1
00:00:00,085 -->
00:00:03,256
Open the pod bay doors, Hal. Welcome to
Primary Technology,
2
00:00:03,256 -->
00:00:04,857
the show about the tech news that matters.
3
00:00:04,857 -->
00:00:07,148
Big news this week, Google had its Pixel
event.
4
00:00:07,148 -->
00:00:08,999
We're gonna get into all the features that
they announced,
5
00:00:08,999 -->
00:00:12,020
the hardware. Mark Zuckerberg wrote an AI
manifesto,
6
00:00:12,020 -->
00:00:13,701
which sounded like a fever dream.
7
00:00:13,701 -->
00:00:16,792
We'll get into that. Claude is gonna
watermark its writing and you'll
8
00:00:16,792 -->
00:00:19,323
be able to tell whether text was actually
generated with AI.
9
00:00:19,323 -->
00:00:21,204
Spotify's trying to kill podcasts,
10
00:00:21,204 -->
00:00:22,985
YouTube made it harder to get monetized,
11
00:00:22,985 -->
00:00:26,676
and a ton more. This episode is brought to
you by Decagon and you,
12
00:00:26,676 -->
00:00:28,207
all of the members who support us
directly.
13
00:00:28,387 -->
00:00:29,539
I'm one of your host Stephen Robles,
14
00:00:29,539 -->
00:00:31,303
joined by my friend Jason Aten.
15
00:00:31,303 -->
00:00:32,303
How's it going, Jason?
16
00:00:32,299 -->
00:00:34,852
It's good because our our espresso maker
works again.
17
00:00:34,852 -->
00:00:36,069
So I'm f I'm
18
00:00:35,427 -->
00:00:36,427
Now listen,
19
00:00:36,069 -->
00:00:37,069
caffeinated.
20
00:00:37,011 -->
00:00:39,055
it always is a good day when you're
caffeinated.
21
00:00:39,055 -->
00:00:41,550
I'm pre caffeinated. I cold brew it up
before every show.
22
00:00:41,550 -->
00:00:42,893
That's how I'm able to come out the gate.
23
00:00:42,583 -->
00:00:43,583
That does
24
00:00:43,065 -->
00:00:44,390
not look like a cup of coffee,
25
00:00:44,390 -->
00:00:46,879
Steven. It looks like you're drinking
26
00:00:45,099 -->
00:00:46,840
Here's the thing. I don't want to.
27
00:00:46,879 -->
00:00:47,879
a smoothie there.
28
00:00:48,001 -->
00:00:51,392
This is gonna be I'm just anticipating
29
00:00:50,787 -->
00:00:51,787
Put it in a room.
30
00:00:51,392 -->
00:00:54,084
the YouTube. I listen, listen.
31
00:00:54,084 -->
00:00:57,125
We have the fancy ninja espresso maker.
32
00:00:57,125 -->
00:00:59,487
I love doing that. But in the morning,
33
00:00:59,487 -->
00:01:05,600
I drink cold brew because there's very
high I drink concentrated,
34
00:01:03,277 -->
00:01:04,277
Concentrated.
35
00:01:05,600 -->
00:01:06,600
non-diluted
36
00:01:06,481 -->
00:01:09,035
chameleon cold brew. You're supposed to
cut it with water.
37
00:01:09,035 -->
00:01:11,493
I don't cut it with anything. I cut it
with my mouth.
38
00:01:11,493 -->
00:01:14,415
All right. And I drink the chameleon cold
brew.
39
00:01:15,085 -->
00:01:18,716
Put some half and f in there. And I know
there's gonna be some comments about this.
40
00:01:18,716 -->
00:01:20,997
I'm just bracing myself for the YouTube
comments.
41
00:01:20,997 -->
00:01:25,579
I put protein powder in the cold brew and
I blend it up,
42
00:01:25,579 -->
00:01:28,841
and that is my breakfast. And it is the
breakfast of champions,
43
00:01:28,841 -->
00:01:32,002
Jason, because that I turn cold brew into
podcasts.
44
00:01:32,002 -->
00:01:34,283
That's literally what I do. Okay?
45
00:01:34,283 -->
00:01:38,154
And videos. But you know. I'm just I
listen.
46
00:01:38,154 -->
00:01:40,435
Whenever I say I put cold protein powder
in cold brew,
47
00:01:40,435 -->
00:01:42,646
people get upset. But listen, it works.
48
00:01:42,646 -->
00:01:44,487
It works for
49
00:01:43,145 -->
00:01:44,527
I mean, I didn't
50
00:01:44,527 -->
00:01:47,019
even know you went that far. So literally
you are drinking a smoothie.
51
00:01:47,019 -->
00:01:48,201
That's the thing.
52
00:01:47,631 -->
00:01:49,103
I mean kind of, kind
53
00:01:49,103 -->
00:01:50,934
of. You probably know the quote from
today.
54
00:01:50,934 -->
00:01:52,596
It's a movie I've seen.
55
00:01:51,766 -->
00:01:52,888
Yes. I I did
56
00:01:53,129 -->
00:01:55,384
just to be clear, I I knew where the movie
was from,
57
00:01:55,384 -->
00:01:58,060
but I looked it up because I feel like the
rest of the conversation is very good.
58
00:01:58,060 -->
00:01:59,814
But anyway, it's from ~ Space Odyssey,
59
00:01:59,814 -->
00:02:01,327
Stanley Kubrick, two thousand and one.
60
00:02:00,812 -->
00:02:02,303
That's that's exactly right.
61
00:02:01,327 -->
00:02:02,409
Yes.
62
00:02:02,924 -->
00:02:08,330
I saw that movie probably in my early
twenties and it felt like a trip.
63
00:02:08,330 -->
00:02:09,511
Like I don't even know what happened.
64
00:02:09,511 -->
00:02:10,793
I remember there's an obelisk.
65
00:02:10,793 -->
00:02:15,017
I know that the the computer traps in the
sh like there's just crazy stuff going
66
00:02:15,017 -->
00:02:16,017
on there. I need to watch it again,
67
00:02:15,958 -->
00:02:18,020
I think. Probably apropos in today's
world.
68
00:02:18,020 -->
00:02:19,020
And I need
69
00:02:18,294 -->
00:02:19,294
Yeah.
70
00:02:18,711 -->
00:02:19,711
to watch iRobot again.
71
00:02:20,219 -->
00:02:21,219
And Jurassic Park.
72
00:02:21,718 -->
00:02:24,239
Jurassic Park. Okay, we had some great
comments by the way.
73
00:02:24,239 -->
00:02:25,620
Everyone watching on YouTube, thanks for
that.
74
00:02:25,620 -->
00:02:29,051
People said Bending Spoons is the
velociraptor of tech companies,
75
00:02:29,051 -->
00:02:30,422
which is an excellent comment.
76
00:02:30,422 -->
00:02:32,532
Actually, correct. Hope you guys enjoyed
that.
77
00:02:32,532 -->
00:02:34,443
And we have a bunch of five star review
shout-outs to give.
78
00:02:34,443 -->
00:02:36,404
We're so close to five hundred reviews.
79
00:02:36,404 -->
00:02:39,305
Thank you for your help. Douglas Sanchez
from Brazil.
80
00:02:39,305 -->
00:02:41,396
We have several international listeners
and viewers.
81
00:02:41,396 -->
00:02:44,777
We actually a picture of our night in
another continent.
82
00:02:44,777 -->
00:02:48,229
We'll get to that. But he said Jason's
humor is hilarious and they make
83
00:02:48,229 -->
00:02:49,229
a fantastic team. Look at that,
84
00:02:49,209 -->
00:02:51,020
Jason. You got a lot of love in in this
week's
85
00:02:51,400 -->
00:02:55,491
Use. They even use your name. They didn't
even say the other guy.
86
00:02:51,414 -->
00:02:52,816
Yeah, they didn't they even use my name.
87
00:02:52,816 -->
00:02:54,239
They didn't say the other guy.
88
00:02:54,239 -->
00:02:55,841
This is
89
00:02:55,491 -->
00:02:56,491
We have to get
90
00:02:56,201 -->
00:02:58,992
the bring the joke back because the maybe
newer listeners don't know.
91
00:02:58,720 -->
00:03:00,141
This is how I know I've made it.
92
00:02:58,992 -->
00:03:01,733
That's it. That's
93
00:03:00,141 -->
00:03:01,190
I have a name.
94
00:03:01,733 -->
00:03:04,224
it. You have a name. So Far and So Bad
from the USA.
95
00:03:04,224 -->
00:03:05,404
Looks forward to each new episode.
96
00:03:05,404 -->
00:03:08,035
Jeff Woo ninety nine from the USA says
these guys great.
97
00:03:08,035 -->
00:03:10,926
Best tech podcast. I listened to Carpick
from India.
98
00:03:10,926 -->
00:03:13,717
Yes, I finally tried Raycast and he gave
us five stars for that.
99
00:03:13,717 -->
00:03:15,197
I have an entire video on Raycast.
100
00:03:15,197 -->
00:03:17,178
Do you what are you a spotlight guy?
101
00:03:17,178 -->
00:03:18,208
Do you have a special launcher?
102
00:03:18,249 -->
00:03:19,566
~ I use Alfred.
103
00:03:20,844 -->
00:03:22,800
Alright, well tune into the next Mac Power
users,
104
00:03:22,800 -->
00:03:24,425
okay, Jason? Cause me and David
105
00:03:23,658 -->
00:03:24,931
I'm not s I'm not switching.
106
00:03:25,289 -->
00:03:27,716
we go head to head. I got some thoughts,
107
00:03:27,716 -->
00:03:28,716
okay.
108
00:03:27,851 -->
00:03:28,851
It's just not
109
00:03:28,491 -->
00:03:30,582
it's not important enough to me to think
about switching.
110
00:03:30,582 -->
00:03:34,615
I do have ~ sh I do have Spotlight
assigned to a different hotkey so
111
00:03:34,615 -->
00:03:37,636
I can still pull it up because the new
version of Spotlight's better.
112
00:03:37,636 -->
00:03:40,708
I just Yeah, yeah, yeah,
113
00:03:38,178 -->
00:03:42,198
~ I mean better than before. Yeah.
114
00:03:40,708 -->
00:03:41,839
yeah, yeah. That's all I'm saying.
115
00:03:41,839 -->
00:03:42,839
It's better
116
00:03:42,209 -->
00:03:44,960
than before 'cause well, it's like
Spotlight was incredible,
117
00:03:44,960 -->
00:03:47,402
then it was useless, and now it's
118
00:03:47,011 -->
00:03:48,011
Sure.
119
00:03:47,402 -->
00:03:48,402
like fine.
120
00:03:48,332 -->
00:03:49,812
Well, and here's here's my thing.
121
00:03:49,812 -->
00:03:54,135
Spotlight was ~ meh. And then in macOS 26,
122
00:03:54,135 -->
00:03:55,715
it's really good. Clipboard history,
123
00:03:55,715 -->
00:03:58,097
file search, quick keys, all very good.
124
00:03:58,097 -->
00:04:03,490
But the macOS 27 Spotlight, I feel like it
gets slower because you got Siri
125
00:04:03,490 -->
00:04:06,441
AI in there at the same time. And they're
always like whenever
126
00:04:06,441 -->
00:04:07,802
you type something into the new spotlight,
127
00:04:07,802 -->
00:04:10,883
Mac OS 27 or on your iPhone, it's always
like fighting.
128
00:04:10,883 -->
00:04:13,445
Like, is this an AI thing? Is he just
trying to open an app?
129
00:04:13,445 -->
00:04:15,826
What are we doing here? And so I talk
about that.
130
00:04:16,256 -->
00:04:20,323
Yeah, I use Alfred and I'm probably not
ever gonna switch unless they stop making
131
00:04:20,323 -->
00:04:25,052
it and whatever because I use it thirty
nine thousand times a day and I just
132
00:04:24,546 -->
00:04:25,546
Yeah, yeah.
133
00:04:25,052 -->
00:04:27,387
cannot afford the mental overhead of
switching to something else.
134
00:04:27,628 -->
00:04:29,142
We're gonna talk about Alfred and
135
00:04:29,142 -->
00:04:30,876
launchers in the bonus episode today.
136
00:04:30,876 -->
00:04:33,232
It's it's been decided. It's now been
decided.
137
00:04:32,692 -->
00:04:34,076
Okay, good. I love
138
00:04:34,076 -->
00:04:36,884
when I can talk you into a bonus episode ~
topic.
139
00:04:35,916 -->
00:04:38,168
Perfect. Perfect. I one always comes one's
always comes up.
140
00:04:36,884 -->
00:04:38,669
Yes.
141
00:04:38,168 -->
00:04:39,366
~
142
00:04:39,369 -->
00:04:42,451
Chris from Washington said, I'm in the
camp of assigning apps
143
00:04:42,451 -->
00:04:45,203
to particular spaces and hope hopping
between them,
144
00:04:45,203 -->
00:04:47,154
whether I'm on a work machine or personal
machine.
145
00:04:47,154 -->
00:04:49,045
So yeah, I'd kind of do that too.
146
00:04:49,045 -->
00:04:51,817
I'm doing it right now. We talked about
spaces on macOS last week.
147
00:04:51,817 -->
00:04:57,361
Right now, I have my primary tech podcast
space where I have Notion,
148
00:04:57,361 -->
00:05:01,324
Riverside, Safari tab group with all the
links I'm going be sharing today,
149
00:05:01,324 -->
00:05:03,636
and a PDF of the sponsor read.
150
00:05:03,926 -->
00:05:05,436
That's my space. And then on the other
space,
151
00:05:05,436 -->
00:05:07,297
I got ~ work stuff. I got final code over
here.
152
00:05:07,297 -->
00:05:09,147
So yeah. You're still not a spaces guy,
153
00:05:09,147 -->
00:05:12,118
right? You haven't tried? Not open right
now,
154
00:05:09,452 -->
00:05:12,652
You have final cut right now? Okay.
155
00:05:12,118 -->
00:05:13,519
but usually that's open. No, no.
156
00:05:12,652 -->
00:05:16,385
Okay. That's the whole point of the dots.
157
00:05:13,519 -->
00:05:16,029
I quit I quit all that. I'm not a monster.
158
00:05:16,890 -->
00:05:18,530
Yeah, I know. I'm not a monster.
159
00:05:18,530 -->
00:05:21,551
essay Digital from the USA gave us six out
of five stars.
160
00:05:21,551 -->
00:05:24,262
Thank you for that. And podcast seven two
eight had a question.
161
00:05:24,262 -->
00:05:27,032
Do we listen to anything in the shower and
what books are we reading?
162
00:05:27,032 -->
00:05:28,113
A little little bit different.
163
00:05:28,113 -->
00:05:30,693
I will say rarely do I listen to something
in the shower,
164
00:05:30,693 -->
00:05:32,134
but when I do, it's a podcast.
165
00:05:32,134 -->
00:05:33,134
And
166
00:05:33,441 -->
00:05:38,854
What books are we reading? I actually just
started reading what is the book?
167
00:05:38,854 -->
00:05:42,333
I think it's a like a Seneca Stoicism
book.
168
00:05:42,333 -->
00:05:45,258
I forget what it's called on the Jason's
face.
169
00:05:45,258 -->
00:05:49,370
It just gave me a crazy look. there's that
and there was I forgot
170
00:05:49,370 -->
00:05:50,881
the books I're reading. I'll have to bring
them for next time.
171
00:05:50,881 -->
00:05:52,222
Are you reading anything, Jason?
172
00:05:52,502 -->
00:05:57,284
Well, I mean I do read I do read books.
173
00:05:57,284 -->
00:05:58,384
I don't read them in the shower.
174
00:05:58,384 -->
00:06:01,685
Just to be clear.
175
00:05:58,986 -->
00:06:00,931
No, no. Two different questions.
176
00:06:00,931 -->
00:06:02,164
Yeah.
177
00:06:01,685 -->
00:06:02,825
~
178
00:06:02,945 -->
00:06:04,606
I are you reading the Daily Stoic?
179
00:06:04,606 -->
00:06:06,006
Are you reading
180
00:06:04,959 -->
00:06:05,981
No no no no no.
181
00:06:06,006 -->
00:06:11,488
Let Letters from a Stoic? Meditations by
Marcus Aurelius.
182
00:06:08,064 -->
00:06:10,537
No, I already read Stillness is the Key.
183
00:06:10,537 -->
00:06:16,276
I've read some of that. There's I don't
know,
184
00:06:11,488 -->
00:06:15,820
Why do I know this many stoicism books?
185
00:06:15,820 -->
00:06:16,820
I can't figure it out.
186
00:06:16,276 -->
00:06:17,276
it's a it's
187
00:06:16,766 -->
00:06:21,784
a book it's the Seneca like life is short
but long if you know how to
188
00:06:21,784 -->
00:06:24,598
use it or whatever. Like whatever that
book is.
189
00:06:24,998 -->
00:06:26,829
That sounds a lot like Zuck's Manifesto.
190
00:06:26,829 -->
00:06:29,922
We'll get to that later. ~ I listen to
podcasts sometimes in the shower.
191
00:06:29,922 -->
00:06:31,823
That's about it. I don't listen to music
in the shower.
192
00:06:31,296 -->
00:06:32,380
Right. Same.
193
00:06:31,823 -->
00:06:32,823
I just
194
00:06:32,404 -->
00:06:34,325
sometimes will listen to a podcast.
195
00:06:34,325 -->
00:06:38,669
~ It's kinda weird because I have a lot of
podcast listening time.
196
00:06:38,669 -->
00:06:40,160
I don't really have to do that in the
shower.
197
00:06:40,160 -->
00:06:43,523
Also, I can only do it in the shower if
there's like no one else around because
then
198
00:06:43,434 -->
00:06:46,076
Yeah. R
199
00:06:43,523 -->
00:06:44,954
they just think I'm talking to some dudes.
200
00:06:46,076 -->
00:06:49,839
Now if you were, yeah. ~ On the Shortness
of Life is the book,
201
00:06:49,839 -->
00:06:53,442
from Seneca. So just I actually want to
get the name there.
202
00:06:53,090 -->
00:06:54,090
Great. I
203
00:06:53,480 -->
00:06:57,145
and the last book I've read is I just it's
taken me a while because I have read
204
00:06:57,145 -->
00:06:59,537
it in spurts. I just finished Apple and
China,
205
00:06:59,537 -->
00:07:00,537
so
206
00:07:00,596 -->
00:07:02,667
there you go. Yeah, I'm in the middle of
that one still.
207
00:07:02,667 -->
00:07:05,629
And ~ my wife and I just finished The Body
Keeps the Score,
208
00:07:05,629 -->
00:07:08,380
which is a heavy book, but is is very
good.
209
00:07:08,380 -->
00:07:11,793
All right, I wanted to thank the pins have
made it to more places.
210
00:07:11,793 -->
00:07:16,555
Andre in South Africa sent me an this
incredible photo.
211
00:07:16,555 -->
00:07:19,917
It is the medieval night riding the polar
bear with Table Mountain
212
00:07:20,318 -->
00:07:22,199
in the background in Cape Town,
213
00:07:22,199 -->
00:07:26,661
South Africa. Amazing. The night is
traversing the globe.
214
00:07:26,661 -->
00:07:27,762
Super fun.
215
00:07:28,474 -->
00:07:33,855
And Steve sent me an email. He and ~ his
spouse went to the Tour de France
216
00:07:34,235 -->
00:07:37,076
and took the medieval knight on the polar
bear,
217
00:07:37,076 -->
00:07:40,897
has a picture of it outside a French Apple
store,
218
00:07:40,897 -->
00:07:42,387
and sent me a bunch of other pictures.
219
00:07:42,387 -->
00:07:44,418
And so very, very kind, very fun.
220
00:07:44,418 -->
00:07:45,918
We're all over the world, Jason.
221
00:07:45,918 -->
00:07:47,719
The knight is making
222
00:07:47,617 -->
00:07:48,617
I love it.
223
00:07:47,719 -->
00:07:48,719
it.
224
00:07:48,299 -->
00:07:51,119
Yeah, it's very cool. on one funny post I
want to say,
225
00:07:51,119 -->
00:07:54,920
you know, we sometimes try to say hey
dingus to not activate everyone's voice
226
00:07:54,920 -->
00:07:57,901
assistance, and then we forget and we say
it a thousand times in an episode anyway.
227
00:07:58,293 -->
00:08:01,435
But Chipionic found a car that has a
license plate,
228
00:08:01,435 -->
00:08:03,360
a vanity license plate that says Dingus.
229
00:08:03,648 -->
00:08:04,648
How bad
230
00:08:04,019 -->
00:08:06,075
do you want to just drive up rolling on
your window and be like,
231
00:08:06,075 -->
00:08:07,075
hey dang us?
232
00:08:06,921 -->
00:08:07,921
Hey Dingus.
233
00:08:07,822 -->
00:08:10,203
That's the that was perfect. So yeah,
234
00:08:10,203 -->
00:08:11,203
that exists in the real world.
235
00:08:11,084 -->
00:08:12,705
All right. There's a ton of pixel news.
236
00:08:12,705 -->
00:08:17,338
Google's hardware, software. Is Alex
Cooper really gonna switch from an iPhone
237
00:08:17,338 -->
00:08:18,989
to an Android? I wanna talk about that.
238
00:08:18,989 -->
00:08:22,251
Because the Google the Made by Google
event was is quite a thing.
239
00:08:22,251 -->
00:08:25,613
So, but real quick, just want to plug
before we get into that.
240
00:08:25,613 -->
00:08:27,855
Apple Immersive is gonna get more content.
241
00:08:27,855 -->
00:08:31,337
MLB games, since we're basically a sports
ball podcast.
242
00:08:31,989 -->
00:08:35,013
Major League Baseball, Friday night
baseball is gonna have several games
243
00:08:35,013 -->
00:08:38,263
in Apple Immersive and while I don't you
know,
244
00:08:38,263 -->
00:08:42,202
I'm not a big sports guy, I am from New
York and I am a Yankees
245
00:08:42,343 -->
00:08:45,146
fan just because that's required if you
live in New York,
246
00:08:45,146 -->
00:08:47,209
unless you're a Mets fan, but nobody's
perfect.
247
00:08:47,209 -->
00:08:49,311
just kidding. I probably just made a bunch
of people mad.
248
00:08:49,311 -->
00:08:50,491
I'm so sorry. I don't know.
249
00:08:50,262 -->
00:08:51,262
Yeah,
250
00:08:50,606 -->
00:08:52,177
yeah, that's not the other guy.
251
00:08:52,177 -->
00:08:53,177
I just want to be clear.
252
00:08:53,089 -->
00:08:54,369
I don't know, it's not the other guy.
253
00:08:54,369 -->
00:08:58,070
But yeah, you'll be able to watch the
Yankees Red Sox game in Apple Vision
254
00:08:58,070 -->
00:08:59,951
Pro live as well as other games.
255
00:08:59,951 -->
00:09:04,782
I'm curious how this will go. Ben Thompson
was already upset without this even
256
00:09:04,782 -->
00:09:09,854
happening yet, because if you had watched
the NBA games in Apple Vision Pro,
257
00:09:09,854 -->
00:09:12,354
they switched the camera angle throughout
the game,
258
00:09:12,354 -->
00:09:14,595
and his argument is that could be jarring.
259
00:09:14,595 -->
00:09:17,436
Just stick me in one place as though I am
actually experiencing it,
260
00:09:17,436 -->
00:09:21,597
which I totally get. I watched an NBA game
in Apple Vision Pro as well.
261
00:09:21,597 -->
00:09:22,597
And
262
00:09:22,569 -->
00:09:25,356
I'm curious how they do it. Are they going
to be switching you all over the park?
263
00:09:25,356 -->
00:09:27,361
Are you going to able to choose a seat?
264
00:09:27,361 -->
00:09:29,669
Probably not, but yeah. I don't know.
265
00:09:29,669 -->
00:09:30,669
I'm going to I'm going to watch it.
266
00:09:30,590 -->
00:09:31,773
I'm going to put on Vision Pro to watch
it.
267
00:09:32,115 -->
00:09:33,716
I wanna just say and I don't I mean,
268
00:09:33,716 -->
00:09:35,678
he's probably right as often as he's not,
269
00:09:35,678 -->
00:09:38,120
but Ben Thompson is a hundred percent
right on this.
270
00:09:38,120 -->
00:09:40,963
Apple has no idea what they're doing with
this.
271
00:09:40,963 -->
00:09:45,037
It's such a stupid idea. They're gonna
have bespoke commentators.
