Techgrumps 3.18 – Down The Stack

1
00:00:00,000 --> 00:00:12,640
Welcome to Techgrumps, Techgrumps 3.18.
2
00:00:12,640 --> 00:00:19,960
And today on Sunday, the 4th of August, we have myself, Ian Forrester.
3
00:00:19,960 --> 00:00:22,160
I go under Cubic Garden pretty much everywhere.
4
00:00:22,160 --> 00:00:26,800
And I'm going to pass to Wendy Grossman.
5
00:00:26,800 --> 00:00:28,960
I'm a freelance writer based in London.
6
00:00:28,960 --> 00:00:32,640
I write mostly about computers, freedom, and privacy.
7
00:00:32,640 --> 00:00:34,480
I do a weekly blog called NetWars.
8
00:00:34,480 --> 00:00:39,880
And you'll find me on Mastodon and other fine social media platforms.
9
00:00:39,880 --> 00:00:42,920
And I'm going to pass to David Eastland.
10
00:00:42,920 --> 00:00:43,880
Hello.
11
00:00:43,880 --> 00:00:49,480
Right now, I seem to write a lot for the text, the new stack.io, as it's called.
12
00:00:49,480 --> 00:00:52,680
As long as I also am writing a game.
13
00:00:52,680 --> 00:00:55,560
But I should give the game a name now, Caravan Sarai.
14
00:00:55,560 --> 00:00:57,680
But it won't come out for whenever.
15
00:00:57,680 --> 00:01:01,960
And you can probably find me still on the terrible ex-Twitter.
16
00:01:01,960 --> 00:01:04,800
I think I'm David J Eastman 1.
17
00:01:04,800 --> 00:01:08,560
But the worst Twitter gets, I'm slightly embarrassed
18
00:01:08,560 --> 00:01:11,040
that I'm probably the only one here.
19
00:01:11,040 --> 00:01:12,760
Who is foolish enough to still be on it?
20
00:01:12,760 --> 00:01:15,200
And I'm stuck now.
21
00:01:15,200 --> 00:01:17,800
I still can't believe you're on that platform.
22
00:01:17,800 --> 00:01:20,320
But I'll say let's spot that for now.
23
00:01:20,320 --> 00:01:24,400
And we have also Andy Piper.
24
00:01:24,400 --> 00:01:25,120
Hello.
25
00:01:25,120 --> 00:01:27,640
Yeah, I'm also based in the UK.
26
00:01:27,640 --> 00:01:29,240
I'm a creative technologist.
27
00:01:29,240 --> 00:01:34,240
I do some work with the Mastodon team on the core Mastodon project.
28
00:01:34,240 --> 00:01:37,120
And I've just started blogging again.
29
00:01:37,120 --> 00:01:39,600
I'm 19 days into a blogging streak.
30
00:01:39,600 --> 00:01:40,440
Oh, wow.
31
00:01:40,440 --> 00:01:43,720
My own website, andypiper.co.uk, or my own blog, at least.
32
00:01:43,720 --> 00:01:44,840
It's WordPress.
33
00:01:44,840 --> 00:01:48,120
But yeah, so that's me.
34
00:01:48,120 --> 00:01:53,920
And our first subject is one from Wendy.
35
00:01:53,920 --> 00:01:58,080
And what would tech crumbs be if it wasn't AI?
36
00:01:58,080 --> 00:02:05,200
This is one that I kind of find pretty scary, like most of the AIs.
37
00:02:05,200 --> 00:02:08,040
But Wendy, can you just kind of explain?
38
00:02:08,040 --> 00:02:10,960
Yeah, well, there's sort of two things.
39
00:02:10,960 --> 00:02:14,640
I think the thing that links them is,
40
00:02:14,640 --> 00:02:19,960
that people are finally getting over some of the AI hype.
41
00:02:19,960 --> 00:02:23,640
And they're beginning to sort of think about what generative AI can really
42
00:02:23,640 --> 00:02:25,800
do to help people and what it can't.
43
00:02:25,800 --> 00:02:28,720
And most things it can't.
44
00:02:28,720 --> 00:02:36,080
And there's a great study in, I'm not sure where it was published.
45
00:02:36,080 --> 00:02:40,040
I found it at archive.arxiv.org.
46
00:02:40,040 --> 00:02:46,800
Researchers set up some workshops in Edinburgh with 20 professional
47
00:02:46,800 --> 00:02:50,160
comedians who are used to working live with audiences,
48
00:02:50,160 --> 00:02:54,520
and set them to experiment with AI.
49
00:02:54,520 --> 00:03:04,280
And they concluded that it's not really very good at making jokes.
50
00:03:04,280 --> 00:03:07,200
And they were very, very surprised.
51
00:03:07,200 --> 00:03:10,240
At making jokes.
52
00:03:10,240 --> 00:03:13,360
And they described it as bland.
53
00:03:13,360 --> 00:03:16,280
I like the description of the humor that it came up with.
54
00:03:16,280 --> 00:03:24,160
Bland 1950s cruise ship style jokes, but a little less racist.
55
00:03:24,160 --> 00:03:25,040
Oh, that's good.
56
00:03:25,040 --> 00:03:29,280
I mean, this is what you go to hear a live comedian for, right?
57
00:03:29,280 --> 00:03:32,080
But the other one was an article in The Guardian,
58
00:03:32,080 --> 00:03:35,280
which was really about a study in The Lancet,
59
00:03:35,280 --> 00:03:42,440
sort of arguing that, I think it was Tony Blair who
60
00:03:42,440 --> 00:03:45,960
has started to say that we should put AI in the NHS,
61
00:03:45,960 --> 00:03:48,720
and it'll make it all more efficient, and it'll fix everything.
62
00:03:48,720 --> 00:03:51,040
And the article in The Lancet basically said,
63
00:03:51,040 --> 00:03:53,120
actually it's a distraction.
64
00:03:53,120 --> 00:03:55,800
The things that need to be done don't need AI.
65
00:03:55,800 --> 00:03:59,640
And they came up with the wonderful phrase,
66
00:03:59,640 --> 00:04:03,960
bionic duckweed, to describe the effect of using AI
67
00:04:03,960 --> 00:04:06,640
the way they were talking about.
68
00:04:06,640 --> 00:04:09,400
So I thought that was, I think it's a good sign
69
00:04:09,400 --> 00:04:11,640
that we're beginning to get people sort of recognizing
70
00:04:11,640 --> 00:04:13,720
that a lot of this is just silly.
71
00:04:13,720 --> 00:04:15,920
I want to know more about the duckweed.
72
00:04:15,920 --> 00:04:17,800
Why is it called bionic duckweed?
73
00:04:17,800 --> 00:04:20,000
Sorry, I want to know more.
74
00:04:20,000 --> 00:04:23,200
Hang on a second.
75
00:04:23,200 --> 00:04:24,920
I don't smoke duckweed, should I?
76
00:04:24,920 --> 00:04:26,760
Or have I got this wrong?
77
00:04:26,760 --> 00:04:27,840
Well.
78
00:04:27,840 --> 00:04:29,560
Everything's a drug reference these days.
79
00:04:29,560 --> 00:04:31,640
Well, here's how he describes bionic duckweed.
80
00:04:31,640 --> 00:04:36,640
It was a term coined by Stan Westlake in 2020
81
00:04:37,680 --> 00:04:40,600
to cover the use, deliberately or otherwise,
82
00:04:40,600 --> 00:04:43,680
of technology that may or may not arrive in the future
83
00:04:43,680 --> 00:04:46,240
to argue against investment in the present.
84
00:04:46,240 --> 00:04:48,640
In other words, it kind of chokes off funds
85
00:04:48,640 --> 00:04:50,800
to things that could actually be useful by the way.
86
00:04:50,800 --> 00:04:52,040
Hey, look at me, look at me.
87
00:04:52,040 --> 00:04:53,960
I'm going to be great someday.
88
00:04:53,960 --> 00:04:58,400
And so his example, his favorite example of bionic duckweed
89
00:04:58,400 --> 00:05:02,680
is Elon Musk's Hyperloop, first proposed more than a decade
90
00:05:02,680 --> 00:05:05,720
ago explicitly to try and discourage California
91
00:05:05,720 --> 00:05:07,520
from going ahead with plans to construct
92
00:05:07,520 --> 00:05:09,120
a high-speed rail line.
93
00:05:10,640 --> 00:05:12,520
So it's not just that it's vaporware,
94
00:05:12,520 --> 00:05:14,760
it's that it's obstructive vaporware.
95
00:05:17,040 --> 00:05:19,560
And I don't know if you've ever seen a pond covered
96
00:05:19,560 --> 00:05:22,520
with duckweed where there's actually a pond like that
97
00:05:22,520 --> 00:05:23,600
very near where I live.
98
00:05:23,600 --> 00:05:26,920
And the entire surface is bright green.
99
00:05:26,920 --> 00:05:30,360
And you try to imagine like a minnow trying to swim in it
100
00:05:30,360 --> 00:05:32,440
and it's just impossible.
101
00:05:32,440 --> 00:05:34,280
I think we have a slightly different,
102
00:05:34,280 --> 00:05:35,640
it's definitely called duckweed as well,
103
00:05:35,640 --> 00:05:37,040
but there's another term that gets used
104
00:05:37,040 --> 00:05:38,600
for exactly the same stuff.
105
00:05:38,600 --> 00:05:41,080
So, but now I realize what you're talking about.
106
00:05:41,080 --> 00:05:43,080
I've never tried to smoke that.
107
00:05:43,080 --> 00:05:43,920
No, I wouldn't do that.
108
00:05:43,920 --> 00:05:46,640
Going back to Ian's initial introduction of this piece,
109
00:05:46,640 --> 00:05:49,240
he said, what would tech crumbs be if it wasn't,
110
00:05:49,240 --> 00:05:50,280
you know, if we weren't talking about this?
111
00:05:50,280 --> 00:05:53,400
Well, tech crumbs existed before the current version
112
00:05:53,400 --> 00:05:55,320
of what we're currently referring to as AI.
113
00:05:55,320 --> 00:05:56,240
And we were grumpy then.
114
00:05:56,240 --> 00:06:01,240
So we can continue to be grumpy and thoughtfully,
115
00:06:03,600 --> 00:06:05,840
you know, against some of the things
116
00:06:05,840 --> 00:06:08,280
that are being discussed here.
117
00:06:08,280 --> 00:06:11,000
I think, yes, I do love that description.
118
00:06:11,000 --> 00:06:13,400
Now that we've dug into a little bit more
119
00:06:13,400 --> 00:06:16,440
the bionic duckweed and the idea that this stuff is,
120
00:06:16,440 --> 00:06:19,560
has to be thrown into everything
121
00:06:19,560 --> 00:06:23,280
to potentially divert other things.
122
00:06:24,360 --> 00:06:26,200
I noted that the paper, the research paper
123
00:06:26,200 --> 00:06:29,520
that all of the conversation initially sprang from
124
00:06:29,520 --> 00:06:34,520
was by four Google deepminders doing this work
125
00:06:34,600 --> 00:06:36,400
with professional comedians.
126
00:06:37,840 --> 00:06:40,280
So it'll be interesting based on the fact
127
00:06:40,280 --> 00:06:44,120
that it's research that that company has funded and done
128
00:06:44,120 --> 00:06:47,400
to see whether that has any impact on the way
129
00:06:47,400 --> 00:06:50,280
that they continue to apply AI.
130
00:06:50,280 --> 00:06:53,920
I know that we've had other things
131
00:06:53,920 --> 00:06:55,920
put into our discussion notes here
132
00:06:55,920 --> 00:06:58,200
about an advert, which I'm aware of.
133
00:06:58,200 --> 00:06:59,960
And I discussed in another podcast this week,
134
00:06:59,960 --> 00:07:03,280
but I haven't seen myself because it's not been shown
135
00:07:03,280 --> 00:07:04,640
in the UK in the Olympics.
136
00:07:04,640 --> 00:07:07,760
I gather it's all over the US Olympics coverage.
137
00:07:07,760 --> 00:07:09,920
I think you can find a way of tying that out.
138
00:07:09,920 --> 00:07:11,720
Oh, I'm sure, I'm sure I can.
139
00:07:11,720 --> 00:07:13,640
It's on YouTube.
140
00:07:13,640 --> 00:07:14,640
Yeah, absolutely.
