:05,840 --> 00:42:10,359
Speaker 2: But it's I was also just like when it when

972
00:42:10,440 --> 00:42:13,000
Speaker 2: reasoning was coming up, going back to twenty twenty three,

973
00:42:13,239 --> 00:42:15,880
Speaker 2: they did was some real shit. The rumors around that

974
00:42:16,320 --> 00:42:19,440
Speaker 2: Q star, Yeah, it was like wit. The reason Samuton

975
00:42:19,520 --> 00:42:21,880
Speaker 2: got fired was they found a terrifying new AI they

976
00:42:21,960 --> 00:42:24,080
Speaker 2: kind of like drummed up. There were leaks about it.

977
00:42:24,120 --> 00:42:26,759
Speaker 2: There were leaks about levels of intelligence. There was all

978
00:42:26,760 --> 00:42:29,640
Speaker 2: of these big leaks. There was really no leaking around

979
00:42:29,680 --> 00:42:32,520
Speaker 2: GPT five other than a Wall Street Journal story towards

980
00:42:32,520 --> 00:42:34,799
Speaker 2: the end of twenty twenty four where it was like, yeah,

981
00:42:34,840 --> 00:42:37,000
Speaker 2: it's costing a shit ton of money. It isn't working

982
00:42:37,160 --> 00:42:40,000
Speaker 2: very well, like the leaks. The reason I because the

983
00:42:40,040 --> 00:42:41,719
Speaker 2: thing is, I mean this to this day. If I

984
00:42:41,760 --> 00:42:43,560
Speaker 2: am wrong about all this, I don't think I am.

985
00:42:43,719 --> 00:42:46,640
Speaker 2: I will admit it. I will explain why. But GPT

986
00:42:46,719 --> 00:42:50,080
Speaker 2: five I wasn't particularly worried about because it did I

987
00:42:50,160 --> 00:42:53,279
Speaker 2: could not fucking tell you what the what it was

988
00:42:53,280 --> 00:42:55,480
Speaker 2: going to be like, no one really if you go

989
00:42:55,600 --> 00:42:57,520
Speaker 2: back to twenty twenty three and you look up GPT

990
00:42:57,680 --> 00:43:00,800
Speaker 2: five stuff, the shit that people are saying is insane.

991
00:43:00,840 --> 00:43:02,800
Speaker 2: That was someone saying it would be completely autonomous and

992
00:43:02,840 --> 00:43:06,319
Speaker 2: it would turn weapons systems against people. There's bonk as ship.

993
00:43:06,880 --> 00:43:08,600
Speaker 2: But getting up to it, yeah, it was kind of

994
00:43:08,600 --> 00:43:11,400
Speaker 2: a proving point, but it was just another fucking model.

995
00:43:11,560 --> 00:43:14,560
Speaker 3: Well, and so that you and I had started exchanging

996
00:43:14,560 --> 00:43:17,200
Speaker 3: emails because whatever, I sorry podcast, and I said them

997
00:43:17,200 --> 00:43:18,600
Speaker 3: on my post over and it was it was it

998
00:43:18,640 --> 00:43:21,000
Speaker 3: was interesting talking to you and then I I when

999
00:43:21,040 --> 00:43:23,880
Speaker 3: that when announcement was going on, I emailed you and

1000
00:43:23,960 --> 00:43:25,640
Speaker 3: you got a lot of emails coming through. But I said,

1001
00:43:26,040 --> 00:43:29,560
Speaker 3: ed ed ed ed ed ed ed. They announced paywalled

1002
00:43:29,719 --> 00:43:30,440
Speaker 3: chat colors.

1003
00:43:30,840 --> 00:43:33,600
Speaker 2: Yes, I go through. No, no, I remember this but go

1004
00:43:33,640 --> 00:43:37,200
Speaker 2: through this. In the announcement of GPT five, the biggest

1005
00:43:37,239 --> 00:43:39,960
Speaker 2: thing ever, They're like, for our paid subscribers, you can

1006
00:43:40,000 --> 00:43:43,840
Speaker 2: turn your chat yellow, which they still haven't released.

1007
00:43:43,920 --> 00:43:46,799
Speaker 3: They still haven't released that. I don't I'm not I'm

1008
00:43:46,840 --> 00:43:49,719
Speaker 3: not a paying subscribers. I've never seen a yellow.