272
00:09:45,037 -->
00:09:46,448
~ you just listen,
273
00:09:45,504 -->
00:09:46,504
Yeah, that's not good.
274
00:09:46,448 -->
00:09:48,312
go l subscribe to Dithering and just
listen to it.
275
00:09:48,312 -->
00:09:49,450
I don't have to get into the whole thing.
276
00:09:49,450 -->
00:09:50,450
But the
277
00:09:49,537 -->
00:09:50,537
Yeah, yeah.
278
00:09:49,981 -->
00:09:54,215
point is you have the most incredible way
to feel like you're actually at the game,
279
00:09:54,215 -->
00:09:57,758
but they're treating it like it's a two D
broadcast and they're gonna just
280
00:09:57,268 -->
00:09:58,268
Right.
281
00:09:58,078 -->
00:10:00,630
And then you're gonna just move you all
over where you should just get to
282
00:10:00,630 -->
00:10:02,981
be able to pick a spot. It it should be
like theater.
283
00:10:02,005 -->
00:10:03,863
Right. Yes.
284
00:10:02,981 -->
00:10:04,312
Isn't that Lystigoris?
285
00:10:04,312 -->
00:10:06,804
Where you can choose where you wanna be
sitting in the theater.
286
00:10:06,227 -->
00:10:09,405
Yes, exactly. Yeah.
287
00:10:06,804 -->
00:10:09,415
You should just be able to choose where
you wanna be at at the game
288
00:10:09,415 -->
00:10:11,096
and you can just watch the whole game.
289
00:10:11,096 -->
00:10:12,417
And if you wanna have a replay,
290
00:10:12,417 -->
00:10:14,459
you just hit rewind and move to a
different seat.
291
00:10:14,459 -->
00:10:15,459
Like
292
00:10:14,953 -->
00:10:15,953
Yeah.
293
00:10:15,709 -->
00:10:16,709
this is so simple.
294
00:10:16,935 -->
00:10:20,838
And I would I long thought, and I actually
sent the team,
295
00:10:20,838 -->
00:10:22,679
I think, this suggestion or whatever,
296
00:10:22,679 -->
00:10:26,082
like orchestra concerts, musicals,
297
00:10:26,082 -->
00:10:31,737
whatever. If you if they had a deal with
some venue or a number of venues where
298
00:10:31,737 -->
00:10:34,609
you could buy a ticket to watch an event
live,
299
00:10:34,609 -->
00:10:37,521
whether it's you know, a pop music type
concert,
300
00:10:37,521 -->
00:10:39,473
if it's actually a stage musical,
301
00:10:39,473 -->
00:10:44,577
rock concert, whatever, and you just get
the center seat.
302
00:10:45,161 -->
00:10:49,293
A little ways back and you can experience
the concert right there,
303
00:10:49,293 -->
00:10:53,505
live, with the live crowd around you and
hearing everything that's going on.
304
00:10:53,505 -->
00:10:56,556
I feel like that would be a particular
selling point.
305
00:10:56,556 -->
00:10:59,627
One, people just can't get out to events
like that.
306
00:10:59,627 -->
00:11:06,060
Artists like I follow ~ Leve because sh ~
my oldest son really likes
307
00:11:06,060 -->
00:11:08,371
her as an artist. I like her music as
well.
308
00:11:08,371 -->
00:11:12,353
She's like a jazz artist. But she doesn't
come to the US super often to tour.
309
00:11:12,921 -->
00:11:15,063
And she's only been here like once in the
last couple of years.
310
00:11:15,063 -->
00:11:16,204
And like that would be an artist for like,
311
00:11:16,204 -->
00:11:18,546
man, it would be awesome if I could
experience a live concert.
312
00:11:18,546 -->
00:11:22,260
My son could experience one in Apple
Vision Pro because you can do that.
313
00:11:22,260 -->
00:11:23,311
It doesn't matter where she was.
314
00:11:23,311 -->
00:11:24,812
You just stick one of the immersive
cameras.
315
00:11:24,812 -->
00:11:27,915
And you just need one then. And I feel
like you'd be able to multiply your
efforts
316
00:11:28,035 -->
00:11:32,469
by just putting one immersive camera in a
venue where you can buy that seat
317
00:11:32,469 -->
00:11:34,511
as if you're there. Because when you're
actually at an event,
318
00:11:34,511 -->
00:11:35,822
you just have that one perspective.
319
00:11:35,822 -->
00:11:37,514
You can't move around. Just do that.
320
00:11:37,514 -->
00:11:38,514
So I don't know.
321
00:11:38,240 -->
00:11:39,240
Yeah,
322
00:11:38,450 -->
00:11:41,648
if only there was like an international
monopoly where you could make a deal
323
00:11:41,648 -->
00:11:44,245
to put these cameras in every single venue
in America.
324
00:11:45,068 -->
00:11:47,205
See, the joke there is Ticketmaster
325
00:11:47,221 -->
00:11:49,413
Right. Live Nation is literally that.
326
00:11:49,413 -->
00:11:50,604
Just I guarantee
327
00:11:50,116 -->
00:11:51,116
That monopoly.
328
00:11:50,604 -->
00:11:53,796
you Apple could get a deal done with Live
Nation and
329
00:11:53,376 -->
00:11:54,376
Yes.
330
00:11:53,796 -->
00:11:55,688
just be like, here's the d well we're
gonna do this,
331
00:11:55,688 -->
00:11:57,459
we'll cut it. We'll cut you guys in.
332
00:11:57,459 -->
00:12:01,292
We'll just like you know, we'll just give
you a cut of every time someone uses this
333
00:12:01,292 -->
00:12:04,815
app. It's like I this is they're thinking
334
00:12:04,213 -->
00:12:05,213
Yeah, yeah, yeah.
335
00:12:04,815 -->
00:12:06,326
of it like T V. It's not T V.
336
00:12:06,815 -->
00:12:08,516
Right, exactly. We should we should be on
the team.
337
00:12:08,516 -->
00:12:11,207
They should do a brain trust and we'll c
we'll come out there.
338
00:12:11,207 -->
00:12:13,576
We'll come out and yeah, we'll talk about
it.
339
00:12:13,576 -->
00:12:17,229
All right. The Pixel Eleven event was last
night.
340
00:12:17,229 -->
00:12:22,021
Odd time for a live keynote, but also all
the YouTube videos came out earlier
341
00:12:22,021 -->
00:12:24,312
in the day, like Mr. Who's the Boss,
342
00:12:24,312 -->
00:12:28,754
MKBHD, a bunch of creators, Brian Tong all
posted videos about the Pixel Eleven
343
00:12:28,754 -->
00:12:33,336
and the entire lineup. And then there was
the live keynote at six PM Eastern
344
00:12:33,656 -->
00:12:36,297
on a Wednesday, hosted by Trevor Noah.
345
00:12:36,907 -->
00:12:40,438
Which was interesting. We'll get into
that.
346
00:12:40,438 -->
00:12:43,920
On the face of it, the hardware,
347
00:12:43,920 -->
00:12:45,641
I don't think, is the biggest story.
348
00:12:45,641 -->
00:12:47,801
You know, there's a new Tensor G6 chip.
349
00:12:47,801 -->
00:12:51,683
The camera hardware is a little better on
the Pixel 1111 Pro.
350
00:12:51,683 -->
00:12:54,024
They are more expensive, $100 more,
351
00:12:54,024 -->
00:12:56,005
but the base storage is now $256.
352
00:12:56,005 -->
00:13:00,907
Ramagedin. The 11 Pro now starts at 12
gigs of RAM instead of 16.
353
00:13:00,907 -->
00:13:03,728
So there's a bunch of like little hardware
updates and improvements.
354
00:13:03,728 -->
00:13:06,439
The fold is out there, slightly improved.
355
00:13:06,899 -->
00:13:11,980
Still thick, but there's a lot of software
things that I think
356
00:13:11,980 -->
00:13:16,112
are pretty interesting. Google's taking
some shots at Whisper Flow
357
00:13:16,552 -->
00:13:18,932
and doing some interesting software things
with the camera.
358
00:13:18,932 -->
00:13:22,093
There is one hardware thing that Joanna
Stern was very excited about,
359
00:13:22,093 -->
00:13:25,414
which is the light on the back of the
Pixel phones.
360
00:13:25,414 -->
00:13:28,065
There's now a BlackBerry
361
00:13:25,516 -->
00:13:26,516
Blackberry light.
362
00:13:28,065 -->
00:13:31,856
light where it's a notification light in
the camera bar.
363
00:13:31,856 -->
00:13:33,937
There's multiple colored LEDs,
364
00:13:33,937 -->
00:13:35,417
so this is not just like a solid
365
00:13:36,065 -->
00:13:39,627
Color flash. This is like actually
different colors that can meld
366
00:13:39,627 -->
00:13:45,421
and do animations kind of. But it can only
tell you when one of your favorites
367
00:13:45,421 -->
00:13:50,054
are calling and when Gemini is doing stuff
on the phone.
368
00:13:50,054 -->
00:13:53,656
Which if you've asked Gemini to do stuff
on the phone,
369
00:13:53,656 -->
00:13:55,017
you're probably looking at the screen,
370
00:13:55,017 -->
00:14:00,331
just you know, figure. And for when your
favorites call,
371
00:14:00,331 -->
00:14:03,723
that's interesting, but I think it would
probably be more beneficial
372
00:14:03,723 -->
00:14:05,624
if you could tell it when they text you.
373
00:14:06,066 -->
00:14:10,100
Or customize what apps like will light up
this thing.
374
00:14:10,100 -->
00:14:13,824
It's called the highlight L E D H I L I G
H T.
375
00:14:13,824 -->
00:14:17,758
I don't know. You can also make a shortcut
that just flashes the flashlight when
376
00:14:17,758 -->
00:14:21,614
you get a call. So that's I'm sorry,
377
00:14:19,678 -->
00:14:21,828
Well, actually isn't that an accessibility
feature?
378
00:14:21,614 -->
00:14:24,635
it's an it's an accessibility feature
because it's not a shortcut's automation
for
379
00:14:21,828 -->
00:14:23,356
Okay.
380
00:14:24,635 -->
00:14:27,577
calls, unfortunately, but it is a
accessibility feature.
381
00:14:27,577 -->
00:14:29,460
You could just have it flash your
flashlight,
382
00:14:29,460 -->
00:14:30,460
but
383
00:14:29,610 -->
00:14:30,610
Right,
384
00:14:29,890 -->
00:14:32,273
because there are a lot of people who
that's the only way they would know that
their
385
00:14:32,273 -->
00:14:34,755
phone is ringing. So that makes a lot of
sense.
386
00:14:33,266 -->
00:14:35,558
Yeah, a hundred percent. Yeah.
387
00:14:34,755 -->
00:14:35,755
So yeah, I
388
00:14:35,666 -->
00:14:37,177
d it is kind of a weird restriction.
389
00:14:37,177 -->
00:14:39,775
It seems like I mean, I don't want this.
390
00:14:39,775 -->
00:14:42,563
I don't want my phone to flash at me when
anything's happening.
391
00:14:42,563 -->
00:14:43,894
I turn most notifications off.
392
00:14:43,894 -->
00:14:46,737
I just don't just like I don't even want
to know people are trying to get
393
00:14:46,737 -->
00:14:49,990
a hold of me. But if I did, it would be
kinda cool just to be able to
394
00:14:49,990 -->
00:14:53,551
see that that is you know the the reason I
hated Blackberry is that
395
00:14:53,551 -->
00:14:54,551
the light just never went.
396
00:14:55,422 -->
00:14:57,811
I always had emails, I always had
notifications.
397
00:14:55,493 -->
00:14:57,172
Right, which is on all the time.
398
00:14:57,172 -->
00:14:58,886
I yeah.
399
00:14:57,811 -->
00:14:59,126
But I do think it should
400
00:14:59,126 -->
00:15:00,522
you should be able to decide what it's
for.
401
00:15:01,158 -->
00:15:04,670
Yeah, exactly. So and they s Google says
they're working on more customizations.
402
00:15:04,670 -->
00:15:07,572
But there's a couple of software things
that I want to focus on.
403
00:15:07,572 -->
00:15:10,114
This first one is called Rambler.
404
00:15:10,114 -->
00:15:13,036
And so this was a segment, this is Trevor
Noah,
405
00:15:13,036 -->
00:15:17,699
and he was using this feature and it's now
going to be built into the Pixel phones.
406
00:15:17,699 -->
00:15:21,962
It is dictation, but it's like dictation
supercharged.
407
00:15:21,962 -->
00:15:27,006
And I've actually been using WhisperFlow a
little more on my Mac,
408
00:15:27,006 -->
00:15:30,408
especially as I've been doing the whole
can do vibe coding.
409
00:15:30,751 -->
00:15:33,383
Because when I have a long thing I want to
tell an LLM to do,
410
00:15:33,383 -->
00:15:37,185
it's just easier to just talk it rather
than type it all,
411
00:15:37,185 -->
00:15:38,366
you know, when there's a lot of things.
412
00:15:38,366 -->
00:15:44,450
And Whisperflow is really good because you
can correct yourself mid-dictation.
413
00:15:44,450 -->
00:15:47,672
You can say, no, actually, I meant this.
414
00:15:47,672 -->
00:15:52,455
And it understands all of those little
quirks in your speech and provides
415
00:15:52,455 -->
00:15:56,678
a very clean message or transcription when
you're ready to send.
416
00:15:56,678 -->
00:16:00,260
Now, the annoyance with whisperflow on
iPhone is.
417
00:16:00,510 -->
00:16:03,581
Is that you have to switch to a
third-party keyboard in order for it
418
00:16:03,581 -->
00:16:08,212
to have that full functionality of pasting
what you just dictated and being able
419
00:16:08,212 -->
00:16:11,363
to turn it on. There's a weird action
button shortcut workaround,
420
00:16:11,363 -->
00:16:13,724
but it's just nowhere near as convenient
as on the Mac.
421
00:16:13,724 -->
00:16:18,305
And Pixel is saying we're gonna build this
just right into our phone.
422
00:16:18,305 -->
00:16:21,266
And so this Rambler feature is basically
that.
423
00:16:21,266 -->
00:16:25,827
It is the Whisperflow style transcription
that is very accurate,
424
00:16:25,827 -->
00:16:29,278
but also smart enough to know when you've
corrected yourself or you say,
425
00:16:29,278 -->
00:16:30,278
actually say this.
426
00:16:30,656 -->
00:16:33,447
Trevor No did a demo and it's it did
exactly what you would expect.
427
00:16:33,447 -->
00:16:35,937
Like correcting himself, changing what
he's saying.
428
00:16:35,937 -->
00:16:39,238
Bottom line, I want this on iPhone.
429
00:16:39,238 -->
00:16:42,529
And I know we're not going to get it
because Apple just improved
430
00:16:42,529 -->
00:16:45,540
the ejectation and all it did was just get
slightly more accurate.
431
00:16:45,540 -->
00:16:50,101
But it is nowhere near the Whisper Flow
self-correction type transcription.
432
00:16:50,101 -->
00:16:53,903
And this is a feature. I would not switch
to a Pixel for this,
433
00:16:53,903 -->
00:16:56,883
but man, I so want this. Or let Apple.
434
00:16:56,883 -->
00:16:59,044
I don't know if
435
00:16:59,156 -->
00:17:04,421
Figure out a way where it can do a third
party dictation service like Whisper
rather
436
00:17:04,421 -->
00:17:05,652
than have to use the built in one.
437
00:17:05,652 -->
00:17:07,364
So I want this feature. That's all.
438
00:17:07,936 -->
00:17:12,718
Yeah, I mean, we are getting to a point
where people just are talking it's weird.
439
00:17:12,718 -->
00:17:15,809
Like we're g we talk at our phones but no
one's there listening.
440
00:17:15,809 -->
00:17:17,310
Like we used to talk at our phones all the
time.
441
00:17:17,310 -->
00:17:18,310
That was the whole point
442
00:17:17,504 -->
00:17:18,504
Sure, sure.
443
00:17:18,220 -->
00:17:20,741
of having a phone, but now people are just
talking at it so you don't have
444
00:17:20,741 -->
00:17:22,572
to talk at someone. It's like just
445
00:17:21,842 -->
00:17:22,842
Yeah, exactly.
446
00:17:22,572 -->
00:17:25,283
take these words, turn into text so I can
send the text somewhere,
447
00:17:25,283 -->
00:17:28,175
whether that's to a human or a robot.
448
00:17:28,175 -->
00:17:32,096
That's fine. I I I don't find like on the
iPhone,
449
00:17:32,096 -->
00:17:36,268
I just use dictation and it's generally
fine for what I'm doing.
450
00:17:35,752 -->
00:17:36,752
It's fine.
451
00:17:36,309 -->
00:17:37,309
Yeah, it's fine.
452
00:17:36,993 -->
00:17:37,993
I
453
00:17:37,093 -->
00:17:40,786
don't know that it's great, but I do feel
like it's mostly fine.
454
00:17:40,786 -->
00:17:43,068
I used and on the Mac I used the
dictation.
455
00:17:43,068 -->
00:17:46,111
I think we talked about this when I had a
problem with like my wrist
456
00:17:46,111 -->
00:17:48,353
and it's like that's a problem 'cause I'm
a writer so I type all day
457
00:17:48,353 -->
00:17:53,077
and I just spent a week having to talk to
my computer and it was fine.
458
00:17:53,077 -->
00:17:56,700
What was it that made you think I need a
third party app to do this?
459
00:17:56,700 -->
00:17:59,042
Like you're doing it that much or?
460
00:17:59,600 -->
00:18:03,601
Well, David Sparks, I mean, just talked it
up so much.
461
00:18:03,601 -->
00:18:08,462
And when it comes to how I've been
prompting Claude for things on my app,
462
00:18:08,462 -->
00:18:12,684
there'll be times where I like I have a
bunch of bullets in a bear note
463
00:18:13,184 -->
00:18:16,705
of changes and bug fixes that people have
reported and they're
464
00:18:16,705 -->
00:18:19,745
not formatted enough where I can just copy
and paste that list into Claude
465
00:18:19,745 -->
00:18:23,406
and say go because there it's lacking some
clarity and context.
466
00:18:23,406 -->
00:18:27,007
And so what I will do is I'll look at that
bulleted list and think to myself,
467
00:18:27,007 -->
00:18:29,048
I could type this all out.
468
00:18:29,722 -->
00:18:33,593
And be more specific, or I could just hold
a key on my keyboard,
469
00:18:33,593 -->
00:18:38,535
talk through these bullet points in a
messy way that's not super precise.
470
00:18:38,535 -->
00:18:42,656
And Whisperflow not only dictates but
cleans it up for me in a way that
471
00:18:42,656 -->
00:18:44,856
I can just hit return and send it to
Claude.
472
00:18:44,856 -->
00:18:48,717
And so that use case has been nice.
473
00:18:48,717 -->
00:18:53,479
And we're gonna get to Claude actually
watermarking text that
474
00:18:53,479 -->
00:18:57,610
it generates and what I feel a little
weird about that.
475
00:18:58,320 -->
00:19:02,573
in because of how I use it. And so I might
start doing it even more because
476
00:19:02,573 -->
00:19:04,438
I want it to be my words and not Claude's.
477
00:19:04,438 -->
00:19:06,303
So for those reasons.
478
00:19:06,826 -->
00:19:09,380
Yeah. Do you think that this is a feature
that's gonna convince people
479
00:19:09,380 -->
00:19:11,394
to buy pixels? Okay.
480
00:19:10,438 -->
00:19:11,438
No, not at all. No one's gonna buy
481
00:19:11,398 -->
00:19:15,212
a pick. I mean, listen, Google stacks this
event with celebrities.
482
00:19:15,212 -->
00:19:16,613
I mean, Stephen Curry's in this,
483
00:19:16,613 -->
00:19:19,956
Trevor Noah's hosting it, Alex Cooper from
Caller Daddy.
484
00:19:19,956 -->
00:19:21,797
I'm gonna about to talk about her segment.
485
00:19:21,797 -->
00:19:26,670
And like, it's hilarious how much juice
Google tries to like put into this Pixel
486
00:19:26,670 -->
00:19:29,684
event, but everyone just still buys the
Samsung Galaxy S26.
487
00:19:29,684 -->
00:19:32,036
So it's like, I don't know. I mean,
488
00:19:32,036 -->
00:19:35,248
they're great, it's great hardware and
they're really interesting features.
489
00:19:35,248 -->
00:19:37,990
I'm gonna talk about all the camera stuff
and the creator suite in a minute,
490
00:19:37,990 -->
00:19:39,770
but yeah, no one's gonna buy a Pixel
because of that.
491
00:19:39,904 -->
00:19:40,904
Yes.
492
00:19:41,387 -->
00:19:42,387
Yeah.
493
00:19:41,707 -->
00:19:44,369
I don't know. I mean, it's in some ways
the same thing,
494
00:19:44,369 -->
00:19:47,831
like Microsoft makes Surface stuff and
they're good laptops.
495
00:19:47,831 -->
00:19:49,793
Google makes very good phones.
496
00:19:49,793 -->
00:19:51,574
It's just not a lot of people buy them,
497
00:19:51,574 -->
00:19:54,576
but they're sort of just putting these
things out there because what's really
498
00:19:54,576 -->
00:19:56,837
important to them is that Android exists.
499
00:19:57,778 -->
00:19:59,509
Right. And it's Google's like,
500
00:19:59,509 -->
00:20:03,291
hey, here's what an ideal Android phone
with Gemini and that's the other thing
501
00:20:03,291 -->
00:20:05,342
is they increase Gemini intelligence.
502
00:20:05,342 -->
00:20:07,413
It works with more apps, a bunch of apps
you never heard of,
503
00:20:07,413 -->
00:20:09,955
plus Zoc Doc and ~ Open Table.
504
00:20:09,955 -->
00:20:12,866
you know, so they're they build those
features in first so they can show
505
00:20:12,866 -->
00:20:14,557
you what's possible. But again,
506
00:20:14,557 -->
00:20:17,139
I don't think people are going to switch
to use it.
507
00:20:17,139 -->
00:20:21,531
But speaking of switching, there was this
entire segment of the Made
508
00:20:21,531 -->
00:20:24,913
by Google event where Alex Cooper goes on
stage and she talks
509
00:20:24,913 -->
00:20:26,804
to somebody from Google and they
510
00:20:27,210 -->
00:20:32,284
Talk about how easy it is to switch to an
Android phone from an iPhone
511
00:20:32,665 -->
00:20:35,437
and how Google worked with Apple earlier,
512
00:20:35,437 -->
00:20:38,930
like throughout this year. So when you
transfer your stuff,
513
00:20:38,930 -->
00:20:41,192
not only it brings your pictures,
514
00:20:41,192 -->
00:20:45,195
it also ~ get brings your passwords over
from Apple Passwords.
515
00:20:45,195 -->
00:20:47,697
And that switching is super easy.
516
00:20:47,697 -->
00:20:49,829
Alex Cooper even asked the guy,
517
00:20:49,829 -->
00:20:52,121
like, So are my group texts gonna be good?
518
00:20:52,121 -->
00:20:54,623
And he was like, Your group texts are
gonna be great.
519
00:20:54,623 -->
00:20:55,623
And it's like, well.
520
00:20:56,432 -->
00:20:59,273
Yes, RCS has gotten way better.
521
00:20:59,273 -->
00:21:03,255
Group techs are better. Still not the same
as iMessage,
522
00:21:03,255 -->
00:21:05,796
but fine. And I just wanted to mention
this segment.
523
00:21:05,796 -->
00:21:08,037
One, Alex Cooper has a very popular
podcast,
524
00:21:08,037 -->
00:21:12,279
Caller Daddy. And so let's check after the
iPhone 18 launch,
525
00:21:12,279 -->
00:21:15,600
a couple weeks after that, what phone is
Alex Cooper using.
526
00:21:15,600 -->
00:21:18,581
That's all I'm saying. I'm saying this is
all fine.
527
00:21:18,581 -->
00:21:22,517
Great. Switching is now easy. Let's see
what happens when the rubber meets the
road.
528
00:21:22,517 -->
00:21:25,374
Because one of the my favorite parts of
old Twitter.
529
00:21:25,940 -->
00:21:31,012
was when someone would make a deal with
Samsung or whoever and talk about
530
00:21:31,012 -->
00:21:37,324
how great the new phone Android phone is
and then like Oprah and then someone looks
531
00:21:37,324 -->
00:21:41,326
at a tweet that they post and it says sent
from iPhone and everybody would just call
532
00:21:41,326 -->
00:21:43,347
that out. Which you don't have that
anymore,
533
00:21:43,347 -->
00:21:46,288
the like the Twitter posted from this
thing.