141
00:07:14,640 --> 00:07:18,120
But I, you know, my colleagues and friends
142
00:07:18,120 --> 00:07:20,800
on my other podcast were talking about this Gemini ad
143
00:07:20,800 --> 00:07:23,640
because they see it all the time in the Olympics coverage
144
00:07:23,640 --> 00:07:28,520
when which AI is asked to help a kid write a letter
145
00:07:28,520 --> 00:07:30,000
to an Olympic athlete.
146
00:07:30,000 --> 00:07:31,680
And it is really sad.
147
00:07:31,680 --> 00:07:33,800
Just go, just say that they were the podcast.
148
00:07:33,800 --> 00:07:38,160
Oh, gamesatwork.biz, we recorded that one on Friday.
149
00:07:38,160 --> 00:07:39,680
And I just say that I make a point
150
00:07:39,680 --> 00:07:41,240
of not watching the Olympics.
151
00:07:42,440 --> 00:07:43,280
Really? Yeah.
152
00:07:43,280 --> 00:07:44,320
No, no, let's not go there.
153
00:07:44,320 --> 00:07:45,160
Let's not go there.
154
00:07:45,160 --> 00:07:46,200
Let's not go there.
155
00:07:47,440 --> 00:07:48,280
Let's just leave that.
156
00:07:48,280 --> 00:07:50,760
We can spend like half an hour on that one alone.
157
00:07:50,760 --> 00:07:54,400
The other thing I mentioned is that I may have been targeted
158
00:07:54,400 --> 00:07:57,120
via LinkedIn advert for the Tony Blair Institute
159
00:07:57,120 --> 00:07:59,720
to do work with them around their AI.
160
00:08:01,240 --> 00:08:02,080
Nice.
161
00:08:02,080 --> 00:08:02,920
This week.
162
00:08:04,240 --> 00:08:05,800
I wrote in fact about the fact that I've been
163
00:08:05,800 --> 00:08:07,800
on LinkedIn for 18 years.
164
00:08:07,800 --> 00:08:08,640
Hooray.
165
00:08:08,640 --> 00:08:10,120
Because I got a notification telling me
166
00:08:10,120 --> 00:08:12,000
I've been on LinkedIn.
167
00:08:12,000 --> 00:08:14,880
Is it like one of their work birthdays?
168
00:08:14,880 --> 00:08:16,480
Yeah, I don't care.
169
00:08:18,000 --> 00:08:20,480
So pivoted so far away from what it was.
170
00:08:20,480 --> 00:08:22,320
Anyway, that's a sidetrack.
171
00:08:22,320 --> 00:08:25,240
I was gonna say, there was a piece that I,
172
00:08:25,240 --> 00:08:26,960
I was gonna add to the show notes actually,
173
00:08:26,960 --> 00:08:31,960
but if you're fair, it was a very small study
174
00:08:32,400 --> 00:08:37,160
and it was around smart TVs
175
00:08:37,160 --> 00:08:41,480
and how they're starting to add AI to,
176
00:08:41,480 --> 00:08:44,840
so I was doing air quotes to all of this stuff
177
00:08:44,840 --> 00:08:49,840
and how consumers have been starting to turn off from the
178
00:08:52,280 --> 00:08:55,080
notion of AI in everything.
179
00:08:55,080 --> 00:08:56,360
It's a very small study.
180
00:08:56,360 --> 00:08:57,960
This is why I didn't include it.
181
00:08:57,960 --> 00:09:01,000
Once they get to, I think it's like a hundred consumers,
182
00:09:01,000 --> 00:09:05,400
which is, you know, not really enough at all.
183
00:09:05,400 --> 00:09:07,120
No, but at the same time,
184
00:09:07,120 --> 00:09:10,120
what they're using AI for in some of these things
185
00:09:10,120 --> 00:09:11,320
is just so ridiculous.
186
00:09:11,320 --> 00:09:14,520
I have a new washing machine and it has an AI feature.
187
00:09:14,520 --> 00:09:19,520
And the AI feature is that it orders the menu on the,
188
00:09:20,040 --> 00:09:22,480
it has a dial to choose different programs
189
00:09:22,480 --> 00:09:26,080
and the AI feature orders them in the order
190
00:09:26,080 --> 00:09:28,440
in which you most often use them.
191
00:09:28,440 --> 00:09:31,920
I'm sorry, that's not AI, that's a custom menu.
192
00:09:31,920 --> 00:09:32,760
Can we just call it what it is?
193
00:09:32,760 --> 00:09:33,600
Exactly, it's an algorithm.
194
00:09:33,600 --> 00:09:35,080
It just sorts stuff.
195
00:09:35,080 --> 00:09:37,640
It's not AI, it really, I agree with you Wendy.
196
00:09:37,640 --> 00:09:40,280
Oh, because it rewrites it as a Shakespeare sonnet,
197
00:09:40,280 --> 00:09:41,560
that's new.
198
00:09:41,560 --> 00:09:45,520
That would be interesting,
199
00:09:45,520 --> 00:09:47,880
but I'm not paying a lot of money just for it.
200
00:09:49,360 --> 00:09:50,680
That is, that's ridiculous.
201
00:09:50,680 --> 00:09:52,720
I've, yeah, that's just.
202
00:09:52,720 --> 00:09:55,760
That's just an, it's a customized menu, big deal.
203
00:09:55,760 --> 00:09:56,760
Yeah, yeah.
204
00:09:56,760 --> 00:09:59,320
Or I'm sorry, I want to be American for a second.
205
00:09:59,320 --> 00:10:00,160
Big whoop.
206
00:10:01,480 --> 00:10:03,080
Big whoop, what's that?
207
00:10:03,080 --> 00:10:04,080
Big whoop.
208
00:10:05,960 --> 00:10:09,560
It would have to be English, I don't know what to understand.
209
00:10:09,560 --> 00:10:11,600
I actually, big whoop.
210
00:10:11,600 --> 00:10:12,760
Yeah, there you go.
211
00:10:12,760 --> 00:10:14,720
Okay, okay.
212
00:10:15,880 --> 00:10:17,600
It's completely over my head.
213
00:10:17,600 --> 00:10:22,200
No, I, we talked about this last month, obviously,
214
00:10:22,200 --> 00:10:25,640
me and Wendy, about how it's, you've seen this everywhere.
215
00:10:25,640 --> 00:10:30,640
And it is pretty ridiculous in some cases.
216
00:10:32,440 --> 00:10:36,480
I do think that actually the bionic duckweed,
217
00:10:36,480 --> 00:10:40,280
I've come across it, but I had not heard that explanation.
218
00:10:40,280 --> 00:10:43,000
And I'm going to use that more.
219
00:10:43,000 --> 00:10:44,360
I like that.
220
00:10:44,360 --> 00:10:48,480
I could definitely see that applied in many different places.
221
00:10:49,880 --> 00:10:51,080
So yeah, I like that.
222
00:10:51,080 --> 00:10:52,880
I'm going to make a note of that one.
223
00:10:53,800 --> 00:10:57,120
Any other thoughts on this one before we move on?
224
00:10:58,000 --> 00:10:59,920
David Eastman?
225
00:10:59,920 --> 00:11:01,320
I was just going to say that,
226
00:11:02,440 --> 00:11:03,640
just continuing what Andy said,
227
00:11:03,640 --> 00:11:06,600
I think tech grumps can continue being grumpy.
228
00:11:06,600 --> 00:11:09,360
And I think we should be grumpy or die trying, you know.
229
00:11:09,360 --> 00:11:14,360
I, to fair, yes, I think you're probably right.
230
00:11:17,280 --> 00:11:19,280
I'll leave that at that.
231
00:11:19,280 --> 00:11:23,280
We're channeling the frustration of millions of people
232
00:11:23,280 --> 00:11:24,640
whose computers hate them.
233
00:11:29,120 --> 00:11:32,080
The next story is mine.
234
00:11:32,080 --> 00:11:35,440
And it's around alternative business models
235
00:11:35,440 --> 00:11:38,280
or actually don't really like that term alternative
236
00:11:38,280 --> 00:11:41,000
because they're just another business model.
237
00:11:41,000 --> 00:11:46,000
And this one is focused around a big retailer in the UK.
238
00:11:49,280 --> 00:11:50,800
I'm not sure if they stretch any further,
239
00:11:50,800 --> 00:11:55,800
but in the UK, there's a group called John Lewis.
240
00:11:56,640 --> 00:11:59,200
Now John Lewis is a partnerships.
241
00:11:59,200 --> 00:12:04,200
That means every one of the employees have a say
242
00:12:04,880 --> 00:12:07,760
and sharing the profits of the company.
243
00:12:08,800 --> 00:12:13,800
The piece from the FT basically talks about how
244
00:12:15,600 --> 00:12:20,600
can this model work in a space where we have,
245
00:12:22,640 --> 00:12:24,520
where it's all basically, not all,
246
00:12:24,520 --> 00:12:28,040
but mostly it's stakeholder capitalism,
247
00:12:28,040 --> 00:12:33,040
where someone is, people are paying a lot of money
248
00:12:33,400 --> 00:12:36,280
into companies and they want huge returns
249
00:12:36,280 --> 00:12:38,920
and they want kind of endless returns.
250
00:12:38,920 --> 00:12:41,240
And yeah, it looks like,
251
00:12:41,240 --> 00:12:43,440
so like John Lewis is very quaint
252
00:12:43,440 --> 00:12:47,840
and actually can't survive in this world.
253
00:12:47,840 --> 00:12:51,560
I found it very interesting because as I said before,
254
00:12:51,560 --> 00:12:53,440
I don't think it's an alternative business model.
255
00:12:53,440 --> 00:12:55,360
I think there are many business models.
256
00:12:55,360 --> 00:12:59,400
And it's just that we kind of been, not just,
257
00:12:59,400 --> 00:13:01,440
I know a lot of people that have started to realize,
258
00:13:01,440 --> 00:13:03,680
start to think about the only business model
259
00:13:03,680 --> 00:13:07,840
is either via Angel, via VC or a stakeholder.
260
00:13:07,840 --> 00:13:10,440
They don't feel about any other business models.
261
00:13:10,440 --> 00:13:14,080
And it kind of scares me that we've kind of gone that route.
262
00:13:14,080 --> 00:13:18,840
I think when you say it's not really
263
00:13:18,840 --> 00:13:20,280
an alternative business model,
264
00:13:20,280 --> 00:13:21,480
there's other business model,
265
00:13:21,480 --> 00:13:22,960
there's a variety of business models.
266
00:13:22,960 --> 00:13:24,120
Of course, that's absolutely correct.
267
00:13:24,120 --> 00:13:26,680
I think what you're getting at though,
268
00:13:26,680 --> 00:13:28,240
is that it's an alternative
269
00:13:28,240 --> 00:13:30,720
to the predominant business model,
270
00:13:30,720 --> 00:13:35,440
which is current form late stage capitalism,
271
00:13:35,440 --> 00:13:36,880
which is what you were just describing.
272
00:13:36,880 --> 00:13:38,440
You start a business,
273
00:13:38,440 --> 00:13:41,640
you get some funding from a VC or somewhere,
274
00:13:41,640 --> 00:13:44,560
and then that drives you in the direction
275
00:13:44,560 --> 00:13:48,560
of permanent constant never ending growth.
276
00:13:48,560 --> 00:13:50,720
And in order to reach that,
277
00:13:50,720 --> 00:13:55,200
you must keep expanding and reducing costs
278
00:13:55,200 --> 00:13:56,560
and maximizing profit,
279
00:13:56,560 --> 00:13:59,600
which is the opposite of what John Lewis,
280
00:13:59,600 --> 00:14:01,040
or different, I should say,
281
00:14:01,040 --> 00:14:04,640
to what John Lewis sought to do in the 19,
282
00:14:04,640 --> 00:14:07,400
in the 1800s when the company was first founded,
283
00:14:07,400 --> 00:14:11,040
the son of the founder really recognized that,
284
00:14:12,600 --> 00:14:15,120
that making sufficient profit
285
00:14:15,120 --> 00:14:19,120
rather than maximizing profit was okay, right?
286
00:14:19,120 --> 00:14:20,960
And sharing that profit.