1009
00:43:49,800 --> 00:43:52,319
Speaker 2: I'm paying for chat JPT plus. See you can turn

1010
00:43:52,360 --> 00:43:53,759
Speaker 2: it yellow. I want to see if I can do this,

1011
00:43:53,840 --> 00:44:03,640
Speaker 2: live on it yellow, pink, green, change my window color well, GPT,

1012
00:44:03,880 --> 00:44:06,480
Speaker 2: GPT No, I'm going to ask it because it's fucking insane.

1013
00:44:06,480 --> 00:44:08,800
Speaker 2: If this doesn't searching the web, yes, you can change

1014
00:44:08,920 --> 00:44:13,080
Speaker 2: they did well. They did well. Can you though? On

1015
00:44:13,120 --> 00:44:16,279
Speaker 2: some platforms you can change the accent color. God, this

1016
00:44:16,360 --> 00:44:19,319
Speaker 2: fucking stinks. The fact that you can't ask a product

1017
00:44:19,640 --> 00:44:20,560
Speaker 2: what it does.

1018
00:44:21,000 --> 00:44:24,719
Speaker 3: Well, if you if you can't ask, like, I don't know,

1019
00:44:24,800 --> 00:44:26,520
Speaker 3: what's it good? But I mean, if you can't ask,

1020
00:44:26,560 --> 00:44:28,319
Speaker 3: if you can't type into a Google doc in two

1021
00:44:28,360 --> 00:44:31,440
Speaker 3: thousand and seven, what does Google docs do? That's unsurprising.

1022
00:44:31,480 --> 00:44:33,520
Speaker 2: But that's because Google Docs is a place to write woods.

1023
00:44:33,560 --> 00:44:34,520
Speaker 2: This is meant to answer the thing.

1024
00:44:34,600 --> 00:44:36,399
Speaker 3: Well, no, that's what I'm saying. But it's like they've

1025
00:44:36,680 --> 00:44:39,359
Speaker 3: they claimed that it's this all you know, all all

1026
00:44:39,400 --> 00:44:42,040
Speaker 3: knowing omniscion thing and it cannot tell you how to turn.

1027
00:44:42,400 --> 00:44:43,719
Speaker 3: Shouldn't it have just done it for you?

1028
00:44:43,840 --> 00:44:44,000
Speaker 1: Yeah?

1029
00:44:44,080 --> 00:44:44,960
Speaker 2: H like give me the bus.

1030
00:44:45,239 --> 00:44:47,880
Speaker 3: It should have it should have said edit, great question,

1031
00:44:47,960 --> 00:44:48,959
Speaker 3: would you like it to be blue?

1032
00:44:49,040 --> 00:44:49,400
Speaker 2: Yellow?

1033
00:44:49,480 --> 00:44:54,160
Speaker 3: Pin or or chars are orange? And and like? But

1034
00:44:54,440 --> 00:44:56,680
Speaker 3: where was the But where was That's what it should

1035
00:44:56,680 --> 00:44:57,000
Speaker 3: have done.

1036
00:44:57,040 --> 00:44:59,160
Speaker 2: And also the idea that that was one of the

1037
00:44:59,160 --> 00:45:01,759
Speaker 2: announcements is very very cool. I love the idea that, like,

1038
00:45:01,840 --> 00:45:03,960
Speaker 2: it's the biggest moment ever and you can now make

1039
00:45:04,040 --> 00:45:08,160
Speaker 2: chair GPT brown. It's insane. It's insane.

1040
00:45:08,680 --> 00:45:12,120
Speaker 3: Brown is one of the supportive colors. Probably not next year. Yeah,

1041
00:45:12,200 --> 00:45:15,000
Speaker 3: that that relies on the compute actually the oracle deal.

1042
00:45:15,280 --> 00:45:18,719
Speaker 2: Yeah it's brown, I'm brown. Oh my god. It's so

1043
00:45:18,760 --> 00:45:21,120
Speaker 2: cool that we've built our entire economy on top of

1044
00:45:21,160 --> 00:45:24,440
Speaker 2: this as well. But but the GPT five thing is

1045
00:45:24,800 --> 00:45:28,919
Speaker 2: it was such a weird moment because watching everyone try

1046
00:45:28,960 --> 00:45:30,960
Speaker 2: and be excited about it was really good. There was

1047
00:45:31,000 --> 00:45:33,680
Speaker 2: the whole THEO white weight, not THEO weight, and that's

1048
00:45:34,000 --> 00:45:39,239
Speaker 2: the information THEO there's this fucking guy. Now I'm gonna

1049
00:45:39,280 --> 00:45:41,160
Speaker 2: I really shouldn't have blanked. I've mentioned him. He did