534
00:21:46,288 -->
00:21:49,269
But when you have a big video podcast like
Alex Cooper,
535
00:21:49,269 -->
00:21:51,390
I'm sure she's gonna hold her phone at
some point.
536
00:21:51,390 -->
00:21:53,070
Let's just all see. I'm just curious.
537
00:21:53,070 -->
00:21:54,301
Let's see. Is it gonna be a pixel?
538
00:21:54,653 -->
00:21:57,005
Or is it gonna be the iPhone eighteen when
that comes out,
539
00:21:57,005 -->
00:21:59,090
~ later Sept late September? So we'll see.
540
00:21:59,090 -->
00:22:00,090
That's all.
541
00:21:59,778 -->
00:22:03,844
To be clear, I think it was actually
Samsung's account that posted something
that
542
00:22:03,844 -->
00:22:07,549
said sent from iPhone. But I also think
543
00:22:05,425 -->
00:22:07,354
Okay, very good. Very good.
544
00:22:07,549 -->
00:22:10,414
that it was like I've El at one point I'm
pretty sure Alicia Keys
545
00:22:10,414 -->
00:22:13,078
was like the creative director for Samsung
and then people like spotted
546
00:22:12,115 -->
00:22:13,115
Yes.
547
00:22:13,078 -->
00:22:14,740
her using an iPhone or something like
that.
548
00:22:14,740 -->
00:22:15,740
So yeah.
549
00:22:15,239 -->
00:22:16,239
It was that and
550
00:22:16,190 -->
00:22:19,516
I r I do think ~ Oprah was one because she
like talked about it
551
00:22:19,516 -->
00:22:21,729
and something and then she tweeted it
whatever was from her iPhone.
552
00:22:21,729 -->
00:22:25,014
There was a bunch of cases, like if you
just Google celebrities that yeah,
553
00:22:25,014 -->
00:22:26,014
anyway.
554
00:22:25,558 -->
00:22:28,881
To be fair, the switching thing is
actually really cool that the Apple
555
00:22:28,521 -->
00:22:30,330
Sure. Yeah.
556
00:22:28,881 -->
00:22:30,377
and Google work together. This
557
00:22:30,377 -->
00:22:31,722
was but it
558
00:22:32,074 -->
00:22:35,495
It's also very self serving. It's like d
we're not a we're a duopoly.
559
00:22:35,495 -->
00:22:37,616
We're not monopolies. Just be clear,
560
00:22:37,616 -->
00:22:38,777
we are not monopolies.
561
00:22:37,927 -->
00:22:38,927
Right.
562
00:22:38,777 -->
00:22:41,018
See, you can easily go back and forth
between the two.
563
00:22:41,018 -->
00:22:43,809
You can't go to anything else because
nothing else exists because we have
564
00:22:43,809 -->
00:22:46,200
the entire market, but we're not a
monopoly.
565
00:22:46,200 -->
00:22:49,331
But it is nice. I've used it obviously
some people like us,
566
00:22:49,331 -->
00:22:54,063
and by people like us, I just mean weirdos
who change phones all the time.
567
00:22:54,063 -->
00:22:55,254
I don't there's
568
00:22:54,345 -->
00:22:55,345
Sure, sure.
569
00:22:55,254 -->
00:22:59,465
nothing special about us. It's just often
I'm carrying multiple devices and
570
00:22:59,113 -->
00:23:00,113
Yeah.
571
00:22:59,666 -->
00:23:00,716
mo often that's
572
00:23:00,918 -->
00:23:03,500
couple times a year going back and forth
between an Android phone and
573
00:23:03,500 -->
00:23:06,382
a and it's very nice that I can just take
my main iPhone,
574
00:23:06,382 -->
00:23:12,606
put a selection of things on a Google f on
an Android device and use it with
575
00:23:12,606 -->
00:23:15,468
all of my information on it. That's super
helpful.
576
00:23:15,468 -->
00:23:19,411
I don't know how many people use this
because they're like,
577
00:23:19,411 -->
00:23:21,522
I've finally done it. I've given up the
iPhone.
578
00:23:21,522 -->
00:23:23,454
I'm going to Android or the other way
around.
579
00:23:24,500 -->
00:23:27,583
Yeah, I don't I mean you would use it once
and then probably
580
00:23:27,623 -->
00:23:29,555
not switch platforms again for a long
time,
581
00:23:29,555 -->
00:23:31,967
but yeah. I don't know. It's it's good
that it's there.
582
00:23:31,967 -->
00:23:36,551
I'm sure after like Apple did all this
work or whatever,
583
00:23:36,551 -->
00:23:40,195
Google also slides the invoice across the
table for Gemini powering Apple
584
00:23:40,195 -->
00:23:41,806
intelligence. It's like, yeah,
585
00:23:41,806 -->
00:23:43,117
by the way, don't forget to pay your bill.
586
00:23:43,117 -->
00:23:46,841
But you know, it's a it's a friendemyship.
587
00:23:46,841 -->
00:23:49,073
Frenemyship. Friendship
588
00:23:48,812 -->
00:23:49,812
They're
589
00:23:49,163 -->
00:23:51,197
they're even better of best of frenemies.
590
00:23:51,197 -->
00:23:52,860
Well that's open AI on Apple, but yeah.
591
00:23:53,245 -->
00:23:54,876
Well, I I think they're just straight
enemies now,
592
00:23:54,876 -->
00:23:56,758
Jason. I don't I don't know if there's any
friend of me.
593
00:23:56,704 -->
00:23:57,994
They need each other, Stephen.
594
00:23:56,758 -->
00:23:57,758
That
595
00:23:58,199 -->
00:23:59,470
that integration is going away,
596
00:23:59,470 -->
00:24:02,743
I'm sure. The open AI one. But so that's
that.
597
00:24:02,743 -->
00:24:04,644
A couple of things on the cameras.
598
00:24:04,644 -->
00:24:07,427
So the cameras have improved hardware in
the pixel devices,
599
00:24:07,427 -->
00:24:10,149
but some really cool software features.
600
00:24:10,149 -->
00:24:13,491
And they're really gearing towards
creators with a couple of them.
601
00:24:13,491 -->
00:24:19,797
One is called Magic Capture. And this is a
feature where in the demo
602
00:24:19,917 -->
00:24:22,577
it shows that you can just hold the Pixel
phone and
603
00:24:23,091 -->
00:24:25,603
And it's basically it looks like it's
capturing a video and has
604
00:24:25,603 -->
00:24:26,794
a couple basketball players doing it.
605
00:24:26,794 -->
00:24:32,199
And the Pixel is just going to choose
moments that are that would make great
photos
606
00:24:32,519 -->
00:24:38,844
and capture those photos. And this is not
the same as pulling still frames from
607
00:24:38,844 -->
00:24:42,257
a video. And in the demo, they zoom into
the photos to show like this
608
00:24:42,257 -->
00:24:45,369
is a high-quality photo. This is not just
a screen grab from a video.
609
00:24:45,369 -->
00:24:49,973
And you don't even like the person demoing
it wasn't even looking at the phone.
610
00:24:49,973 -->
00:24:51,735
Like you're just kind of holding it at
chest height.
611
00:24:52,137 -->
00:24:55,037
Kind of like how your parents or
grandparents hold something when they're
trying
612
00:24:55,037 -->
00:24:57,338
to like capture something in a in the
school,
613
00:24:57,338 -->
00:25:00,039
whatever. They're just kind of like
holding the phone randomly and pointing
614
00:25:00,039 -->
00:25:04,801
it in random directions. But the magic
capture will then just pull great photos
from
615
00:25:04,801 -->
00:25:07,891
moments it decides is really good and you
can decide which ones to keep.
616
00:25:07,891 -->
00:25:10,322
That's a cool feature built in.
617
00:25:10,322 -->
00:25:12,910
Again, it's software, not necessarily,
618
00:25:12,910 -->
00:25:14,733
I don't know if you need the latest
hardware to do it,
619
00:25:14,733 -->
00:25:19,135
but that's a cool feature. And then the
creator focused features,
620
00:25:19,135 -->
00:25:20,455
these are things that
621
00:25:20,618 -->
00:25:22,569
You know, I would love see on an iPhone
too.
622
00:25:22,569 -->
00:25:26,370
Things like you can have a teleprompter
right there while you're recording
623
00:25:26,370 -->
00:25:32,923
in the camera. It has social media safe
lines so you can see what how like frame
624
00:25:32,923 -->
00:25:35,474
it up for vertical video and horizontal.
625
00:25:35,474 -->
00:25:39,865
You can see like what is going when you
crop it to that vertical nine by sixteen,
626
00:25:39,865 -->
00:25:43,637
what it's gonna look like. These are nice
little features that are just added in.
627
00:25:43,637 -->
00:25:48,169
You know, I feel like many times Apple has
started down this path,
628
00:25:48,169 -->
00:25:49,309
like with the Clips app.
629
00:25:49,949 -->
00:25:53,350
And was like, we're gonna make an an app
where you can edit short form video,
630
00:25:53,350 -->
00:25:56,222
but it's gonna be square, and then we're
gonna forget about this app forever.
631
00:25:56,222 -->
00:26:01,594
And it's like if they had if Apple had
just pushed on that for a couple years
632
00:26:01,594 -->
00:26:05,726
and innovated, like that could have been a
useful creator tool.
633
00:26:05,726 -->
00:26:08,066
And Apple is doing some of that,
634
00:26:08,066 -->
00:26:11,828
like into Final Cut, they added the
ability to do captions and overlay it,
635
00:26:11,828 -->
00:26:13,309
but there's not a lot of style.
636
00:26:13,309 -->
00:26:18,161
So just credit to Google actually building
features that I think would help
637
00:26:18,161 -->
00:26:19,161
creators.
638
00:26:19,184 -->
00:26:20,659
If any of them use pixel phones,
639
00:26:20,659 -->
00:26:21,659
but I don't know if they do.
640
00:26:22,188 -->
00:26:27,560
Two things. One, I think like I always
feel very weird,
641
00:26:27,560 -->
00:26:30,020
especially when Google releases camera
features,
642
00:26:30,020 -->
00:26:33,621
because I'm like I don't want to get into
the like what's a photo thing.
643
00:26:33,621 -->
00:26:35,802
But it's like y okay.
644
00:26:34,451 -->
00:26:35,961
We'll we'll touch on that. We'll touch on
that.
645
00:26:36,602 -->
00:26:40,703
But I also think the reality is they just
know how most normal people
646
00:26:41,524 -->
00:26:43,154
how many terrible photos they take.
647
00:26:43,154 -->
00:26:45,995
And st I don't know, have you ever I'm
sure you've had this experience,
648
00:26:45,995 -->
00:26:49,056
but like people will send you photos and
you're like how
649
00:26:49,416 -->
00:26:52,247
How is it possible you took such a
terrible photo of this moment?
650
00:26:52,247 -->
00:26:54,538
Like, do you n do you not
651
00:26:52,568 -->
00:26:53,895
W were you running away?
652
00:26:54,538 -->
00:26:56,338
know how to point your phone correctly?
653
00:26:56,338 -->
00:27:00,030
Like it's like do you not understand
things like maybe I mean,
654
00:27:00,030 -->
00:27:02,020
I understand that the teenagers like
taking pi one,
655
00:27:02,020 -->
00:27:04,661
you point five pictures of their foreheads
and so they do that on purpose.
656
00:27:04,125 -->
00:27:05,354
Yes, yes.
657
00:27:04,661 -->
00:27:05,661
But like, I
658
00:27:05,430 -->
00:27:07,442
don't think that our parents are trying to
do that,
659
00:27:07,442 -->
00:27:09,813
right? And actually I'm not picking on my
parents because my mom
660
00:27:09,873 -->
00:27:11,366
was a photographer and does a great job.
661
00:27:11,366 -->
00:27:12,534
The point is, like
662
00:27:13,430 -->
00:27:14,881
People will look at a photo and be like,
663
00:27:14,881 -->
00:27:16,452
your photo for is so much better.
664
00:27:16,452 -->
00:27:20,495
I'm like, I just pointed the phone at it
and pushed the button.
665
00:27:20,495 -->
00:27:22,276
Like I didn't do it, like I don't I don't
know.
666
00:27:22,143 -->
00:27:26,317
The w the one habit I see is so many that
yeah.
667
00:27:22,276 -->
00:27:25,858
Stop stop pinging to Zoom, stop all that
stuff,
668
00:27:25,858 -->
00:27:27,910
just tap the button. And I think but
669
00:27:26,317 -->
00:27:27,793
Yeah.
670
00:27:27,910 -->
00:27:29,321
Google knows, like, no no no no.
671
00:27:29,321 -->
00:27:30,962
People are terrible at taking photos.
672
00:27:30,962 -->
00:27:32,203
So we're gonna do all this work because
673
00:27:31,839 -->
00:27:32,839
Gorgeous.
674
00:27:32,203 -->
00:27:36,005
what they really want is just to have a
good end result and we can argue about
675
00:27:36,005 -->
00:27:38,287
all of those details and and that's fine.
676
00:27:38,287 -->
00:27:40,739
The other thing I was just gonna say is
like the creator,
677
00:27:40,739 -->
00:27:42,080
like all these tools and stuff,
678
00:27:42,080 -->
00:27:43,080
that's great.
679
00:27:43,713 -->
00:27:45,804
I you I'm sure have this experience.
680
00:27:45,804 -->
00:27:50,005
But like right now, if you took a vertical
video on your phone and you're like,
681
00:27:50,005 -->
00:27:52,350
this would be cool. I want to add
captions.
682
00:27:52,350 -->
00:27:56,433
I want to do this. There isn't actually
like AI should have solved this problem
683
00:27:56,433 -->
00:27:58,635
by now. I mean cap cut, whatever.
684
00:27:58,635 -->
00:27:59,635
Like it's it's
685
00:27:58,834 -->
00:27:59,834
Yeah, yeah.
686
00:27:59,515 -->
00:28:04,890
but it's still like dialing that stuff in
is not super easy anywhere.
687
00:28:04,890 -->
00:28:09,663
And so I actually think the more that pe
the Google and Apple could build these
688
00:28:09,663 -->
00:28:12,445
features in in a way that's really
accessible,
689
00:28:12,445 -->
00:28:13,445
because
690
00:28:13,216 -->
00:28:14,776
That's what people want to do with this
stuff.
691
00:28:14,776 -->
00:28:15,776
They're taking
692
00:28:14,974 -->
00:28:15,974
Yeah.
693
00:28:15,447 -->
00:28:19,489
these short videos so that they can post
them on Instagram or TikTok or whatever.
694
00:28:19,489 -->
00:28:21,510
So I think that's actually a positive,
695
00:28:21,510 -->
00:28:25,323
even though I'm still in the camp of like
people should just be able
696
00:28:25,323 -->
00:28:27,814
to take better photos. The cameras should
stop trying to do so much.
697
00:28:28,786 -->
00:28:32,227
Well, sure. I think you know, the pinching
and zooming is one thing.
698
00:28:32,227 -->
00:28:38,179
I'm also usually I cringe when someone
like moves
699
00:28:38,179 -->
00:28:41,010
the phone so quickly after hitting the
shutter button.
700
00:28:41,010 -->
00:28:44,626
Like so often I see someone like hit the
shutter button and the camera's down.
701
00:28:44,626 -->
00:28:49,093
And like for me, I always like give it a
second just to make sure it's like
captured
702
00:28:49,093 -->
00:28:50,293
and still, you know what I mean?
703
00:28:50,293 -->
00:28:52,094
I don't know if you you see that kind of
stuff,
704
00:28:52,094 -->
00:28:53,094
but
705
00:28:52,214 -->
00:28:53,214
Yeah,
706
00:28:52,404 -->
00:28:56,476
also, honestly, you should just turn off
all the camera control features
707
00:28:56,476 -->
00:28:57,896
and you should use it as a shutter button.
708
00:28:57,896 -->
00:28:59,377
It's the best way to use your camera.
709
00:28:59,377 -->
00:29:04,069
A hundred percent. Just hold it there and
just hit the camera control button
710
00:29:04,069 -->
00:29:05,339
constantly. And that's
711
00:29:04,242 -->
00:29:05,242
Yeah. Yeah.
712
00:29:05,339 -->
00:29:06,840
the people don't do that though.
713
00:29:06,840 -->
00:29:08,781
They're tapping the thing and yeah.
714
00:29:08,168 -->
00:29:09,771
This should I still tap the shutter
button,
715
00:29:08,781 -->
00:29:10,101
No, no, no.
716
00:29:09,771 -->
00:29:10,771
but
717
00:29:10,101 -->
00:29:11,382
Don't do the don't tap the shutter button.
718
00:29:11,504 -->
00:29:15,731
see, I like to do it because pressing the
camera control button I feel like moves
719
00:29:15,731 -->
00:29:18,336
the phone a little too much, even if I try
really hard to hold it still.
720
00:29:18,336 -->
00:29:21,742
Whereas tapping the shutter button I could
do so lightly it doesn't move the phone.
721
00:29:22,806 -->
00:29:24,664
We had this entire conversation once.
722
00:29:24,167 -->
00:29:25,337
I know we had this entire conversation.
723
00:29:24,664 -->
00:29:25,678
Okay. Okay.
724
00:29:25,337 -->
00:29:27,568
I'm just reiterating my habit and
725
00:29:27,748 -->
00:29:28,949
we won't get we won't have to get into it.
726
00:29:28,949 -->
00:29:31,530
But we will say one other thing about the
camera stuff:
727
00:29:31,530 -->
00:29:37,232
Pixel is adding these cam Pixel camera
looks is the feature.
728
00:29:37,232 -->
00:29:41,294
And in the Google Pixels are giving you
more control.
729
00:29:41,294 -->
00:29:44,295
It's not just filters on top of the photo.
730
00:29:44,295 -->
00:29:48,616
This gives like photo gen like photo
styles on iPhone,
731
00:29:48,616 -->
00:29:52,258
where when you decide the look or how you
want the photo
732
00:29:53,448 -->
00:29:59,450
composed, it's actually affecting the
entire photo process on the camera,
733
00:29:59,450 -->
00:30:02,420
like from the sensor, the processing of
HDR,
734
00:30:02,420 -->
00:30:05,041
all of that. So it's not just a filter on
top of a photo.
735
00:30:05,041 -->
00:30:07,302
It will affect how the photo is taken.
736
00:30:07,302 -->
00:30:11,423
And they are giving users more granular
control over these looks.
737
00:30:11,423 -->
00:30:16,314
And for instance, if you look at this
example and it's in the Verge article,
738
00:30:16,314 -->
00:30:17,545
a link in the show notes as well,
739
00:30:17,545 -->
00:30:19,401
you can choose like, you know what,
740
00:30:19,401 -->
00:30:21,446
don't crush the highlights so much.
741
00:30:21,446 -->
00:30:22,696
I kinda want
742
00:30:22,910 -->
00:30:26,151
kind of the more natural look where this
is what it looked like when I
743
00:30:26,151 -->
00:30:30,312
was there in real life. Or you just don't
want the HDR crank so much.
744
00:30:30,312 -->
00:30:33,613
And if you want a ~ the shadows to
actually be shadowy,
745
00:30:33,613 -->
00:30:35,753
or if you want the grain because it's low
light,
746
00:30:35,753 -->
00:30:40,735
it's giving users the option to have less
processing or at least process
747
00:30:40,735 -->
00:30:43,956
it differently. So it doesn't look so much
like a smartphone photo.
748
00:30:43,956 -->
00:30:49,137
And like I've seen a lot on social media
and TikTok people using like
749
00:30:49,137 -->
00:30:52,168
old cameras because it has cool looks
750
00:30:52,530 -->
00:30:56,062
But basically the cool look is just minus
all the computational photography.
751
00:30:56,062 -->
00:30:57,062
Like the
752
00:30:56,640 -->
00:30:57,640
Right.
753
00:30:56,973 -->
00:30:59,684
the cool look is this is an old camera
that all it did was capture what
754
00:30:59,684 -->
00:31:02,936
was in front of it and there were no
settings and no processing.
755
00:31:02,936 -->
00:31:05,287
And especially younger generations are
like,
756
00:31:05,287 -->
00:31:08,459
hey, this is cool. Like I actually want
pictures that look like this.
757
00:31:08,459 -->
00:31:11,761
And so the Pixel is actually doing that
and giving users more control.
758
00:31:11,761 -->
00:31:14,562
Again, photogenic styles gives you some
control on the iPhone,
759
00:31:14,562 -->
00:31:16,523
but the Pixel is going even further,
760
00:31:16,523 -->
00:31:20,375
giving more granular control and being
able to make it look like not
761
00:31:20,375 -->
00:31:21,636
as processed of a photo.
762
00:31:21,810 -->
00:31:22,964
Which I think is good. I mean,
763
00:31:22,964 -->
00:31:24,912
yeah. Nice.
764
00:31:24,672 -->
00:31:25,672
I agree.
765
00:31:25,052 -->
00:31:28,028
I agree. Yeah. And I do think that the
headline of that story is great.
766
00:31:28,028 -->
00:31:30,673
Google's making pixel cameras better by
making them worse.
767
00:31:30,789 -->
00:31:32,280
Exact exactly.
768
00:31:31,555 -->
00:31:32,898
Good job. Good job, David.
769
00:31:33,160 -->
00:31:35,842
Exactly. All right. A couple of last quick
things.
770
00:31:35,842 -->
00:31:37,983
~ I just closed a tab. no.
771
00:31:38,988 -->
00:31:39,988
Show's over.
772
00:31:39,944 -->
00:31:43,045
The show is over. I inadvertently closed a
tab.
773
00:31:43,045 -->
00:31:48,218
No, the Pixel Tag. So Google did announce
a and yeah,
774
00:31:48,218 -->
00:31:49,769
announce because it doesn't come out till
November,
775
00:31:49,769 -->
00:31:52,230
an air tag style tracking tag.
776
00:31:52,230 -->
00:31:55,052
It's called the Pixel Tag. It's gonna be
$29 for one.
777
00:31:55,052 -->
00:31:59,094
It's gonna be $99 for four. It is oblong.
778
00:31:59,376 -->
00:32:01,567
It's not like a perfect circle like an air
tag.
779
00:32:01,567 -->
00:32:05,988
It is a kind of a oval shape. It is still
a coin battery,
780
00:32:05,988 -->
00:32:09,733
year of battery life. Google says it will
be able to ping
781
00:32:09,753 -->
00:32:14,606
the entire Android network. So it seems
like maybe any Android phone doesn't have
782
00:32:14,606 -->
00:32:16,917
to be a Pixel, will be a part of the
network,
783
00:32:16,917 -->
00:32:18,918
the find hub network for Google.
784
00:32:18,918 -->
00:32:21,420
So yeah, if you are in the Pixel universe,
785
00:32:21,420 -->
00:32:23,081
you can have an AirTag like thing.
786
00:32:23,081 -->
00:32:24,322
There you go.
787
00:32:24,192 -->
00:32:25,192
I
788
00:32:24,622 -->
00:32:29,885
I do think this is a perfect example of
how good Apple
789
00:32:29,885 -->
00:32:34,307
is at designing the form of a product
around its limiting factor,
790
00:32:34,307 -->
00:32:36,948
which is the little CR whatever,
791
00:32:36,948 -->
00:32:38,129
twenty thirty two battery.
792
00:32:37,406 -->
00:32:38,542
Twenty thirty two battery. Yeah,
793
00:32:38,542 -->
00:32:39,542
yeah, yeah.
794
00:32:38,909 -->
00:32:40,870
Because this looks like
795
00:32:41,366 -->
00:32:44,798
This actually looks a lot like the one of
the versions of the Google
796
00:32:44,798 -->
00:32:48,731
TV Chromecast remote. Looked a lot like
that.
797
00:32:47,376 -->
00:32:48,507
Yeah, it does, yeah.
798
00:32:48,766 -->
00:32:52,694
It's like this is this is weird in the
sense that like the air tag
799
00:32:52,694 -->
00:32:54,205
is the perfect form factor for this.
800
00:32:54,205 -->
00:32:56,417
And people are like, well, you can make it
smaller if you made it rechargeable.
801
00:32:56,417 -->
00:32:58,112
No, no, no. It's perfect that you can just
twist it,
802
00:32:58,112 -->
00:32:59,139
take out a battery every year,
803
00:32:59,139 -->
00:33:00,339
put it in. I don't have to recharge it.