287
00:14:20,960 --> 00:14:24,720
And certainly as I've been through my career
288
00:14:24,720 --> 00:14:28,640
and recent changes in the last few years,
289
00:14:28,640 --> 00:14:31,520
I've very much got to a point of frustration
290
00:14:31,520 --> 00:14:35,080
with the current system that is predominant
291
00:14:35,080 --> 00:14:37,440
and feeling that,
292
00:14:37,440 --> 00:14:40,000
I was talking about it with my wife on a car journey today,
293
00:14:40,000 --> 00:14:43,160
just the feeling that why should companies
294
00:14:43,160 --> 00:14:46,520
constantly feel that they need to grow and expand
295
00:14:46,520 --> 00:14:48,400
if they do what they enjoy,
296
00:14:48,400 --> 00:14:51,240
they are successful, they make their customers happy,
297
00:14:51,240 --> 00:14:53,240
they meet their customers needs at the scale
298
00:14:53,240 --> 00:14:56,800
at which they can manage and the market can manage.
299
00:14:56,800 --> 00:15:00,960
The challenge is, yeah, what the market,
300
00:15:00,960 --> 00:15:03,120
the market doing the air quotes as well
301
00:15:03,120 --> 00:15:06,040
and thinking of it in big words,
302
00:15:07,480 --> 00:15:12,480
tries to, seems to believe is the only way to operate.
303
00:15:12,800 --> 00:15:16,200
There's another model we have in the UK of the cooperative,
304
00:15:16,200 --> 00:15:18,720
which is also a grocery retailer
305
00:15:18,720 --> 00:15:22,080
that's a competitor to John Lewis in some spaces.
306
00:15:22,080 --> 00:15:26,640
And they also have a model that is very different
307
00:15:26,640 --> 00:15:30,000
to the traditional stock market share.
308
00:15:30,000 --> 00:15:32,000
It's some Northern town in England, isn't it?
309
00:15:32,000 --> 00:15:35,120
Is it Rochdale, which did the original co-op?
310
00:15:35,120 --> 00:15:35,960
Yes.
311
00:15:35,960 --> 00:15:37,480
I believe that might be right.
312
00:15:37,480 --> 00:15:39,600
Just outside of Manchester?
313
00:15:39,600 --> 00:15:41,360
You know, the place where it seems to me
314
00:15:41,360 --> 00:15:44,480
that an alternative business model really is important
315
00:15:44,480 --> 00:15:46,440
is in journalism.
316
00:15:46,440 --> 00:15:49,840
I think when you look at the Guardian and the BBC,
317
00:15:49,840 --> 00:15:53,280
which are funded very differently from the corporate media,
318
00:15:53,280 --> 00:15:56,920
you can see the results in how they cover things
319
00:15:56,920 --> 00:15:59,120
and how they approach things.
320
00:16:01,560 --> 00:16:06,560
You know, they don't, they can publish unpopular things,
321
00:16:06,800 --> 00:16:10,640
you know, because that's not how they measure success.
322
00:16:10,640 --> 00:16:12,680
And I mean, I think the Guardian's approach to the web,
323
00:16:12,680 --> 00:16:15,280
for example, where it doesn't have a paywall,
324
00:16:15,280 --> 00:16:17,680
although it's funny how many American friends
325
00:16:17,680 --> 00:16:18,800
think it does have a paywall
326
00:16:18,800 --> 00:16:20,280
because it keeps putting up these big banners
327
00:16:20,280 --> 00:16:21,120
saying, we need to get this money.
328
00:16:21,120 --> 00:16:23,320
They pay us because we don't have a paywall.
329
00:16:23,320 --> 00:16:25,240
Just assume that they can't just click it away
330
00:16:25,240 --> 00:16:27,480
and go on and read the story.
331
00:16:27,480 --> 00:16:30,440
But you know, the fact that the Guardian is open to the web,
332
00:16:30,440 --> 00:16:32,840
it's because their larger mission
333
00:16:32,840 --> 00:16:36,000
is not to be more profitable than anybody else,
334
00:16:36,000 --> 00:16:39,720
but to have more reach and to influence the world.
335
00:16:39,720 --> 00:16:42,000
And that was the decision they made.
336
00:16:42,000 --> 00:16:44,480
And you know, I'm really glad
337
00:16:44,480 --> 00:16:47,360
that there is a publication like that.
338
00:16:47,360 --> 00:16:49,200
And if you look at, you know, so a couple,
339
00:16:49,200 --> 00:16:53,560
there's been a couple of new investigative journalism outlets
340
00:16:53,560 --> 00:16:56,240
in the last 10, 15 years.
341
00:16:56,240 --> 00:16:57,980
In Bristol, there's the Bristol Cable,
342
00:16:57,980 --> 00:17:00,560
and in Scotland, there's the Ferret,
343
00:17:00,560 --> 00:17:04,040
which are basically cooperatives
344
00:17:04,040 --> 00:17:05,560
among the journalists who write for them.
345
00:17:05,560 --> 00:17:07,040
And I think that's great, you know,
346
00:17:07,040 --> 00:17:09,560
and they're doing really wonderful work, both of them.
347
00:17:09,560 --> 00:17:12,240
I should point out, obviously, the Guardian started from day
348
00:17:12,240 --> 00:17:14,600
nought from scratch, it wouldn't be able to do what it's doing.
349
00:17:14,600 --> 00:17:18,360
It has that big Scott bequest, which is doing very well.
350
00:17:18,360 --> 00:17:19,480
I mean, that's what I'm saying.
351
00:17:19,480 --> 00:17:20,800
It's owned by a trust, not.
352
00:17:20,800 --> 00:17:22,280
Yeah, it's using it very well.
353
00:17:22,280 --> 00:17:25,480
Yeah, it's done differently.
354
00:17:25,480 --> 00:17:27,200
Yeah, that's impressive.
355
00:17:27,200 --> 00:17:30,040
I think your point about journalism is really powerful.
356
00:17:30,040 --> 00:17:34,000
And I think, well, even, I know it's not quite the same,
357
00:17:34,000 --> 00:17:38,240
but like 404 Media, it's regularly.
358
00:17:38,240 --> 00:17:39,920
I am so impressed with them.
359
00:17:39,920 --> 00:17:40,640
They're so great.
360
00:17:40,640 --> 00:17:41,920
Well, they've just won an award.
361
00:17:41,920 --> 00:17:46,680
They've just won the EFF's award, haven't they,
362
00:17:46,680 --> 00:17:48,400
for journalism this year?
363
00:17:48,400 --> 00:17:50,240
What, the Pioneer Award?
364
00:17:50,240 --> 00:17:52,240
I'm just going to check exactly which it is.
365
00:17:52,240 --> 00:18:01,440
But I am, EFF Awards, 404 Media, it is for the,
366
00:18:01,440 --> 00:18:05,000
carry on talking, and I will tell you in a moment.
367
00:18:05,000 --> 00:18:10,200
Usually the awards that EFF gives are the Pioneer Awards.
368
00:18:10,200 --> 00:18:11,720
You know what else is really good?
369
00:18:11,720 --> 00:18:15,080
I'm not sure how they're financed is restofworld.org.
370
00:18:15,080 --> 00:18:16,640
Yeah, another one.
371
00:18:16,640 --> 00:18:20,400
Which is doing a great job of sort of looking up story,
372
00:18:20,400 --> 00:18:23,960
running stories that nobody else would even think about.
373
00:18:23,960 --> 00:18:25,760
Yeah, you're absolutely right.
374
00:18:25,760 --> 00:18:30,520
I, so the reason why I put this story in,
375
00:18:30,520 --> 00:18:32,520
just because I completely agree with all this stuff,
376
00:18:32,520 --> 00:18:37,000
is that I'm meeting people who literally think
377
00:18:37,000 --> 00:18:41,160
there's only one way of doing these things now.
378
00:18:41,160 --> 00:18:43,880
And I kind of tell them about, you know,
379
00:18:43,880 --> 00:18:46,560
have you thought about an alternative business model?
380
00:18:46,560 --> 00:18:48,160
And they're going to go, well, no.
381
00:18:48,160 --> 00:18:50,480
And I think actually Andy's point,
382
00:18:50,480 --> 00:18:53,240
when you were explaining probably better than I was
383
00:18:53,240 --> 00:18:55,640
about this story, is that once you
384
00:18:55,640 --> 00:19:01,360
start to take like a tiny bit of money, or you're kind of hooked,
385
00:19:01,360 --> 00:19:05,000
it's like you're on that ladder now,
386
00:19:05,000 --> 00:19:06,640
and there's no way off it almost.
387
00:19:06,640 --> 00:19:11,760
And I feel that that's kind of that predominant model
388
00:19:11,760 --> 00:19:14,960
is being pushed so hard that people are starting to,
389
00:19:14,960 --> 00:19:18,640
maybe it's like the duck water thing, the duck weeds.
390
00:19:18,640 --> 00:19:22,960
You know, yeah, they literally can't see any other way.
391
00:19:22,960 --> 00:19:24,600
Like, it has to be this way.
392
00:19:24,600 --> 00:19:26,360
There's no other way of doing it.
393
00:19:26,360 --> 00:19:27,440
And that scares me.
394
00:19:27,440 --> 00:19:29,600
That really scares me.
395
00:19:29,600 --> 00:19:33,400
Yeah, on journalism, the other one I think of
396
00:19:33,400 --> 00:19:37,640
is ProPublica, which is a nonprofit as well.
397
00:19:37,640 --> 00:19:42,200
Yeah, it frustrates me the more I think about these things.
398
00:19:42,200 --> 00:19:45,600
I do a lot of stuff around the maker community
399
00:19:45,600 --> 00:19:48,760
and meet a lot of folks who are, like myself, starting out
400
00:19:48,760 --> 00:19:50,480
with really tiny businesses.
401
00:19:50,480 --> 00:19:54,960
And there's kind of a group of people
402
00:19:54,960 --> 00:19:57,400
that feel like they need to expand quickly
403
00:19:57,400 --> 00:20:01,440
and sort of go into mass production of printing
404
00:20:01,440 --> 00:20:04,080
or whatever they are making.
405
00:20:04,080 --> 00:20:08,200
And if you can just get by and communicate
406
00:20:08,200 --> 00:20:11,600
what you're doing at a local level, then I'm all for it.
407
00:20:11,600 --> 00:20:12,480
I think it's great.
408
00:20:12,480 --> 00:20:15,680
It makes it more difficult in many ways.
409
00:20:15,680 --> 00:20:18,640
You don't take advantage of the bigger systems that exist
410
00:20:18,640 --> 00:20:22,920
to support the bigger capitalist system.
411
00:20:22,920 --> 00:20:25,920
I think the fundamental question that you're raising, Ian,
412
00:20:25,920 --> 00:20:31,400
with can John Lewis survive is, well, it has done for 150 years
413
00:20:31,400 --> 00:20:33,840
ish, I think.
414
00:20:33,840 --> 00:20:37,560
And yeah, see whether they pivot.
415
00:20:37,560 --> 00:20:40,080
I remember they were one of the companies
416
00:20:40,080 --> 00:20:42,640
that I was looking at joining out of university.
417
00:20:42,640 --> 00:20:43,720
Oh, really?
418
00:20:43,720 --> 00:20:46,240
One of my relatives said to me, oh, they've
419
00:20:46,240 --> 00:20:48,800
got that really strange thing where everybody's
420
00:20:48,800 --> 00:20:51,400
a partner at the time.
421
00:20:51,400 --> 00:20:55,560
And I didn't really think about that at all at the time.
422
00:20:55,560 --> 00:20:59,240
But you were a child at the time.
423
00:20:59,240 --> 00:21:00,040
Absolutely.
424
00:21:00,040 --> 00:21:00,800
Yeah.
425
00:21:00,800 --> 00:21:03,280
I used to go shopping in John Barnes, which is one of the,
426
00:21:03,280 --> 00:21:05,120
I mean, they had about across the country
427
00:21:05,120 --> 00:21:06,480
quite a few food stores.
428
00:21:06,480 --> 00:21:09,400
But they had this phrase, never knowingly undersold.
429
00:21:09,400 --> 00:21:10,240
And I always just.
430
00:21:10,240 --> 00:21:11,400
Well, Lewis also.
431
00:21:11,400 --> 00:21:13,600
Yeah.
432
00:21:13,600 --> 00:21:16,520
And I never understood what that meant when I was like six.
433
00:21:16,520 --> 00:21:18,560
It took me eight years and years.