1050
00:45:41,200 --> 00:45:44,040
Speaker 2: a whole thing about GPT five. I'm gonna look this

1051
00:45:44,120 --> 00:45:47,600
Speaker 2: up live on this professional show, where he did a

1052
00:45:47,640 --> 00:45:50,840
Speaker 2: whole thing saying GPT five is the most amazing thing ever,

1053
00:45:51,680 --> 00:45:54,200
Speaker 2: and and then had to be like, yeah, actually yes,

1054
00:45:54,239 --> 00:45:57,000
Speaker 2: THEO THEO Brown. There we go, THEO Brown. He did

1055
00:45:57,000 --> 00:45:59,600
Speaker 2: the thing saying I'm scared of how good GPT five is.

1056
00:45:59,680 --> 00:46:02,400
Speaker 2: Then a week laters like actually it's not the same

1057
00:46:02,400 --> 00:46:05,799
Speaker 2: as when I used it, which is craziest that that

1058
00:46:05,880 --> 00:46:09,120
Speaker 2: should have been a scandal, Like why was that the case?

1059
00:46:09,160 --> 00:46:12,319
Speaker 2: But everyone just kind of moved on. But I I

1060
00:46:12,440 --> 00:46:17,960
Speaker 2: don't know what women to be excited about next GPT six. Well, yeah,

1061
00:46:18,280 --> 00:46:20,280
Speaker 2: it's just But also, what's that meant to? Because JPT

1062
00:46:20,360 --> 00:46:23,120
Speaker 2: five was this weird kind of like myth in the future.

1063
00:46:23,120 --> 00:46:25,319
Speaker 2: It's like when we reach this, everything will get better.

1064
00:46:25,560 --> 00:46:28,520
Speaker 2: But now it's like we're gonna get Claudes on it five.

1065
00:46:28,840 --> 00:46:30,920
Speaker 3: I guess yeah, whatever, I mean, you're gonna get the

1066
00:46:30,920 --> 00:46:32,960
Speaker 3: next one. But that's the thing is like, if it's

1067
00:46:33,000 --> 00:46:37,239
Speaker 3: just continued marginal improvement, what am I? Yeah, what what

1068
00:46:37,239 --> 00:46:39,080
Speaker 3: are we doing? It doesn't make it doesn't make me.

1069
00:46:39,280 --> 00:46:41,000
Speaker 3: That did not make me excited. And yes, it can

1070
00:46:41,040 --> 00:46:44,000
Speaker 3: save software engineers typing time and whatever. I mean, if

1071
00:46:44,040 --> 00:46:45,520
Speaker 3: you know what you're doing, you can get a lot down.

1072
00:46:46,000 --> 00:46:48,319
Speaker 3: I guess that's fine. If that's the way you like

1073
00:46:48,360 --> 00:46:49,880
Speaker 3: to work, If you like to type stuff in and

1074
00:46:49,920 --> 00:46:52,480
Speaker 3: wait on moting screens and get your code out and

1075
00:46:52,520 --> 00:46:54,400
Speaker 3: review it, that's a way to do programming.

1076
00:46:54,440 --> 00:46:58,680
Speaker 2: That's fine. Yeah, And it's it's literally fine.

1077
00:46:58,680 --> 00:47:00,799
Speaker 3: I'm actually saying like it sounds super sarcastic, but like

1078
00:47:00,800 --> 00:47:01,560
Speaker 3: that's literally fine.

1079
00:47:01,600 --> 00:47:04,560
Speaker 2: No, but that's literally fine. Would be if this was

1080
00:47:04,560 --> 00:47:06,960
Speaker 2: a ten billion dollar industry, if they were selling it

1081
00:47:07,000 --> 00:47:10,040
Speaker 2: as like the equivalent of virtualization or like some side

1082
00:47:10,120 --> 00:47:15,040
Speaker 2: thing to the great de cloud compute, not the entire

1083
00:47:15,360 --> 00:47:19,319
Speaker 2: future revenue engine because it isn't. I mean.

1084
00:47:19,520 --> 00:47:22,279
Speaker 3: And so you know, I run a business and I

1085
00:47:22,560 --> 00:47:25,520
Speaker 3: think that So a thing that started people say is

1086
00:47:25,560 --> 00:47:27,880
Speaker 3: they say, like you have product market fit, which is like, oh,

1087
00:47:27,920 --> 00:47:30,719
Speaker 3: your product is good. Yeah, if like one of the

1088
00:47:30,840 --> 00:47:34,440
Speaker 3: criteria is if it went away today, would your users

1089
00:47:34,560 --> 00:47:35,399
Speaker 3: like throw a fit?