804
00:33:00,339 -->
00:33:02,641
I don't have to you the battery won't
degrade over time.
805
00:33:02,641 -->
00:33:04,314
It'll just die and I put a new one in.
806
00:33:04,314 -->
00:33:05,314
But I I just
807
00:33:04,594 -->
00:33:05,594
Yeah, yeah.
808
00:33:05,073 -->
00:33:09,386
think not throwing any shade at this thing
other than I have no use for it.
809
00:33:10,206 -->
00:33:14,163
I just think it should we should
appreciate for all of our criticisms of
Apple,
810
00:33:14,163 -->
00:33:16,547
the air tag is actually such a perfect
example of like,
811
00:33:16,547 -->
00:33:19,242
no, the limiting factor of this is the
size and shape of the battery
812
00:33:19,623 -->
00:33:21,556
and they built the perfect device around
that.
813
00:33:21,520 -->
00:33:22,520
Yeah.
814
00:33:21,860 -->
00:33:24,923
And I'm curious like the range of
something like this Pixel One.
815
00:33:24,923 -->
00:33:28,035
I'm sure Lexi Civides, who was on the show
a few weeks ago,
816
00:33:28,035 -->
00:33:32,679
she actually did a whole find my
comparison with the current products.
817
00:33:32,679 -->
00:33:35,981
And so I hope she does a an updated video
with this specifically.
818
00:33:35,981 -->
00:33:37,062
It doesn't come out till November,
819
00:33:37,062 -->
00:33:38,663
but I will be curious the range,
820
00:33:38,663 -->
00:33:41,465
because AirTag two range is insane.
821
00:33:41,465 -->
00:33:45,168
I remember I did a short video and it got
over two hundred feet,
822
00:33:45,168 -->
00:33:48,961
I think, and I still had to precision find
my arrow on screen,
823
00:33:48,961 -->
00:33:50,991
which was pretty wild. So yeah.
824
00:33:51,601 -->
00:33:53,272
That's pretty cool. All right.
825
00:33:53,272 -->
00:33:54,272
Well, that was the Pixel event.
826
00:33:54,263 -->
00:33:58,887
There was also the Pixel Watch 5 that has
breathing emergency detection
827
00:33:59,368 -->
00:34:02,750
and also insulin resistance trends.
828
00:34:02,750 -->
00:34:04,992
Not so much like individual measurements,
829
00:34:04,992 -->
00:34:06,674
but that's there in blood pressure trends.
830
00:34:06,674 -->
00:34:09,408
So you got Pixel Watch 5, which I do I
think it looks nice.
831
00:34:09,408 -->
00:34:10,587
I mean it's a nice circle face.
832
00:34:10,587 -->
00:34:13,239
That's cool. Also, tap to share is going
to be updated,
833
00:34:13,239 -->
00:34:16,952
so you'll be able to just tap two Android
phones together as long
834
00:34:16,952 -->
00:34:19,675
as they're a Galaxy 8, the Z Fold 8,
835
00:34:19,675 -->
00:34:20,675
the Fold 8.
836
00:34:20,593 -->
00:34:22,385
And a Pixel Six series of newer,
837
00:34:22,385 -->
00:34:23,957
which actually going back pretty far.
838
00:34:23,957 -->
00:34:26,610
Shout out Bump. You remember the Bump app,
839
00:34:26,610 -->
00:34:27,741
Jason? Yeah.
840
00:34:26,784 -->
00:34:27,784
Yep, I do.
841
00:34:27,901 -->
00:34:30,865
That was on iPhone too. Like the Bump app
existed on both platforms,
842
00:34:30,865 -->
00:34:31,865
I feel like.
843
00:34:31,040 -->
00:34:32,040
And
844
00:34:31,170 -->
00:34:34,403
it was a big deal before you could just
have those contact posters that
845
00:34:34,403 -->
00:34:36,315
you could share by putting the phone next
to each other.
846
00:34:36,315 -->
00:34:38,748
But in this I mean, bump was like what it
was like,
847
00:34:38,748 -->
00:34:40,490
two thousand ten, eleven maybe.
848
00:34:40,490 -->
00:34:43,110
I can just remember going to conferences
and it's like,
849
00:34:43,110 -->
00:34:45,765
Here you go, wanna bump? Which is a weird
thing to say to someone,
850
00:34:44,391 -->
00:34:45,391
Bump. What a
851
00:34:45,765 -->
00:34:46,765
but
852
00:34:46,102 -->
00:34:48,072
bump. Maybe the name. If it was a better
name,
853
00:34:48,072 -->
00:34:50,322
it might have been better. But anyway.
854
00:34:50,322 -->
00:34:53,183
Alright, Zuck had this huge AI manifesto I
want to get into.
855
00:34:53,183 -->
00:34:55,584
Claude is gonna watermark text that it
generates.
856
00:34:55,584 -->
00:34:56,904
I have some interesting feelings about
that.
857
00:34:56,904 -->
00:34:59,985
And yeah, Spotify is ~ maybe trying to
kill podcasting.
858
00:34:59,985 -->
00:35:01,653
We'll get into it. But before we get to
all that,
859
00:35:01,653 -->
00:35:04,006
we're gonna take a break and thank today's
sponsor,
860
00:35:04,006 -->
00:35:07,087
which is Decagon. Listen, Decagon,
861
00:35:07,087 -->
00:35:10,118
if you have a team and you're dealing with
tons of support tickets,
862
00:35:10,118 -->
00:35:13,089
multiple channels, your customers have
high expectations.
863
00:35:13,393 -->
00:35:16,534
You might have thought about using AI for
customer support and you might
864
00:35:16,534 -->
00:35:20,195
be limited in time and how time consuming
those options are and training it.
865
00:35:20,195 -->
00:35:22,336
Well, that's why Decagon is here for you.
866
00:35:22,336 -->
00:35:27,297
Dekagon helps companies create
personalized concierge-style customer
experiences
867
00:35:27,417 -->
00:35:29,798
with AI agents across chat, email,
868
00:35:29,798 -->
00:35:32,999
voice, and SMS. They're available 24-7,
869
00:35:32,999 -->
00:35:36,890
they feel natural to talk to, and can
resolve customer requests on their
870
00:35:36,890 -->
00:35:41,381
own so businesses can keep up with
requests without losing that personal
touch.
871
00:35:41,798 -->
00:35:46,462
And Dekagon plugs directly into your
existing teams so agents can get live
quickly.
872
00:35:46,462 -->
00:35:49,644
And because workflows can be updated using
natural language,
873
00:35:49,644 -->
00:35:53,788
teams can make changes themselves without
long engineering cycles.
874
00:35:53,788 -->
00:35:57,301
That's huge. You don't need to be a
programmer to affect how it responds
875
00:35:57,301 -->
00:36:00,463
and do you update it. You can just your
team can do it today.
876
00:36:00,463 -->
00:36:03,816
That means less time spent managing
workflows and more time focused
877
00:36:03,816 -->
00:36:05,838
on customers and ROI in weeks.
878
00:36:05,838 -->
00:36:08,520
And once live agents can pull up to date
information,
879
00:36:08,520 -->
00:36:09,671
take action on requests.
880
00:36:09,851 -->
00:36:12,342
And understand the context behind any
conversation.
881
00:36:12,342 -->
00:36:16,765
Also, Decagon gives your team full
visibility into why agents make decisions
882
00:36:16,885 -->
00:36:19,406
and what's happening across every
conversation.
883
00:36:19,406 -->
00:36:23,098
Decagon helps power millions of
conversations every day for brands
884
00:36:23,098 -->
00:36:24,409
you already know and love like Avis,
885
00:36:24,409 -->
00:36:26,610
Affirm, Fanatics, and Aura, like the Aura
Ring.
886
00:36:26,610 -->
00:36:28,901
So if you're ready to transform your
customer support,
887
00:36:28,901 -->
00:36:34,835
go to Decagon.ai slash primary to get a
personalized demo and see what Decagon
888
00:36:34,835 -->
00:36:37,976
can do for your team. So check out Decagon
at Decagon.ai.
889
00:36:37,976 -->
00:36:38,977
~
890
00:36:39,087 -->
00:36:41,178
slash primary. That link's also in the
show notes.
891
00:36:41,178 -->
00:36:44,000
That's decagon.ai slash primary.
892
00:36:44,000 -->
00:36:46,961
Thanks to Decagon for sponsoring this
episode.
893
00:36:46,961 -->
00:36:49,943
And also thanks to all of our members who
support the show,
894
00:36:49,943 -->
00:36:52,504
which just a quick plug, you can become a
member.
895
00:36:52,504 -->
00:36:54,926
There's a link down in the show notes if
you're not already a member.
896
00:36:54,926 -->
00:36:59,328
$2.50 a month or $25 a year. You get an
ad-free version of the show.
897
00:36:59,328 -->
00:37:01,990
You get the bonus episode, you get the
unedited feed,
898
00:37:01,990 -->
00:37:03,491
which also includes a pre-show,
899
00:37:03,491 -->
00:37:05,151
you get primary tech daily, you get all
the things.
900
00:37:05,151 -->
00:37:06,412
So we'd appreciate your support.
901
00:37:06,412 -->
00:37:07,853
Thank you for all those who already do
that.
902
00:37:08,076 -->
00:37:09,076
You
903
00:37:08,217 -->
00:37:10,343
get a sneak peek into something I'm
working on.
904
00:37:10,801 -->
00:37:13,283
That is true. We talk about Jason's next
app,
905
00:37:13,283 -->
00:37:15,444
which is pretty exciting. So, you know,
906
00:37:15,444 -->
00:37:17,586
it's right there. All right, Jason,
907
00:37:17,586 -->
00:37:23,110
I read Mark Zuckerberg's entire AI
manifesto.
908
00:37:23,110 -->
00:37:27,292
It is very long. It's like over 6,500
words.
909
00:37:27,292 -->
00:37:29,534
I have just a quick question for you.
910
00:37:29,534 -->
00:37:32,036
How it's okay if you did. Did you read any
of this?
911
00:37:32,036 -->
00:37:34,377
Okay, you read the whole thing.
912
00:37:32,448 -->
00:37:35,245
I read the whole thing. I even made notes,
913
00:37:35,245 -->
00:37:36,849
Stephen, incontextually.
914
00:37:38,360 -->
00:37:39,360
I made notes too.
915
00:37:39,654 -->
00:37:43,560
I made notes. One, the title is The Future
Is For Everyone.
916
00:37:43,560 -->
00:37:47,787
Subhead The Path to a Positive AI Future.
917
00:37:47,787 -->
00:37:51,103
I don't like that there's the in the title
and the in the subtitle.
918
00:37:51,103 -->
00:37:52,535
I think you should drop one of the the's.
919
00:37:52,535 -->
00:37:53,535
You know what I mean?
920
00:37:55,010 -->
00:37:58,473
Sure, but the Yeah.
921
00:37:55,399 -->
00:37:57,293
That's a that's a joke. That's a social
network joke.
922
00:37:57,293 -->
00:37:58,545
Just throwing
923
00:37:58,545 -->
00:38:01,071
that out there. Do you think you're a
writer,
924
00:38:01,071 -->
00:38:03,445
Jason? I think our listeners know that.
925
00:38:03,445 -->
00:38:05,179
Is Zuck a good writer?
926
00:38:08,551 -->
00:38:11,372
I don't know that it would be fair to say
that Zuck wrote this.
927
00:38:11,372 -->
00:38:14,774
I mean, I don't this is like how many
words was this thing?
928
00:38:13,479 -->
00:38:16,505
Sure. Sixty five hundred? About sixty five
hundred.
929
00:38:14,774 -->
00:38:15,774
A lot.
930
00:38:16,735 -->
00:38:21,767
The guy running the largest advertising
platform in the world doesn't have time
931
00:38:21,767 -->
00:38:23,048
to write sixty five hundred words,
932
00:38:23,048 -->
00:38:24,048
does he?
933
00:38:24,847 -->
00:38:29,158
Well, that's a good question. So high
level,
934
00:38:29,158 -->
00:38:34,860
this is like Zuck trying to I guess make a
call like who do you who is this for?
935
00:38:34,860 -->
00:38:38,831
It seems like maybe this is a call to the
governments and also other
936
00:38:38,831 -->
00:38:44,553
AI companies and also trying to tell the
people that he is he thinks AI
937
00:38:44,553 -->
00:38:47,564
is for the people and needs to be
available to all people,
938
00:38:47,564 -->
00:38:50,745
free and open, and the only way we all
succeed is everybody gets AI.
939
00:38:51,141 -->
00:38:53,334
on their own terms and I'm gonna talk
about the personal agent
940
00:38:53,334 -->
00:38:56,508
he talks about several times. Who who is
this for,
941
00:38:56,508 -->
00:38:57,508
do you think?
942
00:38:57,108 -->
00:38:58,850
Okay, this is helpful context.
943
00:38:58,850 -->
00:39:03,044
Do you wanna start by c going through it
and then we can we can talk about
944
00:39:03,084 -->
00:39:06,808
why this exists? Okay. Okay.
945
00:39:04,935 -->
00:39:06,787
Sure. Let's do that. Let's do that.
946
00:39:06,968 -->
00:39:09,931
'Cause I think that the why it exists is
the only part that matters,
947
00:39:09,931 -->
00:39:12,134
but if we start talking about that we'll
never get to the other part,
948
00:39:12,134 -->
00:39:13,134
so
949
00:39:12,305 -->
00:39:13,305
Okay,
950
00:39:12,695 -->
00:39:14,325
all right. So I'll I'll quickly jam
through this.
951
00:39:14,325 -->
00:39:15,917
I made a bunch of notes, I'm not gonna go
through all of it,
952
00:39:15,917 -->
00:39:19,318
but he does make a call to Steve Jobs
talking about like all
953
00:39:19,318 -->
00:39:20,749
the revolutions in technology,
954
00:39:20,749 -->
00:39:25,041
and he mentions in a line the kid in a
garage who thought personal computer
should
955
00:39:25,041 -->
00:39:27,332
be for everyone. And that I don't know.
956
00:39:27,332 -->
00:39:28,662
I was a little bit like, All right,
957
00:39:28,662 -->
00:39:31,624
all right. You d don't don't invoke the w
he never said Steve Jobs,
958
00:39:31,624 -->
00:39:32,814
but everybody knows what he's talking
about.
959
00:39:32,505 -->
00:39:33,961
but he's always wanted to be Steve Jobs.
960
00:39:34,075 -->
00:39:39,067
Always want to be Steve Jobs. So a couple
poll quotes from this manifesto.
961
00:39:40,241 -->
00:39:43,442
First of all, he's trying to make
superintelligence the term.
962
00:39:43,442 -->
00:39:46,704
And I think he and Meta are the only ones
using that term,
963
00:39:46,704 -->
00:39:49,465
superintelligence. He has the
superintelligence labs,
964
00:39:49,465 -->
00:39:51,686
but fine. He's really trying to make that
a thing.
965
00:39:51,686 -->
00:39:55,927
Some of the bold claims that he has about
the near future of AI.
966
00:39:55,927 -->
00:40:02,550
He says everyone will have an
exceptionally capable personal agent that
understands
967
00:40:02,550 -->
00:40:06,572
you, your goals, and everything you care
about.
968
00:40:06,572 -->
00:40:09,433
That's a bold claim. Like,
969
00:40:09,691 -->
00:40:10,921
Sure, there are agents right now,
970
00:40:10,921 -->
00:40:12,822
but he's saying very personal.
971
00:40:12,822 -->
00:40:15,263
He also believes that everyone will have
one.
972
00:40:15,263 -->
00:40:19,264
He says the economy will become more
entrepreneurial because of this.
973
00:40:19,264 -->
00:40:24,765
Adults will have a superintelligent
learning assistant that knows exactly
974
00:40:24,765 -->
00:40:27,036
how to teach you new job skills,
975
00:40:27,036 -->
00:40:30,807
new languages, new hobbies, or anything
else you're interested in.
976
00:40:30,807 -->
00:40:36,139
Side question: if the superintelligence
can teach me all of that,
977
00:40:36,139 -->
00:40:37,139
why do I need to learn it?
978
00:40:37,590 -->
00:40:38,891
Well, hold on, let's pause.
979
00:40:39,056 -->
00:40:40,056
Okay, sure, please.
980
00:40:40,913 -->
00:40:42,334
Go back to the first thing you said.
981
00:40:42,334 -->
00:40:44,446
Everyone's gonna everyone's gonna
982
00:40:43,057 -->
00:40:44,408
The capable person Yeah.
983
00:40:44,446 -->
00:40:46,948
have a capable This is written by someone
who probably
984
00:40:46,948 -->
00:40:50,521
has fifteen personal assistant type
people,
985
00:40:50,521 -->
00:40:52,002
right? And
986
00:40:50,617 -->
00:40:51,885
~ sure, for sure.
987
00:40:52,002 -->
00:40:56,145
~ also I'm pretty sure that none of them
know everything about him.
988
00:40:56,861 -->
00:40:57,872
Sure, sure.
989
00:40:57,418 -->
00:40:58,418
And I'm pretty
990
00:40:57,899 -->
00:41:01,403
sure he doesn't want that. And I
definitely don't want all of my personal,
991
00:41:01,403 -->
00:41:04,787
everything deep personal, everything about
me living on a server somewhere
992
00:41:05,088 -->
00:41:06,950
so that an AI agent can get to it.
993
00:41:06,950 -->
00:41:09,724
Because what happens when the other agent
breaks into it?
994
00:41:09,724 -->
00:41:11,536
Because it's not like that ever happened
before.
995
00:41:11,536 -->
00:41:16,082
Right? Like i this is a I don't think this
is a future people actually want.
996
00:41:16,935 -->
00:41:18,886
I don't think it's feature people actually
want.
997
00:41:18,886 -->
00:41:23,398
And he also I I think he talks out of both
sides of his mouth because throughout
998
00:41:23,398 -->
00:41:28,801
this entire manifesto, he is trying to
make the plea that AI needs
999
00:41:28,801 -->
00:41:34,184
to not be centralized to a few specific
powerful companies,
1000
00:41:34,184 -->
00:41:35,735
slash people, slash governments.
1001
00:41:35,735 -->
00:41:39,957
It needs to be democratized. But then in a
different breath,
1002
00:41:39,957 -->
00:41:45,340
he says, quote, We Meta will offer free
versions that will be accessible
1003
00:41:45,340 -->
00:41:46,521
to billions of people.
1004
00:41:46,983 -->
00:41:50,215
So it's we want, we don't want a
centralized anything of power,
1005
00:41:50,215 -->
00:41:52,527
but also we will be the one that gives
everyone this power.
1006
00:41:52,527 -->
00:41:55,609
And it's like, I feel like that's a little
contradictory.
1007
00:41:55,609 -->
00:42:01,963
Fine. He also has this wild example of
trying to prove this point
1008
00:42:02,423 -->
00:42:08,307
that if one person had a AI
superintelligent lawyer
1009
00:42:08,607 -->
00:42:11,809
in the courtroom, but the other person
didn't,
1010
00:42:11,809 -->
00:42:13,490
that would be an unfair advantage.
1011
00:42:13,490 -->
00:42:16,232
So the future we need is that everyone
has.
1012
00:42:16,622 -->
00:42:19,814
A superintelligence lawyer on their side,
1013
00:42:19,814 -->
00:42:23,326
and then these superintelligence lawyers
will argue,
1014
00:42:23,326 -->
00:42:28,430
I guess, and do the case. And it's like
recently on Movies on the Side,
1015
00:42:28,430 -->
00:42:31,692
we reviewed the movie Mercy, which is
exactly this situation,
1016
00:42:31,692 -->
00:42:34,283
and it doesn't go well. Just throwing that
out there.
1017
00:42:34,283 -->
00:42:37,092
But just seems like a weird, I don't know,
1018
00:42:37,092 -->
00:42:39,567
it's a weird example. But these are the
kind of examples that he pulls
1019
00:42:39,567 -->
00:42:41,188
to try and prove his point.
1020
00:42:42,184 -->
00:42:47,317
Yeah, I feel like there's this is I think
that didn't they talk about this
1021
00:42:47,317 -->
00:42:52,920
on ~ The Verge cast recently because
Zuckerberg wrote about like what a
1022
00:42:53,241 -->
00:42:55,442
I think it's Neely who talks about
software brain.
1023
00:42:55,442 -->
00:42:56,442
There's
1024
00:42:55,717 -->
00:42:56,717
Yeah, yeah, yeah.
1025
00:42:56,422 -->
00:42:58,043
so much software brain in this.
1026
00:42:58,043 -->
00:43:01,315
There's so much. Also,
1027
00:42:58,598 -->
00:42:59,777
There's a lot of software brain.
1028
00:42:59,777 -->
00:43:01,330
Lot of software brain. ~
1029
00:43:01,315 -->
00:43:02,315
like
1030
00:43:01,515 -->
00:43:04,207
you could have the most super intelligent
lawyer in the entire world.
1031
00:43:04,207 -->
00:43:07,089
Fine. Could be an AI agent, fine.
1032
00:43:07,089 -->
00:43:09,452
But you know who decides cases?
1033
00:43:10,496 -->
00:43:12,588
Just randos off the street, twelve of
them,
1034
00:43:10,928 -->
00:43:13,496
Yeah. That is true.
1035
00:43:12,588 -->
00:43:13,849
sitting in a box, listening
1036
00:43:13,849 -->
00:43:15,891
to those super intelligent AIs for days.
1037
00:43:15,891 -->
00:43:16,891
Or a
1038
00:43:16,252 -->
00:43:17,252
That is true.
1039
00:43:16,771 -->
00:43:21,595
judge who has very little patience already
for AI written briefs that just make
1040
00:43:21,595 -->
00:43:23,157
cra like this
1041
00:43:22,362 -->
00:43:23,362
Yeah, yeah, yeah, yeah.
1042
00:43:23,157 -->
00:43:26,840
is like I'm sure this sounded good to him
when he wrote it,
1043
00:43:26,840 -->
00:43:28,471
but his persp you just have to understand,
1044
00:43:28,471 -->
00:43:31,214
like the perspective of the guy who owns
an island in Hawaii is
1045
00:43:31,214 -->
00:43:33,866
not the perspective of what most of us are
experiencing.
1046
00:43:34,177 -->
00:43:36,338
That's a good data point to keep in mind.
1047
00:43:36,338 -->
00:43:37,856
This dude owns an island in one.
1048
00:43:38,209 -->
00:43:41,411
I just and he and he just keeps up buying
a property and pushing people out.
1049
00:43:41,411 -->
00:43:44,443
Like cause in his mind, like the world
just revolves around.
1050
00:43:44,443 -->
00:43:47,285
Okay, can we just cut to why this exists
and then I'll let you keep going.
1051
00:43:47,285 -->
00:43:49,146
I'm sorry. Okay.
1052
00:43:48,314 -->
00:43:49,847
No no okay, go ahead. Yeah, why does this
exist?
1053
00:43:49,907 -->
00:43:53,700
Cause you said it. His core thesis here is
that it should
1054
00:43:53,700 -->
00:43:58,012
not be that superintelligence should not
be controlled by a few powerful companies.
1055
00:43:58,012 -->
00:44:00,154
Who are those powerful companies?
1056
00:44:00,154 -->
00:44:04,108
Yep. That's it.
1057
00:44:01,102 -->
00:44:04,339
It is OpenAI, Anthropic, Google.
1058
00:44:04,108 -->
00:44:06,068
Well, it's those two, really, right?
1059
00:44:04,339 -->
00:44:06,424
Sure.
1060
00:44:06,464 -->
00:44:08,145
I mean Google Sure, fine, whatever.
1061
00:44:08,145 -->
00:44:10,196
They've already won search, like all that
stuff.
1062
00:44:10,196 -->
00:44:14,437
So what do you say if you are not one of
the companies that is currently
1063
00:44:14,437 -->
00:44:15,999
in the race to control everything?
1064
00:44:15,999 -->
00:44:19,411
You say no one should control everything,
1065
00:44:19,298 -->
00:44:20,369
Exactly. Yes.
1066
00:44:19,411 -->
00:44:20,592
right? It should all
1067
00:44:20,592 -->
00:44:22,913
be open and everyone should have access to
it.
1068
00:44:22,913 -->
00:44:25,684
He thinks superintelligence should be a c
commodity.
1069
00:44:25,684 -->
00:44:30,507
And what do you do when one of your
competitors controls a scarce commodity?