434
00:21:18,560 --> 00:21:22,320
I was so happy that I finally understood what that meant.
435
00:21:22,320 --> 00:21:24,320
But yeah.
436
00:21:24,320 --> 00:21:25,240
This is my thing.
437
00:21:25,240 --> 00:21:28,120
I feel like this is the kind, this is kind of what
438
00:21:28,120 --> 00:21:30,160
the internet needs right now.
439
00:21:30,160 --> 00:21:34,960
And I'm scared that there are people who actually
440
00:21:34,960 --> 00:21:42,160
have genuinely good ideas that are just being hoodwinked
441
00:21:42,160 --> 00:21:47,440
or can't see through their duckweeds.
442
00:21:47,440 --> 00:21:57,080
The only way to move forward is to buy into the existing
443
00:21:57,080 --> 00:21:59,080
VCs and angels.
444
00:21:59,080 --> 00:22:00,760
That's where the money is.
445
00:22:00,760 --> 00:22:02,480
I know, but.
446
00:22:02,480 --> 00:22:03,480
That's where they get it.
447
00:22:03,480 --> 00:22:04,160
No one's got money.
448
00:22:04,160 --> 00:22:06,040
That's the problem.
449
00:22:06,040 --> 00:22:07,080
Where do you get that?
450
00:22:07,080 --> 00:22:08,480
That's what you see first.
451
00:22:08,480 --> 00:22:11,680
But coming back to that, what the internet needs,
452
00:22:11,680 --> 00:22:14,600
what the internet needs, it reminds me of something
453
00:22:14,600 --> 00:22:16,960
that I think we spoke about a few months ago,
454
00:22:16,960 --> 00:22:19,480
of the rewilding the web thing.
455
00:22:19,480 --> 00:22:24,560
It reminds me of the IndieWeb sort of movement.
456
00:22:24,560 --> 00:22:26,480
There's tinyawards.net, which I've just
457
00:22:26,480 --> 00:22:27,840
pasted into our show notes, which
458
00:22:27,840 --> 00:22:35,360
is really fun for discovering non-traditional small websites
459
00:22:35,360 --> 00:22:39,120
that are doing genuinely interesting, creative things.
460
00:22:39,120 --> 00:22:43,520
Rather than just going along with the crowd,
461
00:22:43,520 --> 00:22:47,360
spinning up yet another standard CMS, building a website
462
00:22:47,360 --> 00:22:48,440
in a standard format.
463
00:22:48,440 --> 00:22:49,400
I'm guilty of that.
464
00:22:49,400 --> 00:22:53,280
I'm doing it on a couple of different sites at the moment.
465
00:22:53,280 --> 00:22:56,680
And I kind of regret that I'm going down
466
00:22:56,680 --> 00:22:58,800
the path of least resistance, rather than doing something
467
00:22:58,800 --> 00:22:59,720
interesting.
468
00:22:59,720 --> 00:23:01,680
As soon as you say IndieWeb, I keep
469
00:23:01,680 --> 00:23:03,360
thinking it's Kevin Marks speaking,
470
00:23:03,360 --> 00:23:06,840
because he's been talking about 20 years
471
00:23:06,840 --> 00:23:09,720
talking about IndieWeb.
472
00:23:09,720 --> 00:23:11,520
The other interesting side of this
473
00:23:11,520 --> 00:23:16,800
is what's going on with streaming services, particularly
474
00:23:16,800 --> 00:23:18,960
in the US, where there's a whole bunch of them.
475
00:23:18,960 --> 00:23:21,120
And they started out with this sort of great promise.
476
00:23:21,120 --> 00:23:21,920
There's no ads.
477
00:23:21,920 --> 00:23:23,800
You can watch all this content.
478
00:23:23,800 --> 00:23:26,880
And time has gone on.
479
00:23:26,880 --> 00:23:28,440
There are more competitors.
480
00:23:28,440 --> 00:23:31,680
There are more content owners, rights holders,
481
00:23:31,680 --> 00:23:34,440
who want to run their content on their own streaming service,
482
00:23:34,440 --> 00:23:38,040
instead of licensing it to Netflix, or Amazon, or Apple,
483
00:23:38,040 --> 00:23:39,240
or somebody.
484
00:23:39,240 --> 00:23:41,920
And so all of a sudden, you've got a dozen streaming services.
485
00:23:41,920 --> 00:23:43,640
Nobody can afford to pay for all of them.
486
00:23:43,640 --> 00:23:46,360
And the streaming services themselves
487
00:23:46,360 --> 00:23:48,800
are suddenly realizing that they can make more money
488
00:23:48,800 --> 00:23:55,000
from advertising than they can from subscription money.
489
00:23:55,000 --> 00:23:56,440
Now, this is per user.
490
00:23:56,440 --> 00:23:58,120
And this is actually really interesting,
491
00:23:58,120 --> 00:24:00,960
because we've been here before with online information
492
00:24:00,960 --> 00:24:01,440
services.
493
00:24:01,440 --> 00:24:05,000
This is how AOL and CompuServe were back in the 1990s.
494
00:24:05,000 --> 00:24:09,200
And the difference, there is a guy called the masked scheduler.
495
00:24:09,200 --> 00:24:11,440
His real name is Preston something.
496
00:24:11,440 --> 00:24:15,120
And he was a scheduler for NBC in the States
497
00:24:15,120 --> 00:24:18,080
during the must-see TV period.
498
00:24:18,080 --> 00:24:21,800
He was the guy who scheduled Friends and things like that.
499
00:24:21,800 --> 00:24:25,280
And he says that any new television medium,
500
00:24:25,280 --> 00:24:27,320
he's seen it with cable, for example.
501
00:24:27,320 --> 00:24:28,440
It starts out one way.
502
00:24:28,440 --> 00:24:30,280
And it always eventually regresses
503
00:24:30,280 --> 00:24:33,440
to the template that has been set by commercial broadcast
504
00:24:33,440 --> 00:24:38,880
television, which is live sports, live news, ads,
505
00:24:38,880 --> 00:24:43,720
and least common denominator programming.
506
00:24:43,720 --> 00:24:47,000
And this is what happened on the information services.
507
00:24:47,000 --> 00:24:49,280
They started out as subscription services.
508
00:24:49,280 --> 00:24:51,720
When you have subscriptions, it actually
509
00:24:51,720 --> 00:24:54,480
pays you to develop the kind of content
510
00:24:54,480 --> 00:24:57,400
that attracts devoted fans.
511
00:24:57,400 --> 00:24:59,440
There may not be a huge audience of them,
512
00:24:59,440 --> 00:25:00,760
but they're really fanatical.
513
00:25:00,760 --> 00:25:03,640
And they come back every time there's a new piece of content.
514
00:25:03,640 --> 00:25:05,160
And they really engage with it.
515
00:25:05,160 --> 00:25:06,800
And they're very, very dedicated.
516
00:25:06,800 --> 00:25:09,080
And they will continue paying you subscription money
517
00:25:09,080 --> 00:25:11,720
as long as their favorite stuff is around.
518
00:25:11,720 --> 00:25:14,680
But what happens when you switch to ads
519
00:25:14,680 --> 00:25:20,080
is advertisers don't benefit from small, stable,
520
00:25:20,080 --> 00:25:21,520
fanatical communities.
521
00:25:21,520 --> 00:25:25,240
They benefit from large numbers of eyeballs.
522
00:25:25,240 --> 00:25:31,160
So when AOL switched from subscriptions to advertising,
523
00:25:31,160 --> 00:25:34,320
all of the little niche, I actually had a niche forum.
524
00:25:34,320 --> 00:25:37,200
They had asked us to put together a skeptics forum.
525
00:25:37,200 --> 00:25:40,720
And everything like that was immediately canned.
526
00:25:40,720 --> 00:25:44,240
And they went to things that would pull in large numbers
527
00:25:44,240 --> 00:25:46,360
of people that they could run ads against.
528
00:25:46,360 --> 00:25:47,640
And that is what's going to happen
529
00:25:47,640 --> 00:25:48,760
with these streaming services.
530
00:25:48,760 --> 00:25:53,640
That's what's going to happen with things like Netflix.
531
00:25:53,640 --> 00:25:55,840
Not so much Netflix, because Netflix actually
532
00:25:55,840 --> 00:26:00,960
does sell the whole concept of that they could revive
533
00:26:00,960 --> 00:26:04,720
a series like Suits and have it become a big deal
534
00:26:04,720 --> 00:26:06,920
is an example of what Netflix can do.
535
00:26:06,920 --> 00:26:09,240
But Apple has already said that it wants
536
00:26:09,240 --> 00:26:11,080
to spend less money on content.
537
00:26:11,080 --> 00:26:12,760
All these Hollywood creators who've
538
00:26:12,760 --> 00:26:15,000
signed these giant contracts over the last few years,
539
00:26:15,000 --> 00:26:16,640
I hope they enjoy them because there's not
540
00:26:16,640 --> 00:26:17,760
going to be another one.
541
00:26:17,760 --> 00:26:21,480
So there's a very good YouTube series called
542
00:26:21,480 --> 00:26:25,840
The Rest is Entertainment, Richard Osman, Marina Hyde,
543
00:26:25,840 --> 00:26:28,560
which is good because they come from the inside.
544
00:26:28,560 --> 00:26:31,480
And they were talking about exactly this streamer model
545
00:26:31,480 --> 00:26:34,160
from people who actually write and develop programs.
546
00:26:34,160 --> 00:26:36,280
And they were going over exactly the same thing.
547
00:26:36,280 --> 00:26:38,760
Even if you're desperately interested in the entertainment
548
00:26:38,760 --> 00:26:41,000
industry, that is an excellent YouTube series.
549
00:26:41,000 --> 00:26:42,960
And yeah, they were saying exactly that
550
00:26:42,960 --> 00:26:45,120
of how it was changing and how it was changing
551
00:26:45,120 --> 00:26:47,840
while people were developing.
552
00:26:47,840 --> 00:26:50,120
So basically, they couldn't sell the second series
553
00:26:50,120 --> 00:26:51,040
because it's like, oh no.
554
00:26:51,040 --> 00:26:52,720
Yeah.
555
00:26:52,720 --> 00:26:54,360
Recently, there have been a bunch of things
556
00:26:54,360 --> 00:26:57,480
that were actually ready to air that had been commissioned.
557
00:26:57,480 --> 00:26:58,280
They'd been produced.
558
00:26:58,280 --> 00:26:59,320
They'd been edited.
559
00:26:59,320 --> 00:27:00,640
They were ready to go.
560
00:27:00,640 --> 00:27:02,200
And then they were suddenly pulled
561
00:27:02,200 --> 00:27:06,160
because it'll be more convenient for tax reasons or something.
562
00:27:06,160 --> 00:27:08,000
And it's heartbreaking.
563
00:27:08,000 --> 00:27:11,200
Can you imagine spending years of your life creating something
564
00:27:11,200 --> 00:27:14,000
and then at the last second, the people who commissioned it
565
00:27:14,000 --> 00:27:15,320
said, well, you know what?
566
00:27:15,320 --> 00:27:15,840
Yeah.
567
00:27:15,840 --> 00:27:18,040
Instead, they commissioned Rebel Moon.
568
00:27:18,040 --> 00:27:19,160
Yeah.
569
00:27:19,160 --> 00:27:21,680
I mean, it should be illegal to do that.
570
00:27:21,680 --> 00:27:26,360
Oh, that's quite a, yeah.
571
00:27:26,360 --> 00:27:30,440
The illegal is very interesting.
572
00:27:30,440 --> 00:27:35,400
But you're right about the, I definitely
573
00:27:35,400 --> 00:27:38,000
see what you said about the advertising versus description.
574
00:27:42,000 --> 00:27:49,120
So you might have seen that the CEO of Zoom
575
00:27:49,120 --> 00:27:52,920
wants to do AI clones in meetings.
576
00:27:52,920 --> 00:27:57,120
David, tell us, this can't be for real.
577
00:27:57,120 --> 00:27:57,640
Yeah.
578
00:27:57,640 --> 00:28:00,280
Sadly, it's for real.
579
00:28:00,280 --> 00:28:03,320
We've done a little bit of AI already today.
580
00:28:03,320 --> 00:28:08,280
And so this article appeared on The Verge actually in June.