1090
00:47:35,640 --> 00:47:36,120
Speaker 2: Yeah?

1091
00:47:36,400 --> 00:47:38,239
Speaker 3: Would you throw a fit of chat GPT you got

1092
00:47:38,360 --> 00:47:40,840
Speaker 3: uninstalled from your front You would not, But like you know,

1093
00:47:40,840 --> 00:47:43,000
Speaker 3: would the general person be that upset, and I don't

1094
00:47:43,000 --> 00:47:43,600
Speaker 3: think they would.

1095
00:47:43,719 --> 00:47:45,880
Speaker 2: I think that there would be a contingent of people

1096
00:47:45,920 --> 00:47:48,600
Speaker 2: who'd be very upset if you have a like parasocial

1097
00:47:48,600 --> 00:47:50,680
Speaker 2: exact the right word. Yeah, if you're like in love

1098
00:47:50,719 --> 00:47:53,080
Speaker 2: with your GPT, then that would be like a death

1099
00:47:53,080 --> 00:47:55,880
Speaker 2: in your family and that's very sad, which would be horrible.

1100
00:47:55,920 --> 00:47:59,080
Speaker 2: And indeed they But it's like I've been saying this

1101
00:47:59,120 --> 00:48:01,560
Speaker 2: for what it's like, say it to boosters, like if

1102
00:48:01,560 --> 00:48:04,880
Speaker 2: this disappeared, would your life change? Would it really change

1103
00:48:04,880 --> 00:48:07,480
Speaker 2: that much? And like, well I used it for baby names.

1104
00:48:07,520 --> 00:48:08,799
Speaker 2: I've used it for it's.

1105
00:48:08,560 --> 00:48:11,319
Speaker 3: Like you named your baby after No. I'm just saying, like,

1106
00:48:11,560 --> 00:48:12,319
Speaker 3: if you like, if.

1107
00:48:12,239 --> 00:48:16,080
Speaker 2: You've got a baby name from chat GBT, that's that's tough. Yeah,

1108
00:48:16,120 --> 00:48:18,400
Speaker 2: that's really bad. The more that was said to me

1109
00:48:18,440 --> 00:48:20,120
Speaker 2: by a booster and the more I think about, the

1110
00:48:20,200 --> 00:48:22,920
Speaker 2: more I'm like, brother, one day, your child is going

1111
00:48:22,960 --> 00:48:24,960
Speaker 2: to hear this, because all they do is they.

1112
00:48:24,920 --> 00:48:27,360
Speaker 3: Sell a book called like the Baby Name Book, and

1113
00:48:27,520 --> 00:48:29,000
Speaker 3: it has like a list of names in it.

1114
00:48:29,080 --> 00:48:31,600
Speaker 2: I don't fucking know read some books like just think

1115
00:48:31,600 --> 00:48:35,080
Speaker 2: about it. Yeah, I'm gonna one of the most important

1116
00:48:35,120 --> 00:48:38,239
Speaker 2: choices the identity of a future human. I'm going to

1117
00:48:38,480 --> 00:48:44,520
Speaker 2: send it to incorrect Google search. Yeah, it's it's depressing,

1118
00:48:44,600 --> 00:48:47,239
Speaker 2: but it's also quite funny because I feel like this

1119
00:48:47,440 --> 00:48:50,839
Speaker 2: era has really revealed who just doesn't know anything about

1120
00:48:50,840 --> 00:48:52,840
Speaker 2: sucking it. The people who are just like will believe

1121
00:48:52,880 --> 00:48:55,640
Speaker 2: anything or will just believe that they are smart at

1122
00:48:55,680 --> 00:48:58,399
Speaker 2: something because of machine told them they are, that they

1123
00:48:58,440 --> 00:49:00,359
Speaker 2: got this. You got to you go.

1124
00:49:00,719 --> 00:49:04,239
Speaker 3: So someone online posted I can't wait for the day

1125
00:49:04,280 --> 00:49:07,080
Speaker 3: when there's an AI agent that'll tell me when my

1126
00:49:07,160 --> 00:49:08,239
Speaker 3: friend's birthdays are.

1127
00:49:09,040 --> 00:49:11,160
Speaker 2: Fuck, there's no other way to do that. There's no

1128
00:49:11,239 --> 00:49:14,120
Speaker 2: way to do that. I don't have like some kind

1129
00:49:14,160 --> 00:49:14,880
Speaker 2: of calendar.