1070
00:44:30,507 -->
00:44:33,348
You argue that the commodity should be
free.
1071
00:44:33,348 -->
00:44:36,070
And the other reason that this is
important is
1072
00:44:36,392 -->
00:44:38,794
Is that his entire business well,
1073
00:44:38,794 -->
00:44:43,248
first of all, I think there is definitely
a I want to be seen as cool.
1074
00:44:43,248 -->
00:44:45,109
This is what's cool.
1075
00:44:43,870 -->
00:44:45,026
~ for sure.
1076
00:44:45,469 -->
00:44:47,021
So I have to be in this space.
1077
00:44:47,021 -->
00:44:51,364
But there is actually an argument for why
artificial intelligence
1078
00:44:51,364 -->
00:44:55,417
is important to Facebook, and it is
because generative AI is capable
1079
00:44:55,417 -->
00:44:58,140
of infinity advertising. Right?
1080
00:44:57,404 -->
00:44:58,404
Correct.
1081
00:44:58,460 -->
00:45:00,652
And I've I've heard people talk about this
before.
1082
00:45:00,652 -->
00:45:03,654
I think maybe it's the VerchCast where
they talk about like a YouTube video,
1083
00:45:03,654 -->
00:45:06,036
every single pixel is mod maybe it's
actually Ben Thompson.
1084
00:45:06,366 -->
00:45:08,066
Every single pixel should be monetizable.
1085
00:45:08,066 -->
00:45:12,188
You hold up a product, right? We're people
looking at our video feed right
1086
00:45:12,188 -->
00:45:13,308
now should be able to be like,
1087
00:45:13,308 -->
00:45:14,778
where did that frame come from?
1088
00:45:14,778 -->
00:45:15,949
Where did this lamp come from?
1089
00:45:15,949 -->
00:45:17,029
Where did this book come from?
1090
00:45:17,029 -->
00:45:19,990
Like there should be able to everything
single pixel should be monetizable.
1091
00:45:19,990 -->
00:45:22,571
That is not possible without AI,
1092
00:45:22,571 -->
00:45:24,011
right? Infinity
1093
00:45:23,024 -->
00:45:24,024
Right, right.
1094
00:45:24,011 -->
00:45:26,892
ad content is what what Facebook wants.
1095
00:45:26,892 -->
00:45:30,543
Infinity and so, but they cannot they've
been through the experience.
1096
00:45:30,543 -->
00:45:33,274
It's kind of like people talk about Apple
never wanting to be beholden
1097
00:45:33,274 -->
00:45:34,444
to someone else because like
1098
00:45:34,645 -->
00:45:36,777
Adobe and Microsoft had them over the
barrel,
1099
00:45:36,777 -->
00:45:41,791
right? Zuckerberg just never wants to be
beholden to someone else's platform
because
1100
00:45:41,791 -->
00:45:44,343
that's been the case with Apple for for
decades.
1101
00:45:44,343 -->
00:45:50,277
And so they cannot have a world where all
of this capability
1102
00:45:50,277 -->
00:45:51,949
is controlled by someone else.
1103
00:45:51,949 -->
00:45:52,949
So it has
1104
00:45:52,283 -->
00:45:53,283
Right.
1105
00:45:52,850 -->
00:45:54,921
to be open. So when you read this,
1106
00:45:54,921 -->
00:45:58,174
it sounds super altruistic and super good
for the good of humanity.
1107
00:45:58,174 -->
00:45:59,385
Just understand
1108
00:45:58,545 -->
00:45:59,545
Right.
1109
00:45:59,385 -->
00:46:03,388
that the actual reason is it is an
existential threat to meta.
1110
00:46:04,590 -->
00:46:08,002
It's exactly that. And we've mentioned
before,
1111
00:46:08,002 -->
00:46:11,344
but Zuck missed the mobile revolution.
1112
00:46:11,344 -->
00:46:13,775
Facebook never got to be in the hardware,
1113
00:46:13,775 -->
00:46:18,168
make a phone there. They missed the search
revolution.
1114
00:46:18,168 -->
00:46:22,401
Google got that. And now it feels like
they are missing the AI revolution.
1115
00:46:22,401 -->
00:46:26,333
I mean, I have never heard a soul say,
1116
00:46:26,333 -->
00:46:30,215
I asked the meta AI app about this or
that.
1117
00:46:30,215 -->
00:46:31,356
It is
1118
00:46:32,100 -->
00:46:35,012
90% of the time ChatGPT. Just normal
people,
1119
00:46:35,012 -->
00:46:36,904
not techie people, just normal people,
1120
00:46:36,904 -->
00:46:39,205
they're like, Chat GPT, ChatGPT,
1121
00:46:39,205 -->
00:46:43,248
this and that. And then when you move a
little bit into the nerdier space,
1122
00:46:43,248 -->
00:46:46,391
you'll hear Claude, maybe some Gemini.
1123
00:46:46,391 -->
00:46:49,754
And like there was other news this week
where Gemini reached a billion users,
1124
00:46:49,754 -->
00:46:54,327
and it's like, Google, if you're counting
every Google search as a Gemini user,
1125
00:46:54,327 -->
00:46:58,150
because that's what it is. Right.
1126
00:46:54,432 -->
00:46:55,693
They are because that's what it is.
1127
00:46:55,693 -->
00:46:57,525
So most people are using Gemini,
1128
00:46:57,525 -->
00:46:58,525
they just don't think about
1129
00:46:58,316 -->
00:47:00,779
it as using AI yet because they aren't
going
1130
00:47:00,368 -->
00:47:01,368
Right.
1131
00:47:00,779 -->
00:47:01,950
intentionally into a chat bot.
1132
00:47:02,229 -->
00:47:06,561
Right. But when any I I would argue that
ninety percent of people when they think
1133
00:47:06,561 -->
00:47:09,472
of AI or an AI chat, it's Chat GPT.
1134
00:47:09,472 -->
00:47:14,864
And so I think it's becoming increasingly
clear Meta is losing this battle as well.
1135
00:47:14,864 -->
00:47:17,255
And it feels like that was probably the
impetus for this,
1136
00:47:17,255 -->
00:47:21,657
where Mark Zuckerberg is trying to wave
his arms around and on one hand
1137
00:47:21,657 -->
00:47:24,778
try to be not doomsday-ish,
1138
00:47:24,778 -->
00:47:27,779
but then he also talks about a lot of
other doomsday-ish scenarios that I'll
1139
00:47:27,779 -->
00:47:29,510
get to in a second. And
1140
00:47:30,148 -->
00:47:34,672
Like you're saying, feigning altruism in
the hopes that he could be
1141
00:47:34,933 -->
00:47:37,275
in the room again. And like on one level,
1142
00:47:37,275 -->
00:47:40,959
I almost feel a little bad because it
feels like he's the guy that keeps getting
1143
00:47:40,959 -->
00:47:44,552
left out of the cool kids group because he
got left out of the phone thing,
1144
00:47:44,552 -->
00:47:45,944
got left out of this. But also,
1145
00:47:45,944 -->
00:47:48,297
he has built some of the biggest
companies,
1146
00:47:48,297 -->
00:47:50,468
the most profitable companies in the
world.
1147
00:47:50,468 -->
00:47:53,269
I just don't think he's proud of them.
1148
00:47:54,050 -->
00:47:55,491
Well he built one and he bought two.
1149
00:47:55,491 -->
00:48:01,134
And he in and I I do hear people talk.
1150
00:47:56,879 -->
00:48:01,756
Fair point. Fair point. And the one he
built is dying,
1151
00:48:01,134 -->
00:48:04,276
Yeah, that's good. That's a hundred
percent true,
1152
00:48:01,756 -->
00:48:04,821
just to be clear. Like Facebook the one he
built is dying.
1153
00:48:04,276 -->
00:48:05,276
right?
1154
00:48:05,036 -->
00:48:08,938
Yes. And the one that he the one that he
bought that is doing really well
1155
00:48:09,159 -->
00:48:12,641
is literally just Snapchat, all the good
ideas stolen.
1156
00:48:12,641 -->
00:48:14,912
And TikTok, all the good ideas stolen.
1157
00:48:12,997 -->
00:48:14,942
~ and TikTok 'cause of reels. Yeah,
1158
00:48:14,912 -->
00:48:15,912
Right. And
1159
00:48:14,942 -->
00:48:15,942
exactly.
1160
00:48:15,652 -->
00:48:20,495
so yeah, he's the exactly. They they they
are dominant purely because
1161
00:48:20,495 -->
00:48:22,076
of their network effect that they already
have.
1162
00:48:22,432 -->
00:48:26,785
Right, and that's ~ and in Instagram's
entire user base was built on Twitter.
1163
00:48:26,785 -->
00:48:28,225
Let's not forget that at the time.
1164
00:48:28,225 -->
00:48:32,068
So I 'cause it was a filters app.
1165
00:48:29,061 -->
00:48:32,889
Sure. Yeah. Right.
1166
00:48:32,068 -->
00:48:33,068
The feed wasn't
1167
00:48:32,958 -->
00:48:36,370
the thing in Instagram. You shared your f
Instagram photos on Twitter.
1168
00:48:36,370 -->
00:48:40,813
And then you could and when you you could
also can take all of your Twitter users
1169
00:48:41,033 -->
00:48:43,555
and create your social graph on Instagram
at the beginning.
1170
00:48:43,555 -->
00:48:46,336
So all of these things, like the
innovation here,
1171
00:48:46,336 -->
00:48:47,917
it's not really happening.
1172
00:48:49,010 -->
00:48:50,010
In y the cool
1173
00:48:49,533 -->
00:48:50,533
Yeah.
1174
00:48:49,861 -->
00:48:53,514
kid thing, you're right. He he also
changed the name of his company because
1175
00:48:53,514 -->
00:48:56,766
of his big bet on a thing that no one is
using,
1176
00:48:56,766 -->
00:48:58,418
the metaverse. It's not real.
1177
00:48:57,871 -->
00:48:58,871
Which
1178
00:48:59,752 -->
00:49:05,764
and which feels like the most like he
keeps trying to make big bet after big
bet,
1179
00:49:05,764 -->
00:49:11,276
hoping one of them pays off. It's like the
it's like the guy who either you know,
1180
00:49:11,276 -->
00:49:13,727
at the craps table or behind the lottery
tick,
1181
00:49:13,727 -->
00:49:16,178
he's like, This time I'm gonna hit it big.
1182
00:49:16,178 -->
00:49:19,170
And the thing is like you I don't feel bad
for him because he's filthy rich.
1183
00:49:19,170 -->
00:49:21,561
He owns an island in Hawaii, like he has
successful companies,
1184
00:49:21,561 -->
00:49:26,423
but I also get he's the things that he
built and have bought are
1185
00:49:26,423 -->
00:49:27,483
not the cool things.
1186
00:49:28,155 -->
00:49:31,133
And everyone's talking about AI right now
and no one's talking about meta
1187
00:49:31,133 -->
00:49:32,133
in that conversation.
1188
00:49:32,203 -->
00:49:37,268
Right. And it's because I mean Lama is
maybe a very capable frontier model.
1189
00:49:37,268 -->
00:49:38,839
That's great. And it's not open source.
1190
00:49:38,839 -->
00:49:41,592
It's open weight. There's a there's an
important difference between difference
1191
00:49:41,592 -->
00:49:45,495
there, right? But I think that you're yes,
1192
00:49:45,495 -->
00:49:47,717
the this is not Zuck's manifesto,
1193
00:49:47,717 -->
00:49:49,439
this is Zuck's complex. Right.
1194
00:49:49,439 -->
00:49:51,000
And it's like he
1195
00:49:49,739 -->
00:49:50,739
That's right.
1196
00:49:51,000 -->
00:49:52,722
just you're right, he doesn't want to feel
left out.
1197
00:49:52,722 -->
00:49:54,183
He's he's got the Hamilton complex.
1198
00:49:54,183 -->
00:49:55,525
He wants to be in the room where it
happens,
1199
00:49:55,525 -->
00:49:56,525
right?
1200
00:49:56,037 -->
00:49:57,037
Yes,
1201
00:49:57,100 -->
00:49:58,100
yes, said
1202
00:49:57,116 -->
00:49:58,116
And there was two
1203
00:49:57,816 -->
00:50:00,557
quotes right at the beginning that set the
stage for this.
1204
00:50:00,557 -->
00:50:04,808
He says, historically, hoping that an
absolute power will benevolent benevolent
1205
00:50:04,808 -->
00:50:09,549
benevolently provide for humanity if
sufficiently enlightened has not led
1206
00:50:09,549 -->
00:50:11,730
to safe or positive outcomes. Right?
1207
00:50:11,730 -->
00:50:15,361
Historically, hoping that an absolute
power will be good for us has not
1208
00:50:15,361 -->
00:50:17,692
led to safe outcomes. And then he says,
1209
00:50:17,692 -->
00:50:20,132
as everyone gains more powerful tools,
1210
00:50:20,132 -->
00:50:21,873
this is your super lawyer thing,
1211
00:50:21,873 -->
00:50:25,284
each person will become more capable of
shaping the future,
1212
00:50:25,284 -->
00:50:26,284
not less.
1213
00:50:26,538 -->
00:50:31,755
Has he met people? Do you want to give
everyone more powerful tools?
1214
00:50:31,755 -->
00:50:36,632
Because sometimes being more capable of
shaping the future is not a great thing.
1215
00:50:36,632 -->
00:50:41,559
Do you know what happens when people who
can afford infinity lawyers start hiring
1216
00:50:41,559 -->
00:50:43,641
lawyers? It's not a good thing.
1217
00:50:44,237 -->
00:50:49,859
Yeah. And he also has some awkward like
caveats where he says
1218
00:50:49,859 -->
00:50:51,639
we need to democratize you know,
1219
00:50:51,639 -->
00:50:53,710
AI needs to be in the hands of everyone.
1220
00:50:53,710 -->
00:50:56,971
But that's also gonna create questions of
like,
1221
00:50:56,971 -->
00:51:03,002
well, what laws and what systems do we
lean on as we build this personal
1222
00:51:03,002 -->
00:51:09,224
AI revolution? And he talks about mainly a
Western democratized version
1223
00:51:09,224 -->
00:51:11,454
of economy and government. And I like,
1224
00:51:11,454 -->
00:51:12,847
so
1225
00:51:13,711 -->
00:51:19,254
By everyone, you mean the West Because
those are just like and I don't even know
1226
00:51:19,254 -->
00:51:24,618
if he understands the weirdness of that
those kind of caveats that he's saying.
1227
00:51:24,618 -->
00:51:27,039
So that's why I wonder, like, maybe he did
write this,
1228
00:51:27,039 -->
00:51:29,230
because that's I don't know. I feel like a
PR person would like,
1229
00:51:29,230 -->
00:51:34,584
you know, you just kinda like dissed a
large You just you just slammed like half
1230
00:51:30,400 -->
00:51:33,017
You shouldn't say that out loud.
1231
00:51:33,017 -->
00:51:35,948
Yeah.
1232
00:51:34,584 -->
00:51:35,584
the globe
1233
00:51:35,965 -->
00:51:38,446
while you're trying to say you're making
this for the globe.
1234
00:51:38,446 -->
00:51:39,767
So okay.
1235
00:51:40,708 -->
00:51:42,970
And anyway, I'll just throw out a couple
details because we don't have
1236
00:51:42,970 -->
00:51:44,170
to spend too much time on this,
1237
00:51:44,170 -->
00:51:47,993
but he talks about the good that Meta is
doing.
1238
00:51:47,993 -->
00:51:52,316
And obviously, data centers are a huge
pain point right now here in the States.
1239
00:51:52,316 -->
00:51:55,918
And he talks about that Meta is building a
data center in Louisiana,
1240
00:51:55,918 -->
00:52:00,241
and teachers in that area received a fifty
thousand dollar bonus because
1241
00:52:00,241 -->
00:52:02,382
of the increased tax revenue from our
investment.
1242
00:52:02,382 -->
00:52:05,605
Gruber quoted that and linked to another
article.
1243
00:52:05,605 -->
00:52:09,027
Apparently that is real. Like teachers did
get a fifty thousand dollar bonus.
1244
00:52:09,463 -->
00:52:11,425
in that town because of the data center.
1245
00:52:11,425 -->
00:52:14,327
Great for those teachers. That's a tough
gig.
1246
00:52:14,327 -->
00:52:16,809
Underpaid, like more power to them.
1247
00:52:16,809 -->
00:52:19,631
I don't see how that's going to change
anybody's opinion on data centers
1248
00:52:19,631 -->
00:52:22,023
as a whole and whether they want it in
their backyard,
1249
00:52:22,023 -->
00:52:24,549
because they're not giving fifty thousand
dollars to every resident of this town.
1250
00:52:24,549 -->
00:52:26,400
They're giving it to the teachers of this
town.
1251
00:52:26,400 -->
00:52:27,868
So that's I don't know. Again,
1252
00:52:27,868 -->
00:52:32,682
feels like a weird justification kind of
like I don't know.
1253
00:52:32,682 -->
00:52:33,682
Whatever that is.
1254
00:52:33,504 -->
00:52:36,556
Well, no, what it is is a is a let's like
let's be clear.
1255
00:52:36,556 -->
00:52:41,539
It's just you find a sympathetic group of
h humans and you and you just it's
1256
00:52:41,539 -->
00:52:43,810
a bribe. I'm not saying that the teachers
were the people that made
1257
00:52:42,605 -->
00:52:45,333
It's right, right, right.
1258
00:52:43,810 -->
00:52:45,571
the decision about it, but like you find a
sympathetic
1259
00:52:45,571 -->
00:52:47,923
group of people, you give them a huge
amount of money,
1260
00:52:47,923 -->
00:52:51,054
and then it makes you look good so that
you can do this other thing.
1261
00:52:51,054 -->
00:52:53,676
Like this is that outcome great?
1262
00:52:53,676 -->
00:52:56,448
Sure. But let's just be clear about what's
happening here.
1263
00:52:56,448 -->
00:52:57,448
You're just gonna
1264
00:52:56,763 -->
00:52:57,763
Right, like
1265
00:52:57,268 -->
00:52:59,414
throw money at people to get them to do
what you want.
1266
00:52:59,414 -->
00:53:00,414
It's called a bribe.
1267
00:53:00,653 -->
00:53:03,185
It is thank you. That is what it's called.
1268
00:53:03,185 -->
00:53:06,988
And then he go even goes on to talk about
bioweapons,
1269
00:53:06,988 -->
00:53:11,352
cybersecurity. He talks about China and
how they're bringing online
1270
00:53:11,352 -->
00:53:14,814
one gigawatt plus of nuclear capacity
every other week.
1271
00:53:14,814 -->
00:53:17,296
The US needs to catch up in the
development.
1272
00:53:17,296 -->
00:53:22,801
That the American government needs to
allow AI companies to run and not
1273
00:53:22,823 -->
00:53:26,634
h hold them back. Otherwise we won't be
competitive on the global scale when
1274
00:53:26,634 -->
00:53:30,137
it comes to AI models. And which
1275
00:53:30,723 -->
00:53:33,644
Not surprising, he's very
pro-distillation.
1276
00:53:33,644 -->
00:53:39,427
He said, you know, we need to allow
companies to distill from each other.
1277
00:53:39,427 -->
00:53:42,529
And to ~ I think a fair point,
1278
00:53:42,529 -->
00:53:45,881
AI is was learned from all of human
knowledge.
1279
00:53:45,881 -->
00:53:50,383
Like it was distilled from the the raw
material of what humans have done
1280
00:53:50,383 -->
00:53:54,345
for hundreds of years. And so we should be
able to distill other models.
1281
00:53:54,345 -->
00:53:57,470
But again, it's because he wants his
models to be good too.
1282
00:53:57,470 -->
00:53:59,818
So he's like, I would love to distill from
fable.
1283
00:54:00,118 -->
00:54:01,649
and soul five point six or whatever.
1284
00:54:01,649 -->
00:54:06,973
And so anyway. And then his final promise
is we will build a fully private mode
1285
00:54:06,973 -->
00:54:10,365
where even Meta cannot see or grant access
to your information.
1286
00:54:10,365 -->
00:54:14,288
Hard sell on that, Meta. I don't know if I
believe that.
1287
00:54:14,288 -->
00:54:18,131
And that he says open source is a positive
and important force
1288
00:54:18,131 -->
00:54:23,034
for empowering people. Again, it just
feels like he is wanting things that would
1289
00:54:23,034 -->
00:54:24,955
benefit him and Meta in general,
1290
00:54:24,955 -->
00:54:27,007
like being able to distill other models,
1291
00:54:27,007 -->
00:54:28,378
claim things are open source,
1292
00:54:28,824 -->
00:54:31,908
not have centralized power among anthropic
and open AI,
1293
00:54:31,908 -->
00:54:33,659
although he never mentions them in this
article.
1294
00:54:33,659 -->
00:54:36,933
And so he can be part of this whole thing.
1295
00:54:36,933 -->
00:54:39,326
And it's just I don't know. It fell very
well.
1296
00:54:39,830 -->
00:54:43,433
Well yeah, and it's the whole point,
1297
00:54:43,433 -->
00:54:45,215
it's just important to remember this,
1298
00:54:45,215 -->
00:54:49,439
is that w what Mark Zuckerberg is afraid
of,
1299
00:54:49,439 -->
00:54:54,824
one, there is the I want to be a part of
the cool thing and I cannot beat these
1300
00:54:54,824 -->
00:54:56,686
companies and I need them to not win.
1301
00:54:56,686 -->
00:55:01,651
And two, he needs them to not win because
it is an existential threat threat
1302
00:55:01,651 -->
00:55:04,633
to his business. Right? He even s talks
about like
1303
00:55:05,314 -->
00:55:06,915
He doesn't say this outright, but it's
like,
1304
00:55:06,915 -->
00:55:10,857
yeah, these other guys, their j their goal
is to essentially replace workers.
1305
00:55:10,857 -->
00:55:14,757
But what we want to do is turn everyone
into an entrepreneur by giving them these
1306
00:55:14,757 -->
00:55:18,422
like it's like, I mean, I don't think any
of either of these things are
1307
00:55:18,742 -->
00:55:23,795
are gonna be the case because just bec
just what it what exactly kind
1308
00:55:23,795 -->
00:55:27,527
of an entrepreneur do you want just
generating AI slap videos to put into your
feed
1309
00:55:27,527 -->
00:55:29,959
so that more people will watch them so you
can sell more ads?
1310
00:55:29,959 -->
00:55:32,200
Like let's just be clear about what we're
actually talking about.
1311
00:55:32,536 -->
00:55:35,361
Yeah. Yeah. So anyway, that was his
manifesto.
1312
00:55:35,361 -->
00:55:36,513
You don't have to read it everybody.
1313
00:55:36,513 -->
00:55:38,226
We we we tried to sum it up for you,
1314
00:55:38,226 -->
00:55:39,226
so but we'll
1315
00:55:38,646 -->
00:55:39,646
We didn't do a
1316
00:55:39,247 -->
00:55:41,934
very good job, but that's because it's a ~
1317
00:55:41,851 -->
00:55:44,212
It's hard to summarize. He said a lot of
words.
1318
00:55:42,616 -->
00:55:44,360
it's yeah. It's
1319
00:55:44,360 -->
00:55:45,360
a lot of words.
1320
00:55:45,312 -->
00:55:46,323
We'll link it in the show notes.
1321
00:55:46,323 -->
00:55:48,944
~ but I I'm gonna skip a go things because
I do want to get
1322
00:55:48,944 -->
00:55:53,695
to the anthropic watermarking text that it
generates.
1323
00:55:53,695 -->
00:55:57,537
And I will link to one, an article from
The Economist,
1324
00:55:57,537 -->
00:56:02,058
which I saw Gruber link to. They did this
huge study on how to spot AI writing.
1325
00:56:02,058 -->
00:56:05,350
And they tested ChatGPT, Claude,
1326
00:56:05,350 -->
00:56:09,941
Gemini, and Grok. They had it write
versions of articles and they compared
1327
00:56:09,941 -->
00:56:11,492
it to actual human writing.
1328
00:56:12,126 -->
00:56:14,838
And they've had some surprising findings.
1329
00:56:14,838 -->
00:56:15,999
I'll just read one poll quote.
1330
00:56:15,999 -->
00:56:17,980
It says Our findings are surprising.