581
00:28:08,280 --> 00:28:13,280
But there's a lovely sort of callback
582
00:28:13,280 --> 00:28:18,080
to it by Angela Collier, who's a founder of the company
583
00:28:18,080 --> 00:28:23,560
who's a physicist who does lots of lovely YouTube videos.
584
00:28:23,560 --> 00:28:26,920
She did a classic one where she was talking about string theory
585
00:28:26,920 --> 00:28:28,640
while playing the Binding of Isaac
586
00:28:28,640 --> 00:28:30,560
and winning the Binding of Isaac and complaining
587
00:28:30,560 --> 00:28:33,960
about string theory and how it's been done in physics
588
00:28:33,960 --> 00:28:35,480
at the moment, which is wonderful.
589
00:28:35,480 --> 00:28:39,280
However, she looked at this, what
590
00:28:39,280 --> 00:28:43,600
was a fairly incredible first article about AI clones, which
591
00:28:43,600 --> 00:28:45,680
is the sort of normal stupid.
592
00:28:45,680 --> 00:28:48,800
But when you first read it, you thought, yeah, stupid.
593
00:28:48,800 --> 00:28:50,760
But she went a little bit more into it.
594
00:28:50,760 --> 00:28:52,920
And it really is incredibly bad.
595
00:28:52,920 --> 00:28:56,320
So this is Eric on an interview.
596
00:28:56,320 --> 00:28:58,360
And then he does a podcast later on.
597
00:28:58,360 --> 00:29:01,200
Basically, he sort of ups the idea of,
598
00:29:01,200 --> 00:29:03,640
yeah, instead of going to a Zoom meeting,
599
00:29:03,640 --> 00:29:06,080
you can send your clone, sometimes called
600
00:29:06,080 --> 00:29:10,080
a digital twin, and this will do the meeting for you.
601
00:29:10,080 --> 00:29:14,840
And of course, we are absolutely nowhere near that ability.
602
00:29:14,840 --> 00:29:17,880
We're not even vaguely close to that.
603
00:29:17,880 --> 00:29:18,600
How would it learn?
604
00:29:18,600 --> 00:29:20,640
It would learn off your emails and then turn up
605
00:29:20,640 --> 00:29:23,720
to a meeting of how is it going to create any new ideas.
606
00:29:23,720 --> 00:29:27,600
So that was highly, highly ridiculous.
607
00:29:27,600 --> 00:29:32,480
But it's also quite ridiculous in a business state
608
00:29:32,480 --> 00:29:38,200
because obviously, Zoom, who obviously did pretty well
609
00:29:38,200 --> 00:29:39,920
during the epidemic, but other than that,
610
00:29:39,920 --> 00:29:41,000
aren't particularly good.
611
00:29:41,000 --> 00:29:43,840
They aren't controlling the AI market.
612
00:29:43,840 --> 00:29:49,040
So for them to sort of declare that they're dependent on AI
613
00:29:49,040 --> 00:29:51,080
means that they're basically giving their business away
614
00:29:51,080 --> 00:29:52,600
to OpenAI or whoever uses it.
615
00:29:52,600 --> 00:29:53,920
Because I mean, they're not going
616
00:29:53,920 --> 00:29:54,920
to make any difference.
617
00:29:54,920 --> 00:30:00,280
And she spotted a wonderful quote, which it does resonate.
618
00:30:00,280 --> 00:30:04,000
At one point, the interviewer, who is, as is classic,
619
00:30:04,000 --> 00:30:07,200
they're very relaxed with these guys.
620
00:30:07,200 --> 00:30:09,360
Don't give them sort of a hard time.
621
00:30:09,360 --> 00:30:14,040
And I think the first sort of questioning that
622
00:30:14,040 --> 00:30:18,320
was a bit diverse from the interview was he said,
623
00:30:18,320 --> 00:30:22,480
so is the AI model hallucination problem down there in the stack
624
00:30:22,480 --> 00:30:24,840
or are you investing in making sure
625
00:30:24,840 --> 00:30:27,520
that the rate of hallucinations goes down?
626
00:30:27,520 --> 00:30:30,720
Eric Gion says, oh, I think solving the AI hallucination
627
00:30:30,720 --> 00:30:33,920
problem, I think that'll be fixed.
628
00:30:33,920 --> 00:30:37,000
The interviewer says, yeah, but my question is by who?
629
00:30:37,000 --> 00:30:38,200
Is it by you?
630
00:30:38,200 --> 00:30:40,280
Or is it somewhere down the stack?
631
00:30:40,280 --> 00:30:44,320
And then Eric Gion says, yeah, it's someone down the stack.
632
00:30:44,320 --> 00:30:47,960
So he's not only admitting that he has no idea how
633
00:30:47,960 --> 00:30:49,920
this massive problem would be solved,
634
00:30:49,920 --> 00:30:53,240
he's admitting that it's going to be done somewhere else.
635
00:30:53,240 --> 00:30:54,720
He's admitting this.
636
00:30:54,720 --> 00:30:57,320
He's admitting, what really strikes me
637
00:30:57,320 --> 00:31:00,040
is somebody who thinks that you could just
638
00:31:00,040 --> 00:31:02,040
send an avatar to meetings is somebody
639
00:31:02,040 --> 00:31:04,720
who doesn't listen to anything anybody says in a meeting.
640
00:31:04,720 --> 00:31:07,360
Yeah, I mean, and when he says it'll
641
00:31:07,360 --> 00:31:09,040
be fixed by someone down the stack,
642
00:31:09,040 --> 00:31:10,440
I think he's thinking of the stack
643
00:31:10,440 --> 00:31:12,680
as his stack of company employees.
644
00:31:12,680 --> 00:31:14,520
And it'll be somewhere at a low level
645
00:31:14,520 --> 00:31:16,320
where he doesn't have to have any contact.
646
00:31:16,320 --> 00:31:18,840
And this is what Angela Collier was looking at.
647
00:31:18,840 --> 00:31:20,360
Of course, it wouldn't be anyone in these employees
648
00:31:20,360 --> 00:31:22,360
because they're not doing any AI research.
649
00:31:22,360 --> 00:31:25,600
This would be, if he was lucky, someone in open AI.
650
00:31:25,600 --> 00:31:27,080
So what he's basically saying, he's
651
00:31:27,080 --> 00:31:29,040
outsourcing his problems to another company,
652
00:31:29,040 --> 00:31:31,640
which of course means you're just saying, they can just
653
00:31:31,640 --> 00:31:32,840
take over his business.
654
00:31:32,840 --> 00:31:34,280
Because as far as they're concerned,
655
00:31:34,280 --> 00:31:35,760
they've done the hard bit.
656
00:31:35,760 --> 00:31:38,440
All he does is zoom.
657
00:31:38,440 --> 00:31:44,560
So not only is it incredibly daft, a bit of misunderstanding
658
00:31:44,560 --> 00:31:45,080
of AI.
659
00:31:45,080 --> 00:31:49,320
I mean, he says, the question is,
660
00:31:49,320 --> 00:31:51,920
how long will it take before we would have AI closed?
661
00:31:51,920 --> 00:31:55,960
And he says, five or six years, which is bollocks.
662
00:31:55,960 --> 00:32:02,080
I mean, every dimension is complete bollocks,
663
00:32:02,080 --> 00:32:07,040
especially as he isn't offering to fix the major problem.
664
00:32:07,040 --> 00:32:10,000
Did you see that Ray Kurzweil has a new book out?
665
00:32:10,000 --> 00:32:12,800
The Singularity is Nearer.
666
00:32:12,800 --> 00:32:14,280
No.
667
00:32:14,280 --> 00:32:15,320
Oh my gosh.
668
00:32:15,320 --> 00:32:18,400
By which he means that the date when he expects it to happen
669
00:32:18,400 --> 00:32:19,200
has not changed.
670
00:32:19,200 --> 00:32:21,660
It's just that we've gone on a little bit on the timeline.
671
00:32:21,660 --> 00:32:23,960
So it's fewer years.
672
00:32:23,960 --> 00:32:26,120
Oh, right.
673
00:32:26,120 --> 00:32:27,680
We've gone past the Schwarzfeld radius.
674
00:32:27,680 --> 00:32:29,520
And we're not noticing that we're getting slower and slower.
675
00:32:29,520 --> 00:32:30,040
Yeah.
676
00:32:30,040 --> 00:32:30,560
Yeah.
677
00:32:30,560 --> 00:32:31,320
That's probably it.
678
00:32:31,320 --> 00:32:35,440
So I suggest, again, this thing about AI not really achieving
679
00:32:35,440 --> 00:32:35,960
very much.
680
00:32:35,960 --> 00:32:39,480
So I sometimes read Azim Azhar, who does exponential view.
681
00:32:39,480 --> 00:32:40,280
Yeah.
682
00:32:40,280 --> 00:32:42,560
I was going to say he's sometimes a bit of a wanker.
683
00:32:42,560 --> 00:32:45,360
But I have to be careful here, because I'm a journalist too.
684
00:32:45,360 --> 00:32:46,480
One fact.
685
00:32:46,480 --> 00:32:47,200
One fact.
686
00:32:47,200 --> 00:32:50,200
In the mid-90s, both Azim and I were writing
687
00:32:50,200 --> 00:32:51,800
for the same section of The Guardian.
688
00:32:51,800 --> 00:32:54,920
You see, this is why I need to be careful these days.
689
00:32:54,920 --> 00:32:55,440
Yeah.
690
00:32:55,440 --> 00:32:58,440
It was great when we did this, and we could say,
691
00:32:58,440 --> 00:33:00,000
hell, we wanted.
692
00:33:00,000 --> 00:33:01,440
I can't do that anymore, sorry.
693
00:33:01,440 --> 00:33:03,120
But yeah, no.
694
00:33:03,120 --> 00:33:07,120
He's understanding that, obviously, some of his position
695
00:33:07,120 --> 00:33:09,040
now is AI is great.
696
00:33:09,040 --> 00:33:11,280
But even his understanding of what he's seeing is,
697
00:33:11,280 --> 00:33:12,560
AI not so great.
698
00:33:12,560 --> 00:33:14,320
So he's going through a lot of statistics.
699
00:33:14,320 --> 00:33:17,440
And he's doing it for the number of companies
700
00:33:17,440 --> 00:33:20,880
who have had what I would call AI initiatives that have simply
701
00:33:20,880 --> 00:33:23,280
been dropped, because they're not achieving
702
00:33:23,280 --> 00:33:24,840
what they stated to achieve.
703
00:33:24,840 --> 00:33:26,440
They're absolute rubbish and couldn't possibly
704
00:33:26,440 --> 00:33:28,040
achieve what they're supposed to achieve.
705
00:33:28,040 --> 00:33:30,920
Or they've simply been moved to another project
706
00:33:30,920 --> 00:33:32,800
because the money was a waste of time.
707
00:33:32,800 --> 00:33:36,560
And he's politely undermining his own position
708
00:33:36,560 --> 00:33:39,480
by just saying, there's a lot of AI projects
709
00:33:39,480 --> 00:33:41,280
from companies who said AI was going
710
00:33:41,280 --> 00:33:44,160
to be a major shift who are no longer following it
711
00:33:44,160 --> 00:33:44,880
quite so closely.
712
00:33:44,880 --> 00:33:48,240
Did you see there was a piece this week about how much money
713
00:33:48,240 --> 00:33:50,680
Microsoft is losing on AI?
714
00:33:50,680 --> 00:33:51,800
No.
715
00:33:51,800 --> 00:33:53,080
It was billions.
716
00:33:53,080 --> 00:33:54,120
Yeah.
717
00:33:54,120 --> 00:33:54,920
Yeah.
718
00:33:54,920 --> 00:33:57,440
So it's also happening in meta.
719
00:33:57,440 --> 00:34:01,480
They're losing the money in the more traditional way of.
720
00:34:01,480 --> 00:34:02,560
On meta-nodes.
721
00:34:02,560 --> 00:34:05,200
Yes, yes, by getting it wrong again.
722
00:34:08,080 --> 00:34:11,240
So it's also a lot of people are no longer joining companies
723
00:34:11,240 --> 00:34:13,600
that are talking about AI because they're now
724
00:34:13,600 --> 00:34:16,920
a bit suspicious of what they're going to be doing.
725
00:34:16,920 --> 00:34:19,640
So we've already had this AI specialist,
726
00:34:19,640 --> 00:34:22,960
which we've always wondered what the fuck is an AI specialist.