1130
00:49:15,040 --> 00:49:17,560
Speaker 3: No, No, it's going to be a reminder. And so

1131
00:49:17,640 --> 00:49:20,160
Speaker 3: that's like, that's what's happening now. Is that in the

1132
00:49:20,280 --> 00:49:23,000
Speaker 3: like startup space or just people building technology, it's like, well,

1133
00:49:23,000 --> 00:49:25,600
Speaker 3: we're going to get or if you like watch the

1134
00:49:25,600 --> 00:49:28,200
Speaker 3: ads on the NFL or whatever, it's like you were

1135
00:49:28,239 --> 00:49:31,680
Speaker 3: going to agent is going to do the thing that

1136
00:49:31,719 --> 00:49:34,799
Speaker 3: software is supposed to do. Like software. So like I

1137
00:49:34,840 --> 00:49:38,400
Speaker 3: got sold at one point accounting software. Right, that was AI, Right,

1138
00:49:38,520 --> 00:49:41,439
Speaker 3: the AI is going to categorize your transactions. Sure can't

1139
00:49:41,440 --> 00:49:44,400
Speaker 3: do it I bought. I was at a conference and

1140
00:49:44,440 --> 00:49:47,640
Speaker 3: it was, you know, whatever, May May twentieth, May twenty first,

1141
00:49:47,640 --> 00:49:50,440
Speaker 3: May twenty second. I go to Starbucks, I go to pizza,

1142
00:49:50,520 --> 00:49:53,040
Speaker 3: I go to thing. They're all travel related expenses. Yeah,

1143
00:49:53,080 --> 00:49:56,920
Speaker 3: one is travel and then the Starbucks is client conversation.

1144
00:49:59,080 --> 00:50:01,759
Speaker 2: That's what I decided client conversation. And so I I

1145
00:50:02,120 --> 00:50:04,080
Speaker 2: had a meeting with this the founder of the thing,

1146
00:50:04,080 --> 00:50:07,239
Speaker 2: and I was like, dude, like, what are you? What

1147
00:50:07,320 --> 00:50:09,200
Speaker 2: is it? What is this? What did they say? Well,

1148
00:50:09,200 --> 00:50:10,920
Speaker 2: they were just like, you know, sometimes it makes mistakes.

1149
00:50:10,960 --> 00:50:16,800
Speaker 2: We should get on that. Fuck. Yeah, that's that's my accountant. Yeah, you.

1150
00:50:17,560 --> 00:50:19,680
Speaker 2: My whole thing is I know, it's I think with

1151
00:50:19,760 --> 00:50:21,840
Speaker 2: all of this AI coding stuff in the big in

1152
00:50:21,920 --> 00:50:24,799
Speaker 2: the big tech realm, something's gonna break something. Really, but

1153
00:50:24,880 --> 00:50:28,480
Speaker 2: someone's gonna someone's gonna do something stupid. Yeah.

1154
00:50:28,520 --> 00:50:30,719
Speaker 3: Well, so so back to replet. I mean I think

1155
00:50:30,760 --> 00:50:35,200
Speaker 3: that they I kind of hope that they're I don't

1156
00:50:35,200 --> 00:50:36,480
Speaker 3: know if I hope that they're first to go. I

1157
00:50:36,560 --> 00:50:38,440
Speaker 3: mean whatever, they're nice people working there. So that's the

1158
00:50:38,520 --> 00:50:40,520
Speaker 3: unfortunate thing is there are nice people working at these

1159
00:50:40,600 --> 00:50:43,239
Speaker 3: There are people with jobs like it will involp people.

1160
00:50:43,440 --> 00:50:45,440
Speaker 2: I don't I wouldn't want.

1161
00:50:45,280 --> 00:50:47,440
Speaker 3: To ask for people to get laid off who are

1162
00:50:47,480 --> 00:50:49,880
Speaker 3: hardworking people, and some of them are like cool scientists

1163
00:50:49,920 --> 00:50:51,880
Speaker 3: who have studied hard and they're like the nice people.

1164
00:50:52,000 --> 00:50:54,240
Speaker 2: And it's the grimpa of all this is like people

1165
00:50:54,239 --> 00:50:55,399
Speaker 2: are gonna lose fucking job.

1166
00:50:55,480 --> 00:50:57,400
Speaker 3: But it's I mean, it's the executives who you know

1167
00:50:57,400 --> 00:51:00,239
Speaker 3: obviously pissed me off for just yes lying through their teeth, right,

1168
00:51:00,280 --> 00:51:02,680
Speaker 3: I mean, those people deserved it, and but you know,

1169
00:51:02,719 --> 00:51:05,480
Speaker 3: they're never actually going to have a bad outcome happened.