1331
00:56:17,980 -->
00:56:22,153
AI prose is distinguishable by word and
punctuation choice,
1332
00:56:22,153 -->
00:56:24,855
as well as sentence and paragraph
structure,
1333
00:56:24,855 -->
00:56:27,056
but its hallmarks are not what you might
expect,
1334
00:56:27,056 -->
00:56:30,078
partly because its writing style has
changed with the updates.
1335
00:56:30,078 -->
00:56:33,160
That does not mean that LLMs are great
writers.
1336
00:56:33,160 -->
00:56:37,864
Their prose lacks lucidity and elegance
and is often formulaic.
1337
00:56:37,864 -->
00:56:41,156
So those aspiring to be impressive human
storytellers
1338
00:56:41,466 -->
00:56:44,508
should avoid the following peculiarities
in their own prose.
1339
00:56:44,508 -->
00:56:47,189
I think I have yet to read this whole
study.
1340
00:56:47,189 -->
00:56:51,431
It is behind the Economist. You have to do
like a free trial if you want to read it.
1341
00:56:51,431 -->
00:56:53,913
The gift link ~ expired already.
1342
00:56:53,913 -->
00:56:58,205
But I'm very curious fascinating that they
did this study and I think
1343
00:56:58,205 -->
00:57:03,958
to no surprise, there is formulas to how
AI writes content and text.
1344
00:57:03,958 -->
00:57:09,221
And so that in light of because of
pressure from the EU,
1345
00:57:09,221 -->
00:57:11,142
Anthropic is introducing
1346
00:57:11,634 -->
00:57:17,639
A text watermarking to all text generated
by Claude.
1347
00:57:17,639 -->
00:57:22,964
And so it's saying that new models will
mark AI generated content from day one.
1348
00:57:22,964 -->
00:57:28,548
Claude models launched in the EU on or
after August 2nd will support machine
1349
00:57:28,548 -->
00:57:34,734
readable marking at launch. So this sounds
like the older models or models that have
1350
00:57:34,734 -->
00:57:37,006
already existed, they're still working on
them the marking,
1351
00:57:37,006 -->
00:57:39,798
but any new models going forward will have
this from day one.
1352
00:57:40,269 -->
00:57:44,641
Old models, it's coming soon. And this
watermarking that Claude is going
1353
00:57:44,641 -->
00:57:47,503
to use will apply to all output,
1354
00:57:47,503 -->
00:57:49,504
whether you're using the Claude app,
1355
00:57:49,504 -->
00:57:52,685
Claude Code, Cowork, or even the API,
1356
00:57:52,685 -->
00:57:55,497
which made me think, just curious.
1357
00:57:55,497 -->
00:57:59,209
Like I have a bunch of shortcuts that use
the Anthropic API to
1358
00:57:59,209 -->
00:58:00,669
do things like summarize stuff.
1359
00:58:00,669 -->
00:58:05,352
I am very curious. Like, is that text
watermarked coming through my shortcuts?
1360
00:58:05,352 -->
00:58:07,383
I thought that was very curious.
1361
00:58:07,383 -->
00:58:09,534
And I it seems like the answer is yes.
1362
00:58:09,826 -->
00:58:11,907
according to this Claude support article.
1363
00:58:11,907 -->
00:58:15,029
They're also going to help users detect
Claude marks.
1364
00:58:15,029 -->
00:58:18,371
And they say that existing models are in
progress of being watermarked.
1365
00:58:18,371 -->
00:58:25,325
And the they specify like detecting a
Claude watermark tells
1366
00:58:25,325 -->
00:58:28,036
you that the content may have been
processed by Claude,
1367
00:58:28,036 -->
00:58:31,258
but it doesn't confirm whether it was
fully generated,
1368
00:58:31,258 -->
00:58:33,069
whether it just went through Claude.
1369
00:58:33,069 -->
00:58:37,621
Like for instance, if you write a
newsletter and you just ask Claude
1370
00:58:37,621 -->
00:58:38,692
to proofread it,
1371
00:58:39,414 -->
00:58:44,666
It will mark it. And someone on the
receiving end might have tools from Claude
1372
00:58:44,666 -->
00:58:48,557
to see that it is watermarked and could
assume that you just generated this
1373
00:58:48,557 -->
00:58:53,038
newsletter, but it's you have to
understand,
1374
00:58:53,038 -->
00:58:57,259
like it might just be that someone used it
to proofread or you know,
1375
00:58:57,259 -->
00:58:59,120
process some small part of the text.
1376
00:58:59,120 -->
00:59:02,901
So I find this interesting. Like we've
talked about on this show,
1377
00:59:02,901 -->
00:59:06,362
we do need tools, especially when it comes
to photo.
1378
00:59:06,788 -->
00:59:10,620
video and audio generation that when
something is AI generated.
1379
00:59:10,620 -->
00:59:12,501
And while Spotify is trying to kill
podcasts,
1380
00:59:12,501 -->
00:59:15,213
it's also going to do things to try and
label AI music,
1381
00:59:15,213 -->
00:59:16,213
which I'll get to in a minute.
1382
00:59:16,183 -->
00:59:20,866
But I find this fas is the first time that
I've seen like a text marked,
1383
00:59:20,866 -->
00:59:25,458
a watermark on text. It's still like I
it's hard for me to wrap my mind around
like
1384
00:59:25,458 -->
00:59:30,531
what exactly this is. Is this like
markings on the literal pixels of
1385
00:59:30,531 -->
00:59:34,623
the text that's generated? Is it hidden
metadata behind like I'm
1386
00:59:34,623 -->
00:59:36,134
not exactly sure what it is.
1387
00:59:36,846 -->
00:59:40,589
But it means that whether you just process
or proofread with Claude,
1388
00:59:40,589 -->
00:59:45,283
your content might be marked as AI not
generated,
1389
00:59:45,283 -->
00:59:47,696
but just watermarked that Claude had
something to do with it.
1390
00:59:47,696 -->
00:59:48,837
What what are your thoughts on this?
1391
00:59:48,837 -->
00:59:49,837
I mean you're
1392
00:59:49,662 -->
00:59:52,905
Yeah, a couple of things. I d
congratulations,
1393
00:59:52,905 -->
00:59:55,357
EU, you get to regulate tech across the
world.
1394
00:59:55,357 -->
00:59:57,388
This is what you wanted. So good job.
1395
00:59:57,388 -->
01:00:02,252
Nice nice work. You you know, and so in
his daily update,
1396
01:00:02,252 -->
01:00:06,025
Ben Thompson I this is like the fourth
time w we should get a
1397
01:00:06,025 -->
01:00:08,087
cut of every subscription to Stratekery.
1398
01:00:08,087 -->
01:00:10,219
So ~ the
1399
01:00:11,755 -->
01:00:12,755
Sure.
1400
01:00:12,561 -->
01:00:16,944
anyway, he explains sort of how this
works.
1401
01:00:16,944 -->
01:00:17,944
No well
1402
01:00:17,324 -->
01:00:18,324
Okay.
1403
01:00:17,715 -->
01:00:19,286
no, he explains a his
1404
01:00:19,446 -->
01:00:21,117
how he thinks it works, right?
1405
01:00:20,854 -->
01:00:21,854
Right, right.
1406
01:00:21,117 -->
01:00:22,178
And
1407
01:00:22,178 -->
01:00:26,131
'cause and it has to do with tokens and it
has to do with like lists
1408
01:00:26,131 -->
01:00:31,194
of tokens of how like different things
work and how instead
1409
01:00:31,194 -->
01:00:33,355
of being purely probabilistic,
1410
01:00:33,355 -->
01:00:36,818
the generation becomes somewhat
deterministic in that it chooses from
1411
01:00:36,818 -->
01:00:39,960
a green list more often than it chooses
randomly between the two lists.
1412
01:00:39,960 -->
01:00:44,563
And that somehow that that's detectable by
other tools that are like right.
1413
01:00:44,563 -->
01:00:47,895
Like it's it's it doesn't it supposedly
doesn't change the output,
1414
01:00:47,895 -->
01:00:48,895
but it actually
1415
01:00:49,154 -->
01:00:51,636
Kind of does because of like what you were
just saying,
1416
01:00:51,636 -->
01:00:53,806
like word structure, sentence structure
and stuff.
1417
01:00:53,806 -->
01:00:56,830
So I think it's the reason that this
survives things like copy paste,
1418
01:00:56,830 -->
01:01:00,053
whatever, is it's actually the structure
of the output.
1419
01:01:00,053 -->
01:01:03,216
This is what he says. He says Claude
doesn't explain how watermarking would
work
1420
01:01:03,216 -->
01:01:07,239
beyond that. But the general concept for
text watermarking is described
1421
01:01:07,239 -->
01:01:11,542
in a white paper. In brief, LLMs don't
produce the next token
determinalistically.
1422
01:01:11,542 -->
01:01:15,746
The the model produces a probability
distribution over possible next tokens
1423
01:01:15,746 -->
01:01:18,158
and one is selected with some degree of
randomness.
1424
01:01:18,391 -->
01:01:21,883
the watermark uses a keyed hash to
classify all of those tokens into
1425
01:01:21,883 -->
01:01:24,835
a green list and a red list. And the green
list gets a slight boost
1426
01:01:24,855 -->
01:01:27,837
to its probabil probability, I cannot talk
today,
1427
01:01:27,837 -->
01:01:32,060
of being picked. What that means is that
while an unmark watermarked output would
1428
01:01:32,060 -->
01:01:34,982
choose from the green list and the red
list roughly equally,
1429
01:01:34,982 -->
01:01:39,915
the watermarked output would have a
statistically meaningfully higher
percentage
1430
01:01:39,915 -->
01:01:41,116
of tokens from the green list.
1431
01:01:41,116 -->
01:01:42,346
So there's some like
1432
01:01:42,168 -->
01:01:43,168
Okay.
1433
01:01:42,607 -->
01:01:45,418
it's like a code built into the the
writing,
1434
01:01:45,418 -->
01:01:47,070
literally, but the the problem
1435
01:01:46,351 -->
01:01:47,351
Interesting.
1436
01:01:47,070 -->
01:01:48,070
here is like
1437
01:01:48,744 -->
01:01:51,235
If you're using it for things like
proofing,
1438
01:01:51,640 -->
01:01:52,640
Right.
1439
01:01:52,056 -->
01:01:54,937
that is not the same as if you just
generated text.
1440
01:01:54,937 -->
01:02:01,527
Also, I feel like d AI generated text is
significantly less problematic than
1441
01:02:01,527 -->
01:02:03,342
AI generated video and photos.
1442
01:02:04,565 -->
01:02:09,128
It's less press problematic at scale,
1443
01:02:09,128 -->
01:02:13,190
maybe, but when it's passed as human
writing,
1444
01:02:13,190 -->
01:02:18,443
I think it's problematic and it's easier
to pass it as human writing to most
people.
1445
01:02:18,806 -->
01:02:24,031
But the the the problem with AI generated
text past as human writing
1446
01:02:24,031 -->
01:02:28,335
is different than the problem of AI
generated videos and photos past as real,
1447
01:02:28,335 -->
01:02:31,978
because an AI generated photo and an AI
generated video is being passed
1448
01:02:31,978 -->
01:02:35,081
off as though it is a real capture of a
real thing that happened.
1449
01:02:35,081 -->
01:02:38,043
And anytime you read text about a thing,
1450
01:02:38,043 -->
01:02:40,175
you're one level removed from it,
1451
01:02:40,175 -->
01:02:44,229
right? AI generated fiction, like stories,
1452
01:02:43,289 -->
01:02:44,289
Sure.
1453
01:02:44,509 -->
01:02:45,509
fine.
1454
01:02:45,462 -->
01:02:48,145
The fact that it wasn't generated by a
human might mean that it's
1455
01:02:48,145 -->
01:02:50,627
not an interesting story, it might mean
that it's not a good story,
1456
01:02:50,627 -->
01:02:53,920
it might mean that someone is just using
AI to generate their
1457
01:02:53,920 -->
01:02:55,282
own version of Gulliver's travels,
1458
01:02:55,282 -->
01:02:58,845
like whatever. Like that's different
though than I post a photograph
1459
01:02:58,845 -->
01:03:01,588
of Steven breaking into the back of my car
and stealing my walk.
1460
01:03:02,487 -->
01:03:04,399
Which I did in fact do. The
1461
01:03:03,647 -->
01:03:04,647
Yes, twice.
1462
01:03:04,939 -->
01:03:10,204
So like my my initial concern was the way
I use ALLMs,
1463
01:03:10,204 -->
01:03:12,306
like when I make a YouTube video,
1464
01:03:12,306 -->
01:03:15,649
I don't use AI until the end of the
process.
1465
01:03:15,649 -->
01:03:19,432
I give it my transcript and I ask it for
title and description ideas.
1466
01:03:19,432 -->
01:03:22,495
I A B title test pretty much every video.
1467
01:03:22,495 -->
01:03:26,178
I usually have some titles that I
brainstormed throughout the process.
1468
01:03:26,178 -->
01:03:28,941
I'll use one or two of those. I'll ask
Claude
1469
01:03:29,882 -->
01:03:33,495
For some other title ideas, and sometimes
I'll just copy and paste one in
1470
01:03:33,495 -->
01:03:37,008
the A B test. Sometimes I'll copy one and
change a word or two.
1471
01:03:37,008 -->
01:03:42,092
And so does that mean once this
watermarking is live in all the models,
1472
01:03:42,092 -->
01:03:47,496
will YouTube think that my content is AI
generated?
1473
01:03:47,496 -->
01:03:52,981
Because Anthropic admits in this sport
article that it is not to assume that even
1474
01:03:52,981 -->
01:03:56,494
if something has the watermark that it's
wholly generated,
1475
01:03:56,494 -->
01:03:58,062
but we can't like
1476
01:03:58,435 -->
01:04:03,639
tell you that. Like it won't be
immediately available that this
1477
01:04:03,639 -->
01:04:07,201
was either completely AI generated or this
was processed by AI.
1478
01:04:07,201 -->
01:04:09,873
That that distinction does not exist in
this first iteration.
1479
01:04:09,873 -->
01:04:12,915
So all that to say, me personally,
1480
01:04:12,915 -->
01:04:15,456
I'm probably gonna not copy and paste
stuff anymore.
1481
01:04:15,456 -->
01:04:20,140
Even just simple things like titles for my
YouTube videos.
1482
01:04:20,140 -->
01:04:24,392
I will probably just just literally type
the few words I might see
1483
01:04:24,392 -->
01:04:27,625
in Claude because it feels like I don't
know.
1484
01:04:27,625 -->
01:04:28,625
I don't know.
1485
01:04:28,545 -->
01:04:31,749
I d it it's I'm glad that there's a system
for it,
1486
01:04:31,749 -->
01:04:36,125
but I also I don't know, I've just been
thinking about the Hank Green thing that
1487
01:04:36,125 -->
01:04:39,330
we talked about last week. I listened to a
whole Colin and Samir episode about it.
1488
01:04:39,330 -->
01:04:41,042
I'm gonna link it in top five tech
tomorrow,
1489
01:04:41,042 -->
01:04:45,628
but I don't know. I think people getting a
little weird about stuff.
1490
01:04:45,628 -->
01:04:46,628
That's all.
1491
01:04:46,784 -->
01:04:52,988
Yeah, I just think I understand the the
reasoning behind it.
1492
01:04:52,988 -->
01:04:58,051
The the the thing is is it there's a
difference of what is useful to know
1493
01:04:58,051 -->
01:05:03,704
is AI generated. And I I use AI to I use
like ChatGPT and Claude to proof things
1494
01:05:03,704 -->
01:05:06,145
for me. I use it to fact check things for
me.
1495
01:05:06,145 -->
01:05:09,647
But when I do that, like I've written
something over here,
1496
01:05:09,647 -->
01:05:13,450
I put it in and I say fact check this and
it gives me back information.
1497
01:05:13,450 -->
01:05:14,450
I'm not just
1498
01:05:14,624 -->
01:05:16,317
than publishing that. I'm using it.
1499
01:05:16,161 -->
01:05:17,161
Right.
1500
01:05:16,317 -->
01:05:17,317
So like
1501
01:05:16,638 -->
01:05:18,500
that's fine. That's not gonna be
watermarked.
1502
01:05:18,500 -->
01:05:21,736
Right. But it's it's it's just kind of
like it's weird.
1503
01:05:21,736 -->
01:05:23,158
I don't know. I don't know how I feel
about it,
1504
01:05:23,158 -->
01:05:26,684
except for I don't think that the EU
should be determining how people
1505
01:05:26,684 -->
01:05:28,877
use these tools in this in this way.
1506
01:05:29,793 -->
01:05:32,959
Well, yeah, and it's not necessarily how
you use it,
1507
01:05:32,959 -->
01:05:34,621
but how the tool marks itself,
1508
01:05:34,621 -->
01:05:36,425
I guess. So
1509
01:05:36,576 -->
01:05:40,776
Right, but the the tool marking itself is
not distinguishing between
1510
01:05:40,776 -->
01:05:41,776
how you're using it.
1511
01:05:42,444 -->
01:05:45,295
Right. And that's that's I think the core
the core crux of it.
1512
01:05:45,295 -->
01:05:47,697
And I just I I think I'm pretty clear.
1513
01:05:47,697 -->
01:05:50,138
I mean, I talk pretty openly about what I
use AI for,
1514
01:05:50,138 -->
01:05:52,679
like I just said, like giving me title and
description ideas.
1515
01:05:52,679 -->
01:05:58,431
And I don't want I've I do feel like
people's you know,
1516
01:05:58,431 -->
01:06:00,383
the the internet is always polarizing.
1517
01:06:00,383 -->
01:06:03,945
And so and just seeing what happened to
Hank Green,
1518
01:06:03,945 -->
01:06:07,847
it's like, we're gonna put him in this box
because he apparently used Chat
1519
01:06:07,847 -->
01:06:10,478
GPT for something. And I think there was
confusion about how much
1520
01:06:10,478 -->
01:06:11,478
he actually used AI
1521
01:06:11,661 -->
01:06:14,262
to create his content, but people still
jumped on him and it was like,
1522
01:06:14,262 -->
01:06:16,504
Sheesh. And I understand the
consternation.
1523
01:06:16,504 -->
01:06:22,777
Like I totally get if you what if you read
articles that someone
1524
01:06:22,777 -->
01:06:27,459
was writing for months to years and then
found out that seventy percent
1525
01:06:27,459 -->
01:06:28,890
of their writing was AI generated,
1526
01:06:28,890 -->
01:06:30,391
that would feel like a betrayal.
1527
01:06:30,391 -->
01:06:35,074
I totally get that. And I just don't I
don't want to do anything that might
1528
01:06:35,414 -->
01:06:37,095
instigate the feeling of betrayal,
1529
01:06:37,095 -->
01:06:40,737
even if it's not really n like applicable.
1530
01:06:40,737 -->
01:06:41,737
So
1531
01:06:41,497 -->
01:06:44,529
I don't know. I'm just thinking about how
I might change my habits using that stuff
1532
01:06:44,529 -->
01:06:45,759
going forward. And maybe that's a good
thing.
1533
01:06:45,759 -->
01:06:48,741
I don't know. That's claude. Text
watermarking.
1534
01:06:48,741 -->
01:06:51,182
All right. I wanted to I keep mentioning
the Spotify thing.
1535
01:06:51,182 -->
01:06:53,253
This is a great article from Pod News.
1536
01:06:53,253 -->
01:06:57,246
James Cridland, a friend of Metam at many
podcast movement conferences.
1537
01:06:57,246 -->
01:07:01,228
Spotify is testing a feature in its app.
1538
01:07:01,228 -->
01:07:03,499
It's called the Skip Ahead feature,
1539
01:07:03,499 -->
01:07:10,233
available to premium Spotify users that
will allow you to skip ads in podcasts.
1540
01:07:11,005 -->
01:07:15,086
And even skip over things like membership
plugs.
1541
01:07:15,086 -->
01:07:17,387
So like what I just did earlier in this
episode,
1542
01:07:17,387 -->
01:07:21,438
talking about how you could support the
show and you can get an ad free version
1543
01:07:21,438 -->
01:07:24,989
and all that. This tool from Spotify would
skip even that stuff.
1544
01:07:24,989 -->
01:07:29,470
And allowing Spotify users just to like
totally skip over all the things.
1545
01:07:29,470 -->
01:07:34,752
And Spotify is a large player in the
podcast game as far as how many people use
it.
1546
01:07:34,752 -->
01:07:39,673
Like people use YouTube, and there is
YouTube does have like jump ahead.
1547
01:07:40,065 -->
01:07:41,766
As far as like when an ad starts,
1548
01:07:41,766 -->
01:07:45,028
sometimes you'll see the jump ahead button
show up on the player,
1549
01:07:45,028 -->
01:07:49,371
and it's basically skipping the ad because
it is following the behavior
1550
01:07:49,371 -->
01:07:53,464
of all the viewers before it and assuming
that people jump ahead during this part
1551
01:07:53,464 -->
01:07:55,235
and you can jump forward and skip the ad.
1552
01:07:55,235 -->
01:07:59,718
But this feature specifically built in to
skip ads,
1553
01:07:59,718 -->
01:08:02,399
to even skip the membership plugs,
1554
01:08:02,399 -->
01:08:08,834
feels like ~ you must really not care
about how podcasters make money because
this
1555
01:08:08,834 -->
01:08:09,834
is
1556
01:08:09,420 -->
01:08:14,151
Bypassing all of that. And now our
audience on Spotify is very small.
1557
01:08:14,151 -->
01:08:18,582
A majority of our audience is on Apple
Podcasts and then Overcast and then
YouTube.
1558
01:08:18,582 -->
01:08:20,743
So this is not something that would affect
us specifically,
1559
01:08:20,743 -->
01:08:24,584
but there are some shows where Spotify is
the main one,
1560
01:08:24,584 -->
01:08:26,925
especially like more cross platform shows.
1561
01:08:26,925 -->
01:08:29,986
Spotify had the deal exclusive deals for a
long time,
1562
01:08:29,986 -->
01:08:31,486
and so a lot of people moved to Spotify,
1563
01:08:31,486 -->
01:08:35,627
like once the Joe Rogan show was and Call
Her Daddy was on Spotify exclusively
1564
01:08:35,627 -->
01:08:38,678
for a long time. And a a button that just
by
1565
01:08:39,084 -->
01:08:43,513
default just lets you skip all the
advertising and also the membership
please.
1566
01:08:43,513 -->
01:08:48,322
That doesn't feel great and feels a little
bit anti podcast landscape,
1567
01:08:48,322 -->
01:08:50,587
anti podcast creator kind of thing.
1568
01:08:50,587 -->
01:08:51,587
So
1569
01:08:51,168 -->
01:08:52,168
Yeah,
1570
01:08:51,468 -->
01:08:54,761
and I mean, just to be clear, I know it's
like you're listening to guys
1571
01:08:54,781 -->
01:08:57,553
on a podcast talk complaining about
someone trying to make it harder
1572
01:08:57,553 -->
01:09:00,336
to make money podcasting. I understand
that like fine,
1573
01:09:00,336 -->
01:09:02,087
we're not trying to just be navel gazing
here.
1574
01:09:02,087 -->
01:09:05,650
I do think it's like worthwhile just
mentioning that what's there
1575
01:09:05,650 -->
01:09:10,504
is a inherent conflict between what's
great for the user who doesn't want
1576
01:09:10,504 -->
01:09:13,647
to listen to ads and great for the person
making the content that they
1577
01:09:13,647 -->
01:09:15,919
do want to listen to. There is a tension
there.
1578
01:09:15,919 -->
01:09:19,872
And unless every podcast is going to
become paid.
1579
01:09:20,321 -->
01:09:21,812
You only get to listen to it if you pay
for it.
1580
01:09:21,812 -->
01:09:23,973
Ads are going to be a part of that,
1581
01:09:23,973 -->
01:09:26,115
right? And we like we have ads on our
show.
1582
01:09:26,115 -->
01:09:29,097
We don't make a ton of money on on ads.
1583
01:09:29,097 -->
01:09:31,969
That's just the only way to monetize this,
1584
01:09:31,969 -->
01:09:34,941
other than our members. Now, if all of you
listening want to become members,
1585
01:09:34,941 -->
01:09:37,442
we'll just stop doing ads. That's great.
1586
01:09:36,056 -->
01:09:37,470
Right? Sir.
1587
01:09:37,442 -->
01:09:38,442
If if
1588
01:09:37,943 -->
01:09:39,644
we have zero downloads that are not
members,
1589
01:09:39,644 -->
01:09:40,955
there's no reason for us to do ads.