727
00:34:22,960 --> 00:34:25,000
Because there are people doing AI research, which
728
00:34:25,000 --> 00:34:26,240
is a handful of people.
729
00:34:26,240 --> 00:34:28,120
But by definition, people at that level
730
00:34:28,120 --> 00:34:30,080
don't produce products.
731
00:34:30,080 --> 00:34:33,920
They're doing something down the stack, as has been stated.
732
00:34:33,920 --> 00:34:37,640
So there's just been a massive sort of misdirection
733
00:34:37,640 --> 00:34:40,880
about how AI can actually produce.
734
00:34:40,880 --> 00:34:44,000
So we've all seen what it can do.
735
00:34:44,000 --> 00:34:45,200
And it's smart.
736
00:34:45,200 --> 00:34:45,880
And it's useful.
737
00:34:45,880 --> 00:34:48,040
But it's much smaller than it has been.
738
00:34:48,040 --> 00:34:51,360
OK, Tom's hardware says that Microsoft
739
00:34:51,360 --> 00:34:57,800
lost $20 for every $10 copilot AI subscription.
740
00:34:57,800 --> 00:35:00,080
Costed $20.
741
00:35:00,080 --> 00:35:05,520
Well, yeah, I mean, copilot is another example of developers
742
00:35:05,520 --> 00:35:08,200
saying they tried copilot.
743
00:35:08,200 --> 00:35:09,400
It was good.
744
00:35:09,400 --> 00:35:12,320
Does it produce a great increase?
745
00:35:12,320 --> 00:35:14,800
No, it's a comfortable movement.
746
00:35:14,800 --> 00:35:20,840
And oddly enough, in parallel, Visual Code and Visual Studio,
747
00:35:20,840 --> 00:35:22,720
they were already absorbing a lot of the changes
748
00:35:22,720 --> 00:35:25,600
that were actually useful, like inline updates and so forth.
749
00:35:25,600 --> 00:35:27,840
It didn't actually require copilot.
750
00:35:27,840 --> 00:35:30,720
So copilot has been vaguely successful in that it's
751
00:35:30,720 --> 00:35:34,560
gone from nothing to something, but nowhere near as successful,
752
00:35:34,560 --> 00:35:37,200
to allow them to sort of freely raise the price as much
753
00:35:37,200 --> 00:35:39,960
as they would probably like, or to make it worthwhile.
754
00:35:39,960 --> 00:35:43,720
Yeah, Gizmodo has a headline, so far AI
755
00:35:43,720 --> 00:35:46,160
is a money pit that isn't paying off.
756
00:35:46,160 --> 00:35:47,560
And that's last October.
757
00:35:47,560 --> 00:35:49,920
That was ahead of the curve.
758
00:35:49,920 --> 00:35:53,760
Well, we're now getting this thing as well, where I think,
759
00:35:53,760 --> 00:35:56,280
Zuckerberg said it last, to go from,
760
00:35:56,280 --> 00:35:57,880
I can't remember which animal they use.
761
00:35:57,880 --> 00:35:59,120
Is it llamas or ocelots?
762
00:35:59,120 --> 00:35:59,760
Yeah, llamas.
763
00:35:59,760 --> 00:36:02,480
They go from llama three to llama four or whichever.
764
00:36:02,480 --> 00:36:04,800
It's an exponential rise in llamas.
765
00:36:04,800 --> 00:36:10,200
So the learning models literally have to be 10 times bigger.
766
00:36:10,200 --> 00:36:14,880
Often, amount of computing power that they don't actually have.
767
00:36:14,880 --> 00:36:17,440
And in some respects, he said, in two iterations,
768
00:36:17,440 --> 00:36:18,880
the Earth doesn't have.
769
00:36:18,880 --> 00:36:21,760
So it's just this, what the hell are you talking about?
770
00:36:21,760 --> 00:36:26,800
They say they can't really get those improvements now
771
00:36:26,800 --> 00:36:29,440
is just extremely hard, because the amount of power,
772
00:36:29,440 --> 00:36:32,520
that's obviously going to be a, that basically kicks out
773
00:36:32,520 --> 00:36:34,240
everybody, even if it works.
774
00:36:34,240 --> 00:36:36,120
So just the three big ones.
775
00:36:36,120 --> 00:36:39,320
I loved your confusion about the name of the animal just then,
776
00:36:39,320 --> 00:36:40,400
llama or ocelot.
777
00:36:40,400 --> 00:36:42,120
And I knew the answer was llamas,
778
00:36:42,120 --> 00:36:45,920
but I just checked to see if there was an ocelot AI model
779
00:36:45,920 --> 00:36:46,440
already.
780
00:36:46,440 --> 00:36:46,960
There isn't.
781
00:36:46,960 --> 00:36:49,520
But there is, of course, a company called ocelot,
782
00:36:49,520 --> 00:36:51,400
which has AI stuff.
783
00:36:51,400 --> 00:36:53,440
Because all companies have to have AI stuff.
784
00:36:53,440 --> 00:36:54,440
Yeah, yeah.
785
00:36:54,440 --> 00:36:56,880
There's an ocelot AI communication platform
786
00:36:56,880 --> 00:36:57,640
and other things.
787
00:36:57,640 --> 00:36:58,720
I know nothing about them.
788
00:36:58,720 --> 00:37:01,280
I'm making no comment on them, other than just doing
789
00:37:01,280 --> 00:37:03,080
some research live while we're talking.
790
00:37:03,080 --> 00:37:05,960
I'll probably have to write about ocelot before I know it,
791
00:37:05,960 --> 00:37:08,200
just to cover the angles, even though I
792
00:37:08,200 --> 00:37:09,560
don't know anything about them.
793
00:37:09,560 --> 00:37:15,160
So just the last word about this ridiculous thing
794
00:37:15,160 --> 00:37:19,600
that is the digital clone or the clone of people in it.
795
00:37:19,600 --> 00:37:21,440
Now, as Angela Collier pointed out,
796
00:37:21,440 --> 00:37:23,040
obviously, if they were that smart,
797
00:37:23,040 --> 00:37:24,360
why the fuck would they turn up to a meeting?
798
00:37:24,360 --> 00:37:25,760
They just talk to themselves.
799
00:37:25,760 --> 00:37:27,560
I mean, the whole meeting would be redundant.
800
00:37:27,560 --> 00:37:29,200
So I think that's fairly obvious.
801
00:37:29,200 --> 00:37:33,760
But what is interesting is this gap between Eric Young
802
00:37:33,760 --> 00:37:37,920
seeing some interesting results from AI,
803
00:37:37,920 --> 00:37:41,400
then saying something happens in the stat,
804
00:37:41,400 --> 00:37:46,120
and then like the sort of South Park gnomes saying profit
805
00:37:46,120 --> 00:37:49,160
as number three and a big question mark in number two,
806
00:37:49,160 --> 00:37:51,080
which is where I would usually say,
807
00:37:51,080 --> 00:37:52,200
he's just done witchcraft.
808
00:37:52,200 --> 00:37:54,320
He said something happens, don't know what it is,
809
00:37:54,320 --> 00:37:56,360
and then we'll come out with something.
810
00:37:56,360 --> 00:38:01,040
And this is, I think, the start of the very visible decay
811
00:38:01,040 --> 00:38:02,920
of AI as a product.
812
00:38:02,920 --> 00:38:05,840
Whatever we think of AI as abilities, AI as a product now
813
00:38:05,840 --> 00:38:08,400
is beginning to break people's minds, people who
814
00:38:08,400 --> 00:38:10,080
are running large companies.
815
00:38:10,080 --> 00:38:13,560
No longer saying sensible things.
816
00:38:13,560 --> 00:38:18,800
I feel like I don't want to channel Ed Zitron,
817
00:38:18,800 --> 00:38:25,800
but he talks a lot about how people who make these decisions
818
00:38:25,800 --> 00:38:31,400
and say these things are just so far removed from what's
819
00:38:31,400 --> 00:38:32,840
actually really going on.
820
00:38:32,840 --> 00:38:35,200
They haven't got a fucking clue about what
821
00:38:35,200 --> 00:38:36,600
they're talking about.
822
00:38:36,600 --> 00:38:37,920
It's down the stack.
823
00:38:37,920 --> 00:38:39,280
What the hell does that mean?
824
00:38:39,280 --> 00:38:40,720
Does that mean the basement?
825
00:38:40,720 --> 00:38:43,760
Yeah, yeah.
826
00:38:43,760 --> 00:38:45,840
He calls it the rot economy, doesn't he?
827
00:38:45,840 --> 00:38:48,120
Yes, yeah.
828
00:38:48,120 --> 00:38:50,720
I mean, he's very, I listen to him.
829
00:38:50,720 --> 00:38:53,160
I'm trying to be excited, but I do
830
00:38:53,160 --> 00:38:55,800
think that there's some really interesting points.
831
00:38:55,800 --> 00:39:02,080
All of us on this call have some kind of connection
832
00:39:02,080 --> 00:39:04,640
to the stuff that we're talking about.
833
00:39:04,640 --> 00:39:07,360
We understand what we're talking about.
834
00:39:07,360 --> 00:39:09,600
We're not making up bullshit.
835
00:39:09,600 --> 00:39:12,320
We're not hallucinating.
836
00:39:12,320 --> 00:39:16,200
We're not trying to do a TED talk.
837
00:39:19,200 --> 00:39:22,040
And I think that he's probably right that a whole bunch
838
00:39:22,040 --> 00:39:25,400
of these people, journalists, the journalists who
839
00:39:25,400 --> 00:39:26,880
really know what they're talking about,
840
00:39:26,880 --> 00:39:30,080
know how to do real journalism are being pushed out
841
00:39:30,080 --> 00:39:35,040
in favor of people who just will write whatever
842
00:39:35,040 --> 00:39:36,680
to get the clicks.
843
00:39:36,680 --> 00:39:38,960
Well, real journalists can be really inconvenient.
844
00:39:38,960 --> 00:39:43,480
They insist on facts and things that don't appeal to people.
845
00:39:43,480 --> 00:39:45,640
Yes, yes.
846
00:39:45,640 --> 00:39:48,360
And that's a good summary of what's
847
00:39:48,360 --> 00:39:53,320
going on generally in all of these massive companies.
848
00:39:57,320 --> 00:40:01,240
And for our final story, we talked about profits.
849
00:40:01,240 --> 00:40:03,240
We talked about AI.
850
00:40:03,240 --> 00:40:05,960
Well, there's only one story this month,
851
00:40:05,960 --> 00:40:08,320
the last few months, which we can talk about,
852
00:40:08,320 --> 00:40:09,920
which fits all of those.
853
00:40:09,920 --> 00:40:14,880
Wendy, tell us about Alexa and the profit
854
00:40:14,880 --> 00:40:16,440
or the lack of profit.
855
00:40:16,440 --> 00:40:19,400
Well, Ars Technica has a story up
856
00:40:19,400 --> 00:40:24,560
that is a summary of a Wall Street Journal story
857
00:40:24,560 --> 00:40:28,240
about the amount of money that Amazon is losing on Alexa,
858
00:40:28,240 --> 00:40:32,160
which is, according to the Wall Street Journal,
859
00:40:32,160 --> 00:40:35,200
I think it's $10 billion so far.
860
00:40:35,200 --> 00:40:36,720
But.
861
00:40:36,720 --> 00:40:39,160
Wait a minute.
862
00:40:39,160 --> 00:40:44,080
No, $25 billion between 2017 and 2021.
863
00:40:44,080 --> 00:40:46,120
The business unit that focuses on Alexa,
864
00:40:46,120 --> 00:40:49,360
I don't think they break out the figures very precisely.
865
00:40:49,360 --> 00:40:52,680
So there's probably some uncertainty.
866
00:40:52,680 --> 00:40:55,560
Late 2022, Business Insider reported
867
00:40:55,560 --> 00:40:59,680
that Alexa was set to lose $10 billion that year.
868
00:40:59,680 --> 00:41:01,000
So here's the thing.
869
00:41:01,000 --> 00:41:03,600
The people that I know that have smart speakers
870
00:41:03,600 --> 00:41:04,760
really like them.
871
00:41:04,760 --> 00:41:06,400
But if you look at what they use them for,
872
00:41:06,400 --> 00:41:10,320
one of my friends in the States, they
873
00:41:10,320 --> 00:41:12,560
have a Google Home instead of an Alexa,
874
00:41:12,560 --> 00:41:13,720
but it's the same thing.