1170
00:51:05,200 --> 00:51:07,080
Speaker 2: To them this, which is why we need to write

1171
00:51:07,080 --> 00:51:09,400
Speaker 2: things to put their it because at some point there

1172
00:51:09,440 --> 00:51:12,080
Speaker 2: needs to be a record of this, of course. Yeah,

1173
00:51:12,160 --> 00:51:14,319
Speaker 2: so I'm going to wrap it there. Charlie whek came

1174
00:51:14,360 --> 00:51:14,920
Speaker 2: people find you?

1175
00:51:15,280 --> 00:51:19,480
Speaker 3: So I have a blog blog dot Charlie Meyer dot co,

1176
00:51:19,680 --> 00:51:21,960
Speaker 3: which is like where kind of my writings go. But

1177
00:51:21,960 --> 00:51:24,640
Speaker 3: I'm also trying to set up a newsletter. So that's

1178
00:51:24,719 --> 00:51:28,360
Speaker 3: Csmeyer dot substack dot com. And my name is spelled

1179
00:51:28,920 --> 00:51:29,960
Speaker 3: m E y e R.

1180
00:51:30,200 --> 00:51:33,160
Speaker 2: Hell yeah, and I of course amed Zytron. You can

1181
00:51:33,200 --> 00:51:35,399
Speaker 2: find me on the internet at Google dot com. That's

1182
00:51:35,440 --> 00:51:37,080
Speaker 2: where I live. I will put all the links to

1183
00:51:37,120 --> 00:51:39,359
Speaker 2: Charlie's stuff of course in the episode notes, but it's

1184
00:51:39,400 --> 00:51:41,480
Speaker 2: good for you to hear it now. And yes, should

1185
00:51:41,480 --> 00:51:43,640
Speaker 2: have a monologue coming up this week. I know I

1186
00:51:43,680 --> 00:51:45,680
Speaker 2: did an announcement where I said I was going to

1187
00:51:45,680 --> 00:51:48,200
Speaker 2: have a big story that is on hold, not because

1188
00:51:48,239 --> 00:51:51,000
Speaker 2: anything went wrong, but because the scale of the information

1189
00:51:51,080 --> 00:51:54,960
Speaker 2: I got has changed dramatically. When I eventually talk about this,

1190
00:51:55,000 --> 00:51:57,000
Speaker 2: it will be a lot of fun. Otherwise, catch you

1191
00:51:57,080 --> 00:52:07,520
Speaker 2: sim Thank you for listening to Better Offline.

1192
00:52:07,680 --> 00:52:10,120
Speaker 5: The editor and composer of the Better Offline theme song

1193
00:52:10,160 --> 00:52:12,840
Speaker 5: is Matasowski. You can check out more of his music

1194
00:52:12,840 --> 00:52:16,480
Speaker 5: and audio projects at Matasowski dot com, M A T

1195
00:52:16,480 --> 00:52:20,960
Speaker 5: T O s O W s ki dot com. You

1196
00:52:21,000 --> 00:52:23,480
Speaker 5: can email me at easy at better offline dot com,

1197
00:52:23,600 --> 00:52:25,920
Speaker 5: or visit better Offline dot com to find more podcast

1198
00:52:25,960 --> 00:52:29,320
Speaker 5: links and of course my newsletter. I also really recommend

1199
00:52:29,320 --> 00:52:31,239
Speaker 5: you go to chat dot Where's youreed dot at to

1200
00:52:31,320 --> 00:52:34,120
Speaker 5: visit the discord, and go to our slash Better Offline

1201
00:52:34,160 --> 00:52:35,240
Speaker 5: to check out our reddit.

1202
00:52:36,040 --> 00:52:37,400
Speaker 2: Thank you so much for listening.

1203
00:52:38,200 --> 00:52:40,920
Speaker 1: Better Offline is a production of cool Zone Media. For

1204
00:52:41,040 --> 00:52:44,200
Speaker 1: more from cool Zone Media, visit our website cool Zonemedia

1205
00:52:44,239 --> 00:52:47,120
Speaker 1: dot com or check us out on the iHeartRadio app,

1206
00:52:47,160 --> 00:52:49,800
Speaker 1: Apple Podcasts, or wherever you get your podcasts.

1207
00:53:02,480 --> 00:53:02,640
Speaker 5: FI