1590
01:09:40,955 -->
01:09:41,955
That's totally fine.
1591
01:09:41,356 -->
01:09:42,356
Right. Right.
1592
01:09:42,506 -->
01:09:47,789
I think that I just think that Spotify
should not be putting itself in that
position
1593
01:09:48,396 -->
01:09:49,396
Yeah.
1594
01:09:48,710 -->
01:09:49,710
to
1595
01:09:48,970 -->
01:09:49,970
do that.
1596
01:09:50,060 -->
01:09:52,795
Correct. That it would be and yeah,
1597
01:09:52,795 -->
01:09:55,300
it would be like as if YouTube would
introduce a feature like this.
1598
01:09:55,300 -->
01:09:56,893
That would be terrible.
1599
01:09:56,982 -->
01:09:58,233
Well, because d and the reason
1600
01:09:58,233 -->
01:10:02,556
I say that is Spotify benefits from we
give Spotify our content for free.
1601
01:10:02,556 -->
01:10:03,556
They don't pay us
1602
01:10:03,041 -->
01:10:04,041
Right.
1603
01:10:03,487 -->
01:10:06,489
to let you listen to this show in their
app.
1604
01:10:06,489 -->
01:10:10,833
We are giving that to them for free and
they are monetizing
1605
01:10:10,833 -->
01:10:15,336
it through subscriptions. And then w the
people who pay them for that subscription,
1606
01:10:15,336 -->
01:10:19,469
they are going to give those people the
ability to skip the thing that monetizes
1607
01:10:19,469 -->
01:10:22,982
this show, but they will not be sharing
any of that subscription with us.
1608
01:10:23,272 -->
01:10:27,933
Exact exactly. That. And that's the
differen like YouTube,
1609
01:10:27,933 -->
01:10:29,094
which we'll talk about in a second.
1610
01:10:29,094 -->
01:10:32,995
Like YouTube premium allows you to not
have the YouTube ads,
1611
01:10:32,995 -->
01:10:36,175
but creators get paid from that.
1612
01:10:36,175 -->
01:10:40,137
And actually get paid a little more from
premium subscribers viewing their content
1613
01:10:40,137 -->
01:10:44,688
than the ads anyway. And like that is how
you do that system right.
1614
01:10:44,688 -->
01:10:48,999
That's how you do it both for the user's
experience and for the creator.
1615
01:10:48,999 -->
01:10:52,610
So that's so that's a thing. But on
1616
01:10:52,790 -->
01:10:56,891
The complete opposite side, Spotify is
also introducing tools that
1617
01:10:56,891 -->
01:11:00,592
are going to label AI music, which is very
necessary.
1618
01:11:00,592 -->
01:11:04,023
And I talked to my oldest son and he'll
use Spotify and he's like,
1619
01:11:04,023 -->
01:11:08,334
Man, I just get these playlists and
there's just a ton of AI generated music
1620
01:11:08,334 -->
01:11:10,555
in there. And like it's very clear it's AI
generated.
1621
01:11:10,555 -->
01:11:12,425
It's just not labeled as such.
1622
01:11:12,425 -->
01:11:16,727
And so Spotify is going to introduce AI
persona labels.
1623
01:11:16,727 -->
01:11:19,847
So if an entire artist only produces
AI-generated music,
1624
01:11:19,847 -->
01:11:21,788
you'll see that AI persona label.
1625
01:11:22,209 -->
01:11:25,110
And also in playlists and on tracks and in
search,
1626
01:11:25,110 -->
01:11:27,811
they're going to label AI generated
content.
1627
01:11:27,811 -->
01:11:32,333
Artists could appeal if somehow it's
labeled incorrectly.
1628
01:11:32,333 -->
01:11:37,675
But for as bad as the the podcast ad
skipping feature would be,
1629
01:11:37,675 -->
01:11:41,396
kudos to Spotify for this feature because
I do think this is needed.
1630
01:11:41,396 -->
01:11:42,747
And like, yeah, let's be clear,
1631
01:11:42,747 -->
01:11:46,888
AI generated. I know that almost feels
antithetical to I was just talking about
1632
01:11:46,888 -->
01:11:49,239
the text watermarking and I have mixed
feelings about that.
1633
01:11:49,239 -->
01:11:51,320
But for this I don't know
1634
01:11:51,776 -->
01:11:54,589
I don't know, I can Yeah.
1635
01:11:51,852 -->
01:11:54,734
Well, there's a there's a there's a degree
here because what
1636
01:11:54,734 -->
01:12:00,319
we're saying is that if you ran your text
through Claude and you asked
1637
01:12:00,319 -->
01:12:03,542
it to correct all your grammar or your
typos or whatever.
1638
01:12:03,542 -->
01:12:06,454
You created all the content, you're just
asking it to do that piece,
1639
01:12:06,454 -->
01:12:08,636
it's gonna get labeled as AI. Right?
1640
01:12:08,342 -->
01:12:09,342
Right.
1641
01:12:08,736 -->
01:12:12,360
That's different than I just had Suno make
a song and I uploaded it
1642
01:12:12,360 -->
01:12:14,241
to Spotify and I'm gonna collect
royalties.
1643
01:12:14,241 -->
01:12:16,372
I don't think those two things are the
same.
1644
01:12:14,998 -->
01:12:17,295
Right. Yes.
1645
01:12:16,372 -->
01:12:17,372
I don't think it's like
1646
01:12:17,604 -->
01:12:20,076
contradictory to s hold both of those
beliefs.
1647
01:12:20,258 -->
01:12:22,069
Thank thank you, Jason. Thank you.
1648
01:12:22,198 -->
01:12:26,549
Right. As opposed to if AI if Spotify was
marking like you're an indie artist
1649
01:12:26,549 -->
01:12:30,551
and you were just using AI to create a
drum loop and they mark that as AI,
1650
01:12:30,551 -->
01:12:32,600
then I feel like that's that's a little
bit different because like,
1651
01:12:32,600 -->
01:12:34,722
no, this is a real person, this is a real
artist,
1652
01:12:34,156 -->
01:12:35,432
Right. Right, right, right.
1653
01:12:34,722 -->
01:12:35,722
they really
1654
01:12:35,633 -->
01:12:41,244
r made this song, they just needed to use
like in logic the whatever
1655
01:12:41,244 -->
01:12:43,425
fe thing that would generate a drum loop
for
1656
01:12:43,382 -->
01:12:44,392
Right, right, right, right, right,
1657
01:12:43,425 -->
01:12:44,696
you. That's I feel
1658
01:12:44,392 -->
01:12:45,392
right.
1659
01:12:44,696 -->
01:12:45,916
like that's different.
1660
01:12:46,175 -->
01:12:47,747
I think it's different too. I think it's
different too.
1661
01:12:47,747 -->
01:12:48,747
You know it when you see it.
1662
01:12:50,475 -->
01:12:51,475
Or hear it.
1663
01:12:51,220 -->
01:12:54,938
or hear it. OpenAI says they had a really
powerful model and it's putting
1664
01:12:54,938 -->
01:12:58,088
the brakes on it because it has critical
cybersecurity capabilities.
1665
01:12:58,088 -->
01:12:59,789
Just wanna point this up because hey
again,
1666
01:12:59,789 -->
01:13:04,995
let's just n let's not. Let's just not
have the velociraptors while stop.
1667
01:13:02,722 -->
01:13:04,525
Stop breeding the Velociraptors.
1668
01:13:04,525 -->
01:13:05,938
What what
1669
01:13:04,995 -->
01:13:05,995
Stop.
1670
01:13:05,938 -->
01:13:08,893
is the one in like Jurassic Park three
that's like bigger than the T Rex?
1671
01:13:08,928 -->
01:13:09,928
It's the
1672
01:13:09,648 -->
01:13:12,650
~ ultra mega don. It's the
1673
01:13:12,192 -->
01:13:13,192
Yeah.
1674
01:13:12,790 -->
01:13:15,711
there's another one too in the the
Jurassic World things.
1675
01:13:15,711 -->
01:13:19,373
Jurassic World, big monster. That's what
I'm googling right now.
1676
01:13:19,373 -->
01:13:25,036
Distortus Rex the Distortus Rex is the one
that has like the weird head in the
1677
01:13:24,650 -->
01:13:26,131
Gigantosaurus
1678
01:13:26,131 -->
01:13:28,192
is the one in Jurassic World Dominion.
1679
01:13:28,192 -->
01:13:30,573
Indominus Rex, that's what I'm thinking
of.
1680
01:13:28,616 -->
01:13:33,398
Okay. Indominus Rex. The Indominus Re
Don't make Don't make that.
1681
01:13:30,573 -->
01:13:33,814
They specifically
1682
01:13:33,814 -->
01:13:36,686
genetically engineered a hybrid designed
to be bigger,
1683
01:13:36,686 -->
01:13:38,677
smarter, and more dangerous than a T-Rex.
1684
01:13:38,677 -->
01:13:40,778
That's literally what OpenAI is doing.
1685
01:13:40,790 -->
01:13:41,981
That is what they're doing. Stop it.
1686
01:13:41,981 -->
01:13:46,708
Stop it. Don't do that. You you you know
the the Bob Newhart stop it skit from SNL?
1687
01:13:46,228 -->
01:13:47,228
We do n
1688
01:13:46,719 -->
01:13:50,017
yeah. But all I know is I don't need an
Indominus Rex in my computer.
1689
01:13:50,113 -->
01:13:52,484
That's right. Stop it. Just stop it.
1690
01:13:52,484 -->
01:13:54,395
And you know that have you seen that skit
though?
1691
01:13:54,395 -->
01:13:55,395
You know what I'm talking
1692
01:13:54,475 -->
01:13:55,475
Yes.
1693
01:13:54,983 -->
01:13:55,983
Yeah, it's pretty, it's pretty.
1694
01:13:55,936 -->
01:13:58,648
It's a good one. Stop it or I'll bury you
alive in a box.
1695
01:13:58,648 -->
01:14:01,409
~ last thing before personal tech.
1696
01:14:01,409 -->
01:14:03,130
That's from the skit, by the way.
1697
01:14:03,130 -->
01:14:09,684
YouTube is increasing the requirements to
become monetized
1698
01:14:09,684 -->
01:14:11,615
or become part of the YouTube partner
program.
1699
01:14:11,615 -->
01:14:17,158
Previously, up until now, and until I
think it's February first
1700
01:14:17,158 -->
01:14:18,158
of twenty twenty seven.
1701
01:14:18,645 -->
01:14:22,036
It's 1,000 subscribers and 4,000 watch
hours.
1702
01:14:22,036 -->
01:14:26,077
They are changing that for long form
videos on YouTube.
1703
01:14:26,077 -->
01:14:28,958
Same subscriber count, 1,000 subscribers,
1704
01:14:28,958 -->
01:14:33,699
but it's doubling the amount of watch
hours to 8,000 in the calendar year
1705
01:14:34,019 -->
01:14:37,980
to become monetized. And if you want to be
monetized specifically on shorts,
1706
01:14:37,980 -->
01:14:44,342
you need twenty million qualified shorts
views in the last ninety days
1707
01:14:44,802 -->
01:14:47,623
to become monetized as a shorts creator.
1708
01:14:48,311 -->
01:14:52,564
Now there's a lot of dust up about this
because if you have ever tried
1709
01:14:52,564 -->
01:14:55,196
to build a YouTube channel and get
monetized,
1710
01:14:55,196 -->
01:14:57,877
the 1,000 subscribers is a challenge,
1711
01:14:57,877 -->
01:15:03,351
but attainable. The 4,000 watch hours
feels like an insurmountable task
1712
01:15:03,351 -->
01:15:07,964
as a creator because you look at that
count day after day and you're like,
1713
01:15:07,964 -->
01:15:12,507
this number barely moves. And I understand
it can be very disheartening.
1714
01:15:12,507 -->
01:15:15,219
And so to double that to 8,000 hours,
1715
01:15:15,219 -->
01:15:17,811
total like I understand.
1716
01:15:18,772 -->
01:15:23,715
I understand that could be very ~
discouraging and maybe even discourage
some
1717
01:15:23,715 -->
01:15:28,337
creators to try. I will also here's the
thing about all of that.
1718
01:15:28,337 -->
01:15:32,640
Whether you are monetized or not,
1719
01:15:32,640 -->
01:15:35,841
if you just make those requirements,
1720
01:15:35,841 -->
01:15:39,663
even the 8,000 hours, you're still making
almost no money.
1721
01:15:39,663 -->
01:15:44,246
As an example, I don't think a lot of
YouTube channels do this.
1722
01:15:44,246 -->
01:15:47,728
I am going to show what this channel
makes.
1723
01:15:48,204 -->
01:15:49,245
The primary technology channel,
1724
01:15:49,245 -->
01:15:50,767
what we make on YouTube in a month.
1725
01:15:50,767 -->
01:15:52,448
Are you ready for that? Are you ready?
1726
01:15:51,978 -->
01:15:53,229
I didn't even know we made anything.
1727
01:15:54,190 -->
01:15:58,425
Well, yeah. I mean, I've basically
included in what I've sent you because
it's it's
1728
01:15:57,568 -->
01:15:59,546
No I'm not I'm not questioning you,
1729
01:15:58,425 -->
01:16:02,479
nothing. I know. But so
1730
01:15:59,546 -->
01:16:01,675
Steven. I'm just like, okay.
1731
01:16:02,619 -->
01:16:04,401
this is a channel that is monetized.
1732
01:16:04,401 -->
01:16:06,543
We get well, I'll just show you.
1733
01:16:06,543 -->
01:16:08,325
This way, I'm gonna share my screen.
1734
01:16:08,325 -->
01:16:09,656
And for those listening, don't worry,
1735
01:16:09,656 -->
01:16:10,656
I'm going to
1736
01:16:10,838 -->
01:16:13,583
And we probably have higher than average
watch hours,
1737
01:16:13,583 -->
01:16:17,510
I would assume, because people probably
watch all fourteen hours of this show
1738
01:16:18,071 -->
01:16:19,113
as opposed to like
1739
01:16:18,454 -->
01:16:19,454
Rights.
1740
01:16:19,113 -->
01:16:21,718
people who only watch the first three
minutes of a YouTube video.
1741
01:16:21,975 -->
01:16:27,355
Correct. So this channel has made $72 in
the last month.
1742
01:16:28,128 -->
01:16:29,825
That's way more than I thought.
1743
01:16:29,825 -->
01:16:31,573
I just kidna
1744
01:16:30,369 -->
01:16:31,369
Seventy two dollars.
1745
01:16:32,100 -->
01:16:37,884
So and that is for in the last month we've
had two point five thousand watch hours.
1746
01:16:37,884 -->
01:16:42,107
So that means we probably make five
thousand watch hours every two months,
1747
01:16:42,107 -->
01:16:46,319
and it would take us over three months it
to get monetized on this
1748
01:16:46,319 -->
01:16:48,851
new eight thousand hour thing.
1749
01:16:49,142 -->
01:16:51,509
But we're two whatever, almost three years
into this.
1750
01:16:51,927 -->
01:16:52,927
Correct.
1751
01:16:52,492 -->
01:16:56,543
So you'd have to be doing it for that long
and then in one of those years
1752
01:16:56,543 -->
01:16:57,543
hit that number.
1753
01:16:57,623 -->
01:17:02,656
Correct. So I understand the
discouragement of it might take longer
1754
01:17:02,656 -->
01:17:04,848
to get monetized from YouTube,
1755
01:17:04,848 -->
01:17:10,031
but in order for YouTube AdSense revenue
to be anything significant,
1756
01:17:10,031 -->
01:17:12,553
you need way more hours anyway.
1757
01:17:12,553 -->
01:17:13,553
Like
1758
01:17:12,673 -->
01:17:13,673
Yeah.
1759
01:17:13,834 -->
01:17:16,496
if you want to actually make a YouTube
channel a thing,
1760
01:17:16,496 -->
01:17:20,609
you need what you just need way more watch
hours for it to amount to any number.
1761
01:17:20,609 -->
01:17:23,771
And so I understand the discouraging side
of that,
1762
01:17:23,771 -->
01:17:26,333
meaning like if you were like if you're
right now,
1763
01:17:26,333 -->
01:17:27,333
you have like
1764
01:17:27,032 -->
01:17:30,474
3,000 watch hours and now you're like
stressing because you want to
1765
01:17:30,474 -->
01:17:33,676
hit it before February. So you get into
the YouTube partner program.
1766
01:17:33,676 -->
01:17:35,937
I totally get that. I was there.
1767
01:17:35,937 -->
01:17:39,018
I was there, like trying to get those
4,000 watch hours.
1768
01:17:39,018 -->
01:17:44,672
But I also made almost no money from
YouTube bec even in after I was monetized,
1769
01:17:44,672 -->
01:17:49,694
because it just takes time. Part of it is
like if I look at any one of
1770
01:17:49,694 -->
01:17:53,666
my videos in the last month, no one video
makes a significant amount of money.
1771
01:17:53,666 -->
01:17:54,666
But
1772
01:17:54,857 -->
01:17:57,928
I have now 1,000 videos uploaded on my
channel,
1773
01:17:57,928 -->
01:18:00,389
which is a wild number. I didn't even
realize it was that many.
1774
01:18:00,389 -->
01:18:02,829
But I have like 1,000 videos uploaded,
1775
01:18:02,829 -->
01:18:07,571
and the collective watch time across all
those videos over the course
1776
01:18:07,571 -->
01:18:09,911
of a month generates some income.
1777
01:18:09,911 -->
01:18:12,772
Still not a ton, still not the main source
of income.
1778
01:18:12,772 -->
01:18:17,193
I still make more from pretty much every
other revenue stream in my business,
1779
01:18:17,193 -->
01:18:20,764
but it's just is it takes time and it
takes a lot of watch hours.
1780
01:18:20,764 -->
01:18:23,265
So totally identify because I was there.
1781
01:18:23,647 -->
01:18:28,040
wanting to get monetized, but I also get
like even if you attain
1782
01:18:28,040 -->
01:18:30,832
the YouTube partner program, you're not
gonna make significant money anytime soon.
1783
01:18:30,832 -->
01:18:33,514
So that's all. I just want to point that
out.
1784
01:18:33,514 -->
01:18:36,356
And but yeah, I think still go for it.
1785
01:18:36,356 -->
01:18:38,567
You know, I think if people are wondering
like is it too late to start
1786
01:18:38,567 -->
01:18:40,298
a YouTube channel? I don't think so.
1787
01:18:40,298 -->
01:18:42,940
You know, if I I started fairly recently,
1788
01:18:42,940 -->
01:18:45,442
like three or four years ago, I started
putting a sort of effort.
1789
01:18:45,442 -->
01:18:46,823
And I still think there's room today.
1790
01:18:46,823 -->
01:18:49,856
I still see like Christian Sealig.
1791
01:18:49,856 -->
01:18:50,856
Are you familiar with him?
1792
01:18:51,190 -->
01:18:52,190
Yes.
1793
01:18:51,687 -->
01:18:52,687
Maker of
1794
01:18:53,527 -->
01:18:56,638
~ the Apollo Reddit app, which is now
defunct because of Reddit,
1795
01:18:56,638 -->
01:19:02,971
but Christian Sealig, he last year he he
releases like
1796
01:19:02,991 -->
01:19:06,292
one video a year, okay, on his YouTube
channel.
1797
01:19:06,292 -->
01:19:09,494
And he knocks it out of the park every
time.
1798
01:19:09,494 -->
01:19:12,595
And so I'll put a link to his YouTube
channel.
1799
01:19:12,595 -->
01:19:15,176
But he released a video last year.
1800
01:19:15,176 -->
01:19:20,238
It's a gaming piece. It's like I built I
made the PC I couldn't buy.
1801
01:19:20,238 -->
01:19:22,209
Short title, here's the thumbnail.
1802
01:19:22,823 -->
01:19:28,627
8.4 million views, and it got him like
over 200,000 subscribers with that one
video,
1803
01:19:28,627 -->
01:19:33,210
and it is an excellent video, it is
incredibly well done.
1804
01:19:33,210 -->
01:19:36,011
He is an engaging narrator and
storyteller.
1805
01:19:36,011 -->
01:19:37,953
Not once is his face on screen,
1806
01:19:37,953 -->
01:19:39,634
it is a faceless YouTube video,
1807
01:19:39,634 -->
01:19:41,935
but his visuals are super fun and
interesting,
1808
01:19:41,935 -->
01:19:47,459
and so you can totally go viral on a
long-form YouTube video
1809
01:19:47,899 -->
01:19:52,362
and probably make that 8,000 watch hours
and 1,000 subscribers.
1810
01:19:52,562 -->
01:19:54,434
In one video. Like it's very possible.
1811
01:19:54,434 -->
01:19:56,156
He did it. Like Christian Sealy.
1812
01:19:55,710 -->
01:19:56,710
I mean it
1813
01:19:56,172 -->
01:19:59,848
did help that he had an incredibly popular
app that people loved.
1814
01:19:59,848 -->
01:20:01,173
In a name right I'm just
1815
01:20:01,396 -->
01:20:06,121
Yes, but that does not g like none of that
matters on YouTube,
1816
01:20:06,121 -->
01:20:07,615
honestly. Because if you
1817
01:20:07,146 -->
01:20:08,146
You don't think
1818
01:20:07,827 -->
01:20:11,613
him being able to post on to his hundreds
of thousands of followers
1819
01:20:11,613 -->
01:20:13,555
on Twitter and threads that I made a
video,
1820
01:20:13,555 -->
01:20:14,777
there's no carryover?
1821
01:20:15,498 -->
01:20:19,726
His Goodbye Apollo video got thirty eight
thousand views,
1822
01:20:19,726 -->
01:20:23,313
which is not bad, but also nothing.
1823
01:20:23,313 -->
01:20:25,378
F five years ago,
1824
01:20:23,490 -->
01:20:25,162
Sure. Okay. I mean and by the way,
1825
01:20:25,162 -->
01:20:26,162
I
1826
01:20:25,555 -->
01:20:26,778
I still have Apollo on my phone,
1827
01:20:26,778 -->
01:20:28,196
I just want you to know I won't delete it.
1828
01:20:27,974 -->
01:20:28,974
Sure
1829
01:20:28,214 -->
01:20:31,046
I'm sure. But like he posted lots of
videos before this.
1830
01:20:31,046 -->
01:20:32,997
Like he has an M1 MacBook Air video,
1831
01:20:32,997 -->
01:20:39,441
six thousand views. So that's the weird
thing about YouTube is like just because
1832
01:20:39,642 -->
01:20:42,764
you have tons of followers elsewhere
doesn't necessarily mean they're gonna
come
1833
01:20:42,764 -->
01:20:47,487
over and watch. And this gaming like this
gaming video has nothing to
1834
01:20:47,487 -->
01:20:50,329
do with Reddit or Apollo or the apps he
made.
1835
01:20:50,329 -->
01:20:52,290
He's just talking about building a PC.
1836
01:20:53,054 -->
01:20:56,416
And he did like great work. He like 3D
printed the casing and
1837
01:20:56,416 -->
01:21:00,229
he like customized pieces. Like it is a tr
incredible video.
1838
01:21:00,229 -->
01:21:02,821
Totally makes sense why. And he did it
again.
1839
01:21:02,821 -->
01:21:05,593
He just released or or or two years ago,
1840
01:21:05,593 -->
01:21:09,615
he released a video about I built the my
dream keyboard from scratch.
1841
01:21:09,615 -->
01:21:11,146
That video got two point four million
views.
1842
01:21:11,146 -->
01:21:13,841
So all I'm saying is even today,
1843
01:21:13,841 -->
01:21:16,670
a year ago, Christian Sealag did this with
one video.
1844
01:21:16,670 -->
01:21:20,573
And so part of one of the things as if you
want to be a creator,
1845
01:21:20,573 -->
01:21:22,034
you have to reconcile with is like
1846
01:21:22,955 -->
01:21:25,677
Do I have the chops? Do I have the ideas?
1847
01:21:25,677 -->
01:21:28,259
And do I have the execution to be able to
do this?
1848
01:21:28,259 -->
01:21:31,101
And if those three things are true,
1849
01:21:31,101 -->
01:21:35,395
you can still do it today. And even if the
watch time requirement is
1850
01:21:35,395 -->
01:21:38,307
to 8,000 hours versus 4,000, you can still
do it today.
1851
01:21:38,307 -->
01:21:40,809
That's all I'm saying. So be encouraged.