875
00:41:13,720 --> 00:41:17,440
They use it as a timer to play music
876
00:41:17,440 --> 00:41:19,520
and to turn lights on and off.
877
00:41:19,520 --> 00:41:23,240
And so you hear them say things like, Google
878
00:41:23,240 --> 00:41:26,920
set a timer for four minutes or set a timer for one minute.
879
00:41:26,920 --> 00:41:28,720
And I'm sitting there thinking, one minute?
880
00:41:28,720 --> 00:41:30,800
You can't remember one minute?
881
00:41:30,800 --> 00:41:34,680
But it's just sort of convenient because it's
882
00:41:34,680 --> 00:41:35,200
hands-free.
883
00:41:37,680 --> 00:41:42,240
And the problem is that I think what Amazon imagined
884
00:41:42,240 --> 00:41:46,280
was that Alexa would be a vector for getting people
885
00:41:46,280 --> 00:41:47,880
to order stuff.
886
00:41:47,880 --> 00:41:51,840
And instead, it would be more convenient.
887
00:41:51,840 --> 00:41:53,640
Instead of having to fire up a browser
888
00:41:53,640 --> 00:41:55,440
and choose a product and all the rest of it,
889
00:41:55,440 --> 00:41:59,240
you could just say to Alexa, send me an X,
890
00:41:59,240 --> 00:42:01,240
whatever the thing is.
891
00:42:01,240 --> 00:42:03,400
Well, don't send me an X because I don't really
892
00:42:03,400 --> 00:42:06,080
want a clone of the Twitter platform
893
00:42:06,080 --> 00:42:09,280
to arrive at my house tomorrow.
894
00:42:09,280 --> 00:42:10,840
Send me a flashlight.
895
00:42:10,840 --> 00:42:14,280
Send me a new battery for my Fairphone.
896
00:42:14,280 --> 00:42:17,240
Send me whatever.
897
00:42:17,240 --> 00:42:19,680
But that's not actually how people shop.
898
00:42:19,680 --> 00:42:22,000
Most times, you want to look at choices.
899
00:42:22,000 --> 00:42:27,400
And it's very hard to do that over an audio interface.
900
00:42:27,400 --> 00:42:29,920
And I remember one friend of mine
901
00:42:29,920 --> 00:42:32,680
got an Alexa for his mother because she
902
00:42:32,680 --> 00:42:34,120
was suffering from dementia.
903
00:42:34,120 --> 00:42:36,240
She had trouble typing and so forth.
904
00:42:36,240 --> 00:42:37,760
And he thought it would work for her.
905
00:42:37,760 --> 00:42:40,200
It turned out that because she didn't understand
906
00:42:40,200 --> 00:42:42,360
how to operate the computer without one,
907
00:42:42,360 --> 00:42:47,280
she struggled to understand how to operate a computer with one.
908
00:42:47,280 --> 00:42:50,080
You need the background knowledge and understanding
909
00:42:50,080 --> 00:42:51,760
of how stuff works.
910
00:42:51,760 --> 00:42:55,120
Anyway, so people don't use it the way they hoped.
911
00:42:55,120 --> 00:43:00,960
And so it doesn't follow the Gillette business model
912
00:43:00,960 --> 00:43:05,120
of you sell a razor and then for the years afterwards,
913
00:43:05,120 --> 00:43:06,640
you profit from the razor blades.
914
00:43:06,640 --> 00:43:09,000
Everybody has to buy to service the thing.
915
00:43:09,000 --> 00:43:11,880
And it's just not working out.
916
00:43:11,880 --> 00:43:14,240
But the people who have them really like them.
917
00:43:14,240 --> 00:43:17,840
So then the question becomes, what happens to a technology
918
00:43:17,840 --> 00:43:21,120
that people like but that is unprofitable for the company
919
00:43:21,120 --> 00:43:22,440
that makes it?
920
00:43:22,440 --> 00:43:23,960
What keeps it alive?
921
00:43:23,960 --> 00:43:26,760
I think the universe just stops and restarts because nothing.
922
00:43:26,760 --> 00:43:31,920
I look, the numbers and the finances here
923
00:43:31,920 --> 00:43:35,760
are the sort of eye popping things.
924
00:43:35,760 --> 00:43:42,480
And I think the reasoning behind what's happened is solid.
925
00:43:42,480 --> 00:43:47,960
But I think Amazon had a history of trying a bunch of stuff
926
00:43:47,960 --> 00:43:48,680
to see what worked.
927
00:43:48,680 --> 00:43:51,200
They were trying out those smart buttons
928
00:43:51,200 --> 00:43:55,680
that you'd attach to your washing machine
929
00:43:55,680 --> 00:43:58,640
to reorder your washing powder and stuff.
930
00:43:58,640 --> 00:44:01,040
Those were trying to get you into that subscription
931
00:44:01,040 --> 00:44:07,520
or into that immediate just reorder convenience place.
932
00:44:07,520 --> 00:44:09,320
They, I don't know if they still make them,
933
00:44:09,320 --> 00:44:12,200
they carried on making those for a while.
934
00:44:12,200 --> 00:44:14,880
I agree with you that people clearly
935
00:44:14,880 --> 00:44:18,400
haven't used the voice devices for ordering stuff.
936
00:44:18,400 --> 00:44:23,480
The lack of ability to compare and contrast and decide
937
00:44:23,480 --> 00:44:27,320
what you want to get is not there.
938
00:44:27,320 --> 00:44:29,960
I definitely have them and use them
939
00:44:29,960 --> 00:44:33,240
for switching on lights and all those kind of things as well.
940
00:44:33,240 --> 00:44:35,120
So I've got always on microphones,
941
00:44:35,120 --> 00:44:38,320
which I'm sure Wendy wouldn't approve of in the house, which
942
00:44:38,320 --> 00:44:40,440
are those things that let me switch on my lights.
943
00:44:40,440 --> 00:44:41,840
And me.
944
00:44:41,840 --> 00:44:44,560
Oh yeah, and me.
945
00:44:44,560 --> 00:44:49,880
Yeah, you can kind of see it as the weird thing
946
00:44:49,880 --> 00:44:51,800
is how long they've persisted with it.
947
00:44:51,800 --> 00:44:53,720
They kind of were trying some things.
948
00:44:53,720 --> 00:44:55,160
There was a moment when it looked
949
00:44:55,160 --> 00:44:57,480
like they were getting them out everywhere
950
00:44:57,480 --> 00:44:59,840
and it could be successful if they pivoted them
951
00:44:59,840 --> 00:45:00,840
in the right direction.
952
00:45:00,840 --> 00:45:05,480
But they've been kind of caught out by large language models.
953
00:45:05,480 --> 00:45:09,880
And now if they want to add the chat GPT style,
954
00:45:09,880 --> 00:45:14,240
advanced chat GPT style, Scarlett Johansson hack stuff
955
00:45:14,240 --> 00:45:16,080
that they showed off at their conference
956
00:45:16,080 --> 00:45:20,240
a couple of months ago, which is what I think some people might
957
00:45:20,240 --> 00:45:25,440
want from some kind of voice assistant.
958
00:45:25,440 --> 00:45:29,040
They now want to potentially start adding a subscription
959
00:45:29,040 --> 00:45:32,640
to their Alexa services, which people don't want because
960
00:45:32,640 --> 00:45:34,600
they've already paid for the hardware.
961
00:45:34,600 --> 00:45:37,140
And they're not necessarily going to see the value of then
962
00:45:37,140 --> 00:45:40,240
paying yet more on top of Amazon Prime and Amazon,
963
00:45:40,240 --> 00:45:44,320
whatever else you're paying for subscription to Amazon
964
00:45:44,320 --> 00:45:47,480
to justify the value of those devices.
965
00:45:47,480 --> 00:45:49,880
I find it really interesting and odd
966
00:45:49,880 --> 00:45:52,560
that they carry on bringing them out and changing the form
967
00:45:52,560 --> 00:45:56,120
factors and trying out different things and ways to get them
968
00:45:56,120 --> 00:45:58,640
into different rooms.
969
00:45:58,640 --> 00:46:00,640
The article on Ars Technica here also
970
00:46:00,640 --> 00:46:05,120
mentions the Astro robot, which was that kind of little trolley
971
00:46:05,120 --> 00:46:08,640
robot with a little happy looking face, which hasn't
972
00:46:08,640 --> 00:46:10,640
been made generally available.
973
00:46:10,640 --> 00:46:12,440
Was I think a couple of months ago,
974
00:46:12,440 --> 00:46:14,320
there was a story about them pivoting it just
975
00:46:14,320 --> 00:46:17,600
to be industrial kind of security type monitoring
976
00:46:17,600 --> 00:46:21,480
rather than in home assistance.
977
00:46:21,480 --> 00:46:25,120
So I'm not totally surprised that they've
978
00:46:25,120 --> 00:46:27,640
thrown some money at the wall to see what they could
979
00:46:27,640 --> 00:46:28,600
make a success of.
980
00:46:28,600 --> 00:46:32,000
I find it incredible that they've carried on doing it
981
00:46:32,000 --> 00:46:34,920
and just treating the money as disposable
982
00:46:34,920 --> 00:46:36,680
over that period of time.
983
00:46:36,680 --> 00:46:39,000
Suppose it is somewhat disposable.
984
00:46:39,000 --> 00:46:42,320
They have that sort of stickiness.
985
00:46:42,320 --> 00:46:45,080
It just has no further business line.
986
00:46:45,080 --> 00:46:49,320
So exactly as you say, all the orders I get,
987
00:46:49,320 --> 00:46:51,720
Amazon tells me, I've got this really good book
988
00:46:51,720 --> 00:46:52,920
that you're interested in.
989
00:46:52,920 --> 00:46:54,120
But of course, I have a joint account.
990
00:46:54,120 --> 00:46:54,720
So it's not me.
991
00:46:54,720 --> 00:46:55,320
It's my partner.
992
00:46:55,320 --> 00:46:56,840
And I have no idea what the book is.
993
00:46:56,840 --> 00:46:59,680
And I just say, no.
994
00:46:59,680 --> 00:47:01,600
But the thing is, it doesn't work
995
00:47:01,600 --> 00:47:03,200
because they've got the stickiness.
996
00:47:03,200 --> 00:47:04,520
They don't have the business line.
997
00:47:04,520 --> 00:47:06,200
As you say, they keep on, yes, trying
998
00:47:06,200 --> 00:47:07,880
to put it into different rooms.
999
00:47:07,880 --> 00:47:10,720
But the majority of us, I just use it to turn on Radio 4
1000
00:47:10,720 --> 00:47:12,800
and ask for the weather, which I'm very happy to do
1001
00:47:12,800 --> 00:47:14,880
because I don't like doing that on the phone.
1002
00:47:14,880 --> 00:47:19,840
But for them, it's not going to be any sort of money mover.
1003
00:47:19,840 --> 00:47:22,320
I'd just like to clarify that I'm also
1004
00:47:22,320 --> 00:47:24,560
against voice assistants.
1005
00:47:24,560 --> 00:47:27,280
I think actually there should be some kind of sider
1006
00:47:27,280 --> 00:47:31,120
to say that you are entering a room with this running.
1007
00:47:31,120 --> 00:47:37,040
But just to clarify, I do find, no, under your right,
1008
00:47:37,040 --> 00:47:38,600
they've tried lots of different things.
1009
00:47:38,600 --> 00:47:40,280
I remember the bar and stuff like that.
1010
00:47:40,280 --> 00:47:44,800
My thoughts are, though, that I know that obviously Amazon
1011
00:47:44,800 --> 00:47:46,320
has got a lot of money.
1012
00:47:46,320 --> 00:47:50,800
And I'm wondering when they will kill this.
1013
00:47:50,800 --> 00:47:52,240
Or will they kill it at all?
1014
00:47:52,240 --> 00:47:53,520
Or will they pivot it?
1015
00:47:53,520 --> 00:47:55,240
How will they pivot it?
1016
00:47:55,240 --> 00:48:00,680
So I have thoughts that I wonder how many people would use it.
1017
00:48:00,680 --> 00:48:05,560
So Andy and David who have one.