1852
01:21:40,809 -->
01:21:44,302
That's all. Okay, my my app,
1853
01:21:44,302 -->
01:21:47,805
Jason. Personal tech is app cornered this
week.
1854
01:21:48,148 -->
01:21:49,148
It
1855
01:21:48,269 -->
01:21:50,046
just says in our in our rundown,
1856
01:21:50,046 -->
01:21:51,753
Knightmare.
1857
01:21:53,355 -->
01:21:54,846
I did that because I thought it was a fun
plan.
1858
01:21:54,846 -->
01:21:57,979
It's not been a nightmare. Thank you all
for supporting Can Do,
1859
01:21:57,979 -->
01:22:02,928
my task management app. It hit number one
in the Mac App Store and iOS App Store.
1860
01:22:02,928 -->
01:22:04,704
I got a lot of positive feedback,
1861
01:22:04,704 -->
01:22:07,707
t a lot of downloads. Really appreciate
it.
1862
01:22:07,707 -->
01:22:11,119
I just wanted to share a couple things
about the development process that
1863
01:22:11,119 -->
01:22:12,831
I feel like you will identify with.
1864
01:22:12,831 -->
01:22:19,176
And ~ yeah. So one, I've received infinite
feature requests
1865
01:22:19,176 -->
01:22:21,318
and bug fix reports, which
1866
01:22:21,824 -->
01:22:24,166
Totally understand. Thank you all.
1867
01:22:24,166 -->
01:22:27,388
I'm implementing some. There's just so
many feature requests.
1868
01:22:27,388 -->
01:22:28,799
There's no way I would be able to build
them all.
1869
01:22:28,799 -->
01:22:31,551
Maybe if I have Codex and Cloud working on
it simultaneously.
1870
01:22:31,551 -->
01:22:35,313
We'll see. But probably not. But I just
wanted to talk a little bit about
1871
01:22:35,313 -->
01:22:40,307
the actual troubleshooting process because
there was one night,
1872
01:22:40,307 -->
01:22:42,508
Jason. It was a dark night of the soul for
me.
1873
01:22:42,508 -->
01:22:48,543
And that was last Friday. I was receiving
emails
1874
01:22:48,543 -->
01:22:50,514
and bug reports that iCloud sync.
1875
01:22:50,996 -->
01:22:54,747
was not working properly. And I would
test.
1876
01:22:54,747 -->
01:22:56,728
Sometimes it would work, sometimes it
wouldn't.
1877
01:22:56,728 -->
01:22:58,128
So I knew there were problems.
1878
01:22:58,128 -->
01:23:01,239
And I will reiterate again, I'm not a
developer,
1879
01:23:01,239 -->
01:23:04,050
so I can't read the code. I don't know
what it's doing.
1880
01:23:04,050 -->
01:23:08,471
And iCloud also feels like a black box of
what's happening here.
1881
01:23:08,471 -->
01:23:12,772
So I was up till 3 a.m.
1882
01:23:12,772 -->
01:23:16,053
Friday, just going back and forth with
Claude,
1883
01:23:16,053 -->
01:23:18,654
Fable, burning tokens all night long,
1884
01:23:18,654 -->
01:23:20,224
saying
1885
01:23:20,939 -->
01:23:24,060
Here's what's happening. The sync is not
working.
1886
01:23:24,060 -->
01:23:28,941
I'll send screenshots. I'll send screen
recordings to Claude of the various
devices.
1887
01:23:28,941 -->
01:23:30,781
It's running stuff on the Mac.
1888
01:23:30,781 -->
01:23:34,922
And I will send it to Claude. Sometimes it
would work for 20 minutes.
1889
01:23:34,922 -->
01:23:38,563
It would just go. And I'm like,
1890
01:23:38,563 -->
01:23:41,064
I see the token count going. And I'm just
sitting here.
1891
01:23:41,064 -->
01:23:44,845
I'm just staring at it because it's two in
the morning and I have nothing else
1892
01:23:44,845 -->
01:23:47,075
to do. And I'm like watching it go.
1893
01:23:47,075 -->
01:23:49,146
And then it'll come back, Jason.
1894
01:23:49,146 -->
01:23:50,186
And Claude will be like,
1895
01:23:51,156 -->
01:23:53,107
I got it. This was the problem.
1896
01:23:53,107 -->
01:23:55,869
This was a ~ issue with the sync engine.
1897
01:23:55,869 -->
01:23:58,430
This was a thing with the app writing the
boards.
1898
01:23:58,430 -->
01:24:01,732
Fixed it all. Committed the build to
GitHub.
1899
01:24:01,732 -->
01:24:04,633
You're ready to deploy. I said,
1900
01:24:04,633 -->
01:24:06,834
great, Claude. That sounds good.
1901
01:24:06,834 -->
01:24:11,457
So I go in, I go straight to a test flight
build.
1902
01:24:11,457 -->
01:24:14,819
I understand you can deploy the app
directly on your iPhone and Mac.
1903
01:24:14,819 -->
01:24:18,263
I do that as well to test. But with the
iCloud Sync specifically,
1904
01:24:18,263 -->
01:24:20,572
I wanted to test it on the test flight.
1905
01:24:20,936 -->
01:24:25,027
install because that uses the production
iCloud sync,
1906
01:24:25,027 -->
01:24:29,629
not the developer iCloud Sync paradigm in
iCloud Developer Console.
1907
01:24:29,629 -->
01:24:33,870
Anyway, so I would wait for it to hit my
personal test flight.
1908
01:24:33,870 -->
01:24:35,480
I would update my Mac, my iPhone,
1909
01:24:35,480 -->
01:24:37,491
my iPad, I would try it again,
1910
01:24:37,491 -->
01:24:40,112
and inevitably it would still be broken.
1911
01:24:40,112 -->
01:24:43,893
And I would the same process, it was like
Sisyphus rolling up
1912
01:24:43,893 -->
01:24:46,834
the iCloud the mountain, like just trying
to figure this out.
1913
01:24:46,834 -->
01:24:47,904
And it was
1914
01:24:48,744 -->
01:24:52,295
It boggled my mind how often Claude just
is gaslighting me.
1915
01:24:52,295 -->
01:24:55,967
And I understand like it is thought it
found the problem and fixed
1916
01:24:55,967 -->
01:24:58,098
the problem and is very confident.
1917
01:24:58,098 -->
01:25:02,230
I finally understood your take on like
Claude just being the overconfident,
1918
01:25:02,230 -->
01:25:06,981
whatever. And it would probably a dozen or
more times that night,
1919
01:25:06,981 -->
01:25:10,143
it'd be the same process. I would tell
Claude the issue be like,
1920
01:25:10,143 -->
01:25:13,604
~ I missed this one thing, it's fixed now,
1921
01:25:13,604 -->
01:25:16,705
no problem. I would do the whole test
flight process,
1922
01:25:16,705 -->
01:25:18,066
still broken.
1923
01:25:18,943 -->
01:25:21,974
And it was maddening. And at one point,
1924
01:25:21,974 -->
01:25:26,185
Jason, I almost lost it. Because it was
about 2 a.m.
1925
01:25:26,185 -->
01:25:30,086
And I sent the thing to Claude.
1926
01:25:30,086 -->
01:25:31,275
It was thinking. And it came back.
1927
01:25:31,275 -->
01:25:33,867
It said, We've done a lot of good work
today.
1928
01:25:33,867 -->
01:25:36,048
I think I just fixed the iCloud bug.
1929
01:25:36,048 -->
01:25:39,809
It's it's almost 2 A.m. Let's call it a
night.
1930
01:25:39,809 -->
01:25:41,890
This is what Claude this is what Claude
told me.
1931
01:25:41,890 -->
01:25:42,890
Like, this is what Claude told me.
1932
01:25:42,850 -->
01:25:45,941
And I I was like.
1933
01:25:46,443 -->
01:25:48,767
I couldn't whisper flow 'cause it was two
in the morning and my office
1934
01:25:48,767 -->
01:25:53,625
is close to my kids. But just imagine me
typing furiously Don't tell me what to do,
1935
01:25:53,625 -->
01:25:56,360
Claude. You don't tell me it's late.
1936
01:25:56,360 -->
01:25:58,904
Okay? Don't tell me what time it is.
1937
01:25:58,904 -->
01:25:59,956
I know what time it is.
1938
01:25:59,990 -->
01:26:02,320
Time is a human construct and you are a
robot,
1939
01:26:02,320 -->
01:26:03,404
so you just stop.
1940
01:26:03,774 -->
01:26:05,544
And I d I didn't say all that to Claude,
1941
01:26:05,544 -->
01:26:09,255
but I did say, Claude, don't make any
judgment calls about the time
1942
01:26:09,255 -->
01:26:12,096
I'm working or what don't don't tell me
when we're done.
1943
01:26:12,096 -->
01:26:14,137
I tell you when you're done, okay?
1944
01:26:14,137 -->
01:26:16,467
And like it almost sent me over the edge,
1945
01:26:16,467 -->
01:26:19,938
I'll be honest. Kept working, thought it
was fixed.
1946
01:26:19,938 -->
01:26:24,620
It's mostly fixed. I'm still been fixing
issues this entire week.
1947
01:26:24,620 -->
01:26:26,900
And I'm at a point now where I'm also
installing Codex.
1948
01:26:26,900 -->
01:26:29,221
I'm gonna see may let's let's put it to
the test.
1949
01:26:29,221 -->
01:26:31,181
Let's see clo if codex can fix this.
1950
01:26:31,181 -->
01:26:33,312
I did get a very nice email.
1951
01:26:33,642 -->
01:26:39,085
From Phil. He sent me a markdown file with
instructions for LLMs
1952
01:26:39,165 -->
01:26:41,286
on building iCloudSync into the app.
1953
01:26:41,286 -->
01:26:44,818
I gave that markdown file to Claude and I
said,
1954
01:26:44,818 -->
01:26:46,990
please compare this to what you've been
doing and see if there's
1955
01:26:46,990 -->
01:26:49,221
any improvements here. And once I clan,
1956
01:26:49,221 -->
01:26:50,682
I don't know if Claude's gaslighting me or
not,
1957
01:26:50,682 -->
01:26:56,195
but it's like, this part I totally didn't
know about this part about pushing
1958
01:26:56,655 -->
01:26:58,276
something when changes are made.
1959
01:26:58,276 -->
01:27:01,838
I got it now. When it made a bunch of
changes.
1960
01:27:02,376 -->
01:27:03,970
Again, still not for totally fixed,
1961
01:27:03,970 -->
01:27:06,015
but my thank you to Phil. It did help.
1962
01:27:06,015 -->
01:27:08,261
It is more reliable. Still working on the
fix.
1963
01:27:08,261 -->
01:27:11,710
But man, I almost I almost lost it on
Claude Jason.
1964
01:27:12,342 -->
01:27:15,825
There was two two things that happened for
me recently that are similar
1965
01:27:15,982 -->
01:27:17,947
d in building this second app.
1966
01:27:17,947 -->
01:27:19,558
Actually I guess it's the third app,
1967
01:27:19,558 -->
01:27:23,331
but the sec the first there's one that
won't you won't no one'll see it,
1968
01:27:23,331 -->
01:27:28,674
but it's fine. the I s I got to a point
where I was basically the
1969
01:27:28,674 -->
01:27:31,398
to the token window was like nine hundred
thousand tokens and the tok
1970
01:27:31,398 -->
01:27:33,089
the the limit is a million and I was like,
1971
01:27:33,089 -->
01:27:35,221
Okay, simple problem to solve.
1972
01:27:35,221 -->
01:27:36,382
I'm just gonna start a new chat.
1973
01:27:36,382 -->
01:27:39,344
Right? I don't so I said I said,
1974
01:27:39,344 -->
01:27:40,625
Great, you fixed that problem.
1975
01:27:40,625 -->
01:27:41,625
Awesome.
1976
01:27:41,847 -->
01:27:44,918
I'm gonna start a I'm gonna start over in
a new chat because the context window
1977
01:27:44,918 -->
01:27:47,318
is almost full. What info do I need to
bring with me?
1978
01:27:47,318 -->
01:27:49,649
So it'll make a claude.markdown file for
1979
01:27:49,300 -->
01:27:50,300
Right.
1980
01:27:49,649 -->
01:27:52,810
you and it puts it in the project so that
the next chat can just see
1981
01:27:52,810 -->
01:27:55,850
all that kind of stuff. And its response
was,
1982
01:27:55,850 -->
01:27:58,911
Congratulations on shipping exclamation
point.
1983
01:27:58,911 -->
01:28:02,092
I'm like, dude, no,
1984
01:28:02,092 -->
01:28:05,823
we didn't and then it goes through this
whole list of here are your
accomplishments.
1985
01:28:05,823 -->
01:28:08,533
You did this. I'm like, dude, we've only
been working on this for four hours.
1986
01:28:08,533 -->
01:28:10,594
Like, we are nowhere near shipping.
1987
01:28:10,594 -->
01:28:11,594
I just
1988
01:28:11,848 -->
01:28:14,700
I just am telling you like we're gonna
keep this conversation going,
1989
01:28:14,700 -->
01:28:17,062
but I need to start a new chat and I just
need to
1990
01:28:16,660 -->
01:28:17,660
Yes.
1991
01:28:17,923 -->
01:28:23,757
s succinctly tell what it what it has
because it's like it it has no no idea.
1992
01:28:23,757 -->
01:28:29,611
It's just completely crazy. ~ and then it
was the number
1993
01:28:29,611 -->
01:28:31,813
of times where just like you're saying,
1994
01:28:31,813 -->
01:28:36,996
it will it will tell you that it's figured
something out.
1995
01:28:36,996 -->
01:28:39,178
And it goes, I guarantee it's fixed.
1996
01:28:39,562 -->
01:28:42,764
And I just take a screenshot of of what it
thinks is fixed
1997
01:28:42,089 -->
01:28:43,089
Ha ha.
1998
01:28:43,165 -->
01:28:45,036
and I send it and I go, This is ugly,
1999
01:28:45,036 -->
01:28:47,448
bro. And it's like, Yeah, fair.
2000
01:28:47,448 -->
01:28:50,380
It's pretty ugly. And it's like it's
2001
01:28:49,341 -->
01:28:50,511
See. And that that
2002
01:28:50,511 -->
01:28:53,963
was another time where after I gave it
that markdown file about iCloud Sync,
2003
01:28:53,963 -->
01:28:57,985
it said, Yeah, this is definitely helps.
2004
01:28:57,985 -->
01:28:59,605
We should definitely implement it.
2005
01:28:59,605 -->
01:29:03,667
But it's pretty late. We can implement
this in the next iteration.
2006
01:29:03,667 -->
01:29:05,668
I think you should ship it as is.
2007
01:29:05,668 -->
01:29:08,889
That's what it told me. And and to add
insult to injury,
2008
01:29:08,889 -->
01:29:12,410
it put a little cruise ship emoji in its
message.
2009
01:29:12,410 -->
01:29:16,212
I kid you not. I kid you not. And when I
saw that again,
2010
01:29:16,212 -->
01:29:18,293
that cruise ship emoji, I almost
2011
01:29:18,535 -->
01:29:21,388
I almost flipped a table. It was I was
like,
2012
01:29:21,388 -->
01:29:25,352
No, I'm not gonna ship if there are still
more improvements that need to be done.
2013
01:29:25,352 -->
01:29:28,425
Like why are you tired of working?
2014
01:29:28,425 -->
01:29:32,549
Like are you are you saying like so
anyway,
2015
01:29:32,549 -->
01:29:34,491
that was that was just maddening and I
couldn't deal.
2016
01:29:34,039 -->
01:29:35,039
Well, and there
2017
01:29:34,539 -->
01:29:37,742
was a point where it was having a hard
time solving a particular problem.
2018
01:29:37,742 -->
01:29:40,864
And I'm thinking, You're Fable Five,
2019
01:29:40,864 -->
01:29:43,967
the most capable LLM in the world for
doing this.
2020
01:29:43,967 -->
01:29:47,229
Figure it out. And it did.
2021
01:29:48,200 -->
01:29:54,598
It did. Okay, well good. Now you have not
you have not like been tempted
2022
01:29:54,598 -->
01:29:56,431
to try Codex for anything yet?
2023
01:29:56,938 -->
01:30:03,291
Not on this app, no. I mean I've used it
in X code for contextually this
2024
01:30:03,291 -->
01:30:05,331
was before Fable exist existed,
2025
01:30:05,331 -->
01:30:08,283
but when there were s issues that I felt
like it just kept going around
2026
01:30:08,283 -->
01:30:11,224
in circles and was not getting any new
pers and part of that could be because
2027
01:30:11,224 -->
01:30:13,004
I was sending the whole context window
back,
2028
01:30:13,004 -->
01:30:16,126
so it's just continuing on the same train
of thought or thread.
2029
01:30:16,126 -->
01:30:18,887
So there was times when I would just bump
into X code,
2030
01:30:18,887 -->
01:30:21,858
go over to the AI chat thing, load codecs
there and be like,
2031
01:30:21,858 -->
01:30:22,909
Could you look at this problem?
2032
01:30:22,909 -->
01:30:24,489
And it'd be like, Yeah, it looks like it's
probably this,
2033
01:30:24,489 -->
01:30:25,770
this, and this, and be like, Sweet.
2034
01:30:25,770 -->
01:30:26,770
And then
2035
01:30:26,634 -->
01:30:29,146
I would get to a point where that couldn't
figure it out anymore and
2036
01:30:29,146 -->
01:30:30,179
I'd just switch back to Claude.
2037
01:30:30,179 -->
01:30:32,102
But mo this entire app has been Claude.
2038
01:30:32,605 -->
01:30:35,536
Okay. And mine has been clawed up until
now completely.
2039
01:30:35,536 -->
01:30:38,218
I am going to share I just want to show
you this conversation.
2040
01:30:38,218 -->
01:30:40,449
I just need to prove it and I'll put this
as the chapter artwork.
2041
01:30:40,449 -->
01:30:41,570
Look at look at this, look at this.
2042
01:30:41,570 -->
01:30:43,451
Ship it.
2043
01:30:42,082 -->
01:30:45,225
Ship it. Ship this Claude.
2044
01:30:45,992 -->
01:30:48,414
I said, I'm not shipping without these
improvements.
2045
01:30:48,414 -->
01:30:50,775
Make the changes per this document and I
will retest.
2046
01:30:50,775 -->
01:30:55,838
Agreed. Let's do it properly. Like listen.
2047
01:30:55,502 -->
01:30:56,502
Ha ha ha
2048
01:30:56,878 -->
01:30:59,129
Merlin Mann was a guest on MacPower Users
recently.
2049
01:30:59,129 -->
01:31:01,261
He's a very divisive figure. People have
strong feelings.
2050
01:31:01,261 -->
01:31:02,261
We had
2051
01:31:02,171 -->
01:31:03,582
Literally the polar opposite comments.
2052
01:31:03,582 -->
01:31:05,564
Like, this was the best episode in a
while.
2053
01:31:05,564 -->
01:31:07,746
And people are like, I couldn't listen to
five seconds of this episode.
2054
01:31:07,746 -->
01:31:12,129
Totally understand. But the one thing from
that episode that Merlin gets into
2055
01:31:12,250 -->
01:31:17,934
is all the instructions that he has built
into his Claude and of like MD file
2056
01:31:18,295 -->
01:31:21,017
on how to talk. I understand it now.
2057
01:31:21,017 -->
01:31:26,422
I underst I understand why you would be
that particular because sometimes
2058
01:31:26,422 -->
01:31:29,224
it is absolutely maddening, especially
when you're trying to fix a problem.
2059
01:31:29,224 -->
01:31:30,355
And I understand
2060
01:31:30,730 -->
01:31:32,461
People are gonna say, like, this is a vibe
coded app,
2061
01:31:32,461 -->
01:31:35,114
you're not fixing anything. Sure,
2062
01:31:35,114 -->
01:31:38,516
I guess, but I am part of this testing
process.
2063
01:31:38,516 -->
01:31:41,139
Like, I am the one like dealing
2064
01:31:39,146 -->
01:31:40,146
Yeah.
2065
01:31:41,139 -->
01:31:43,891
with data. I'm using the can do app for my
own video planning.
2066
01:31:43,891 -->
01:31:46,803
And so as stuff is breaking, like my own
data is breaking.
2067
01:31:46,803 -->
01:31:50,016
And so, like, this is like we're in it
together,
2068
01:31:50,016 -->
01:31:52,828
whether or not I'm actually writing the
code with my fingers or not.
2069
01:31:52,828 -->
01:31:55,630
Like, this is it is a whatever process.
2070
01:31:55,630 -->
01:31:58,483
And like it was the way Claude talks
sometimes just makes me insane.
2071
01:31:58,483 -->
01:31:59,483
So.
2072
01:31:58,604 -->
01:31:59,604
Well,
2073
01:31:58,844 -->
01:32:02,276
and here two things. We I wanna have this
conversation at some point about
2074
01:32:02,636 -->
01:32:05,247
the stigma ~ the perception of vibe coded
apps,
2075
01:32:05,247 -->
01:32:07,158
whether they're less valuable because of
that.
2076
01:32:07,158 -->
01:32:10,569
But I I really like the way Mark Armin was
talking about it recently when
2077
01:32:10,569 -->
01:32:13,631
he was talking about the number of things
he was able to get fixed in
2078
01:32:13,631 -->
01:32:17,102
his app and like how he spun up some new
stuff
2079
01:32:17,161 -->
01:32:18,161
Yeah.
2080
01:32:17,563 -->
01:32:21,254
entirely based and he was thinking of
himself as a director of AI,
2081
01:32:21,254 -->
01:32:22,645
and I actually think that's super useful.
2082
01:32:22,645 -->
01:32:25,066
That's the first thing. Second thing is
2083
01:32:26,120 -->
01:32:32,035
Every single piece of software you're
using today is being written in some part
2084
01:32:32,115 -->
01:32:34,197
at this point by either Codex or Claude.
2085
01:32:34,197 -->
01:32:37,900
Every single thing, whether it's coming
from the largest,
2086
01:32:37,900 -->
01:32:40,623
especially when it's coming from the
largest companies in the world.
2087
01:32:40,623 -->
01:32:41,623
Now,
2088
01:32:41,041 -->
01:32:42,041
yeah.
2089
01:32:41,684 -->
01:32:43,565
they didn't like have one guy sit down and
be like,
2090
01:32:43,565 -->
01:32:45,367
make me a Kanban board app, right?
2091
01:32:45,367 -->
01:32:47,179
That's not but that's what we think of as
vibe coded,
2092
01:32:47,179 -->
01:32:51,292
but these are the tools that are being
used by the smartest developers in the
world.
2093
01:32:51,326 -->
01:32:52,326
Yeah, yeah.
2094
01:32:51,693 -->
01:32:53,334
And they are just very good directors.
2095
01:32:54,129 -->
01:32:56,231
Exactly. And that is what yeah,
2096
01:32:56,231 -->
01:32:58,473
that's what it is. So I just want to get
that off my chest.
2097
01:32:58,473 -->
01:33:01,395
I thought hopefully that was entertaining
and ship it.
2098
01:33:01,395 -->
01:33:04,538
Listen, Claude. I wanted to ask Jason
about Offred.
2099
01:33:04,538 -->
01:33:05,979
So we're gonna go record a bonus episode.
2100
01:33:05,979 -->
01:33:07,150
You can get the unedited feed,
2101
01:33:07,150 -->
01:33:10,452
which I'm sure is over two hours this week
and usually is close to two hours.
2102
01:33:10,452 -->
01:33:12,194
You can listen to that, our daily show,
2103
01:33:12,194 -->
01:33:14,466
get an ad-free version, click that link
down in the show notes.
2104
01:33:14,466 -->
01:33:16,999
We'd appreciate a five-star rating and
review in Apple Podcast.
2105
01:33:16,999 -->
01:33:20,572
We are extremely close to ~ five hundred
ratings there.
2106
01:33:20,572 -->
01:33:21,572
I'm gonna check it real quick.
2107
01:33:21,422 -->
01:33:22,422
We are at
2108
01:33:22,527 -->
01:33:26,488
494. So we only need six more five-star
ratings.
2109
01:33:26,488 -->
01:33:27,509
So if you haven't done that yet,
2110
01:33:27,509 -->
01:33:30,193
head over to Apple Podcasts, give us a
five-star rating and review there.
2111
01:33:30,193 -->
01:33:32,224
Appreciate it. And thanks for tuning in.
2112
01:33:32,224 -->
01:33:33,325
We'll catch you next time.