1018
00:48:05,560 --> 00:48:08,400
If you're a device server, you're
1019
00:48:08,400 --> 00:48:15,000
if your device suddenly started, I'll
1020
00:48:15,000 --> 00:48:20,720
use the controversial term of listened to you
1021
00:48:20,720 --> 00:48:23,480
beyond what it currently does.
1022
00:48:23,480 --> 00:48:27,960
Would you carry on using it as a timer and to access Radio 4?
1023
00:48:32,600 --> 00:48:35,040
I think you're implying if it, yeah, I understand
1024
00:48:35,040 --> 00:48:37,520
what you're implying with that.
1025
00:48:37,520 --> 00:48:40,600
I would probably be put off by that.
1026
00:48:40,600 --> 00:48:42,200
OK.
1027
00:48:42,200 --> 00:48:44,600
So end-shittification goes so far.
1028
00:48:44,600 --> 00:48:46,640
Yeah, yeah.
1029
00:48:46,640 --> 00:48:49,040
I think we're all at the bottom of it.
1030
00:48:49,040 --> 00:48:50,360
Tech no positive.
1031
00:48:50,360 --> 00:48:52,080
I mean, despite the fact this is tech grumps,
1032
00:48:52,080 --> 00:48:56,200
because our ability to use grumpies.
1033
00:48:56,200 --> 00:48:57,920
How long would you keep your Alexa
1034
00:48:57,920 --> 00:49:02,160
if Amazon decided to support it by running ads on it?
1035
00:49:02,160 --> 00:49:04,320
I mean, that would be out the window.
1036
00:49:04,320 --> 00:49:06,320
I literally would just snip it for out the window.
1037
00:49:06,320 --> 00:49:09,120
I don't want ads in my head when I'm trying to just like,
1038
00:49:09,120 --> 00:49:10,320
yeah, tell me what the weather is.
1039
00:49:10,320 --> 00:49:11,280
I don't want an effing ad.
1040
00:49:11,280 --> 00:49:11,920
No.
1041
00:49:11,920 --> 00:49:13,480
I mean, it's funny.
1042
00:49:13,480 --> 00:49:15,600
If I just go back in time, I remember
1043
00:49:15,600 --> 00:49:17,680
that I used Amazon a bit.
1044
00:49:17,680 --> 00:49:18,960
I actually quite like Timo.
1045
00:49:18,960 --> 00:49:21,360
I know this is an advert show for different delivery
1046
00:49:21,360 --> 00:49:22,120
services.
1047
00:49:22,120 --> 00:49:25,080
But Timo was brilliant.
1048
00:49:25,080 --> 00:49:27,160
Yeah, yeah, yeah.
1049
00:49:27,160 --> 00:49:32,080
And the interesting thing was I wasn't an Amazon Prime member.
1050
00:49:32,080 --> 00:49:33,360
I should mention it twice.
1051
00:49:33,360 --> 00:49:36,640
My partner saw that Amazon had these terrible ads that says,
1052
00:49:36,640 --> 00:49:38,320
do you want to get this immediately?
1053
00:49:38,320 --> 00:49:39,560
Then click here.
1054
00:49:39,560 --> 00:49:42,440
And they were doing the sort of typical sort
1055
00:49:42,440 --> 00:49:46,240
of black web dark web pattern of not saying, yeah,
1056
00:49:46,240 --> 00:49:49,000
but that connects to a service that you've got to pay for.
1057
00:49:49,000 --> 00:49:51,080
I suddenly realized I was paying for Amazon Prime.
1058
00:49:51,080 --> 00:49:53,480
I was like, I never signed for Amazon Prime.
1059
00:49:53,480 --> 00:49:54,920
And I just looked back and then finally
1060
00:49:54,920 --> 00:49:55,920
found out what had happened.
1061
00:49:55,920 --> 00:49:57,960
Now, of course, later on I joined Amazon Prime
1062
00:49:57,960 --> 00:49:59,500
because that was actually quite good.
1063
00:49:59,500 --> 00:50:01,760
But that wasn't the reason I joined.
1064
00:50:01,760 --> 00:50:05,840
So I really was a little bit resentful of them for a bit,
1065
00:50:05,840 --> 00:50:08,280
despite the fact that, by the way, I think Prime,
1066
00:50:08,280 --> 00:50:11,560
because of the delivery guys, is particularly good.
1067
00:50:11,560 --> 00:50:13,760
Although I live in London and I prefer to go with shop
1068
00:50:13,760 --> 00:50:14,920
if I possibly can.
1069
00:50:14,920 --> 00:50:17,160
So I said for me, just to summarize,
1070
00:50:17,160 --> 00:50:18,080
the answer would be no.
1071
00:50:18,080 --> 00:50:19,880
If I had ads, I'd just throw it out the window.
1072
00:50:19,880 --> 00:50:23,240
I'm already getting annoyed with the amount of ads on YouTube.
1073
00:50:23,240 --> 00:50:24,360
I mean, I've got the choice.
1074
00:50:24,360 --> 00:50:25,800
I can pay for YouTube.
1075
00:50:25,800 --> 00:50:26,520
What would happen?
1076
00:50:26,520 --> 00:50:28,560
YouTube would still put ads on even though I'm paying for it
1077
00:50:28,560 --> 00:50:31,480
because we already know what's happened.
1078
00:50:31,480 --> 00:50:36,160
Yeah, it's really hard to watch things on YouTube anymore.
1079
00:50:36,160 --> 00:50:36,800
Yeah.
1080
00:50:36,800 --> 00:50:38,400
So what they do is they like it.
1081
00:50:38,400 --> 00:50:40,400
If you find it really small, fortunately,
1082
00:50:40,400 --> 00:50:42,720
the things that I go, I don't go there for many things.
1083
00:50:42,720 --> 00:50:45,400
And the things that I do go there for tend to be small,
1084
00:50:45,400 --> 00:50:46,920
sort of very niche things.
1085
00:50:46,920 --> 00:50:48,520
And so the channels are too small.
1086
00:50:48,520 --> 00:50:49,280
It doesn't have any money.
1087
00:50:49,280 --> 00:50:49,960
Yes, yes.
1088
00:50:49,960 --> 00:50:52,360
I do watch philosophy shows with no audience.
1089
00:50:52,360 --> 00:50:54,560
So they always have no ads.
1090
00:50:54,560 --> 00:50:58,520
The other way around is to watch very long things
1091
00:50:58,520 --> 00:51:02,160
because the ad thing is we'll give you an ad every hour.
1092
00:51:02,160 --> 00:51:04,280
It'll last a few minutes, but we won't interrupt it
1093
00:51:04,280 --> 00:51:06,280
all the time because we recognize you're watching it
1094
00:51:06,280 --> 00:51:09,840
because it's long, which does mean I watch absolute trash
1095
00:51:09,840 --> 00:51:11,560
sometimes, but it is really long.
1096
00:51:11,560 --> 00:51:14,960
I mean, bizarre explanations of Warhammer 40,000,
1097
00:51:14,960 --> 00:51:17,080
which is a lot of nonsense and that.
1098
00:51:17,080 --> 00:51:20,080
But there's lots of them and they're all really long.
1099
00:51:20,080 --> 00:51:23,960
I mean, I watched one that was eight hours.
1100
00:51:23,960 --> 00:51:25,080
Not in one sitting.
1101
00:51:25,080 --> 00:51:28,240
It's disappointing.
1102
00:51:28,240 --> 00:51:30,920
So yeah.
1103
00:51:30,920 --> 00:51:33,880
Andy, did you have anything that you thought
1104
00:51:33,880 --> 00:51:35,800
would be in the spot?
1105
00:51:35,800 --> 00:51:40,600
I agree that inserting ads into it would be frustrating.
1106
00:51:40,600 --> 00:51:45,120
I do occasionally get it offering to help me more
1107
00:51:45,120 --> 00:51:46,880
as it tries to upsell additional features.
1108
00:51:46,880 --> 00:51:50,200
In fact, when I walked into this room to sit down to record,
1109
00:51:50,200 --> 00:51:52,680
I used it to switch on some lights
1110
00:51:52,680 --> 00:51:54,360
and then it offered to set up a schedule
1111
00:51:54,360 --> 00:51:56,280
to switch off all the lights at the end of the day.
1112
00:51:56,280 --> 00:51:57,520
And I just thought, no, shut up.
1113
00:51:57,520 --> 00:51:58,520
No way.
1114
00:51:58,520 --> 00:51:59,520
I didn't ask you to do that.
1115
00:51:59,520 --> 00:52:01,520
Therefore, get out of my face.
1116
00:52:01,520 --> 00:52:03,480
You know, there actually is.
1117
00:52:03,480 --> 00:52:07,120
Years ago, I was at a conference of some kind
1118
00:52:07,120 --> 00:52:08,840
and I ended up having dinner.
1119
00:52:08,840 --> 00:52:13,960
It was sitting at dinner with Ben Schneiderman, who's
1120
00:52:13,960 --> 00:52:15,520
at the University of Maryland and he's
1121
00:52:15,520 --> 00:52:17,240
a specialist on user interfaces.
1122
00:52:17,240 --> 00:52:20,640
And we got into this conversation about the fact
1123
00:52:20,640 --> 00:52:23,480
that there's sort of two kinds of people,
1124
00:52:23,480 --> 00:52:25,920
sort of two kinds of approaches to software.
1125
00:52:25,920 --> 00:52:29,720
One is the software that is quite rule-bound,
1126
00:52:29,720 --> 00:52:30,800
but very predictable.
1127
00:52:30,800 --> 00:52:34,440
You know, if you get the command wrong, it won't do anything.
1128
00:52:34,440 --> 00:52:36,520
But if you get the command right,
1129
00:52:36,520 --> 00:52:39,040
it will do exactly what you expect it to do every time.
1130
00:52:39,040 --> 00:52:40,560
And then there's the kind of software
1131
00:52:40,560 --> 00:52:42,120
that tries to adapt to you.
1132
00:52:42,120 --> 00:52:45,280
And you know, the industry is always
1133
00:52:45,280 --> 00:52:48,000
talking about software agents that are going
1134
00:52:48,000 --> 00:52:49,520
to buy you plane tickets.
1135
00:52:49,520 --> 00:52:52,800
And the latest of this is today's kinds
1136
00:52:52,800 --> 00:52:54,520
of AI and machine learning, where, oh, it's
1137
00:52:54,520 --> 00:52:57,400
going to know what you want before you know yourself
1138
00:52:57,400 --> 00:52:59,560
and all that kind of predictive policing and all that kind
1139
00:52:59,560 --> 00:53:00,760
of thing.
1140
00:53:00,760 --> 00:53:03,960
And both Ben and I were of the, I
1141
00:53:03,960 --> 00:53:06,640
wanted to do what it says it's going to do and nothing else.
1142
00:53:06,640 --> 00:53:07,920
Just leave me alone.
1143
00:53:07,920 --> 00:53:10,280
Let me tell you what I want and you do it.
1144
00:53:10,280 --> 00:53:14,680
You know, so but there is this whole part of the industry
1145
00:53:14,680 --> 00:53:17,240
that's interested in adaptive software.
1146
00:53:17,240 --> 00:53:20,280
And we thought it should be a Harper's Magazine forum,
1147
00:53:20,280 --> 00:53:22,760
but we never got we never figured out
1148
00:53:22,760 --> 00:53:23,720
how to make them do it.
1149
00:53:23,720 --> 00:53:27,640
But there is this sort of constant tension
1150
00:53:27,640 --> 00:53:31,040
between these two approaches to software.
1151
00:53:31,040 --> 00:53:35,800
I think the amazing success of predictive text on a phone
1152
00:53:35,800 --> 00:53:37,960
has fooled a lot of people because it's exactly
1153
00:53:37,960 --> 00:53:41,640
that process that is literally, it's guessing with ease
1154
00:53:41,640 --> 00:53:44,240
because it's seen you write a million times about the text.
1155
00:53:44,240 --> 00:53:45,640
And it mainly gets that right.
1156
00:53:45,640 --> 00:53:48,160
And I can write on one hand and I make the mistake,
1157
00:53:48,160 --> 00:53:51,320
but I still am more productive with one hand,
1158
00:53:51,320 --> 00:53:54,560
although you might say, why am I texting people with one hand?
1159
00:53:54,560 --> 00:53:58,160
But it nevertheless, that is the thing it gets right.
1160
00:53:58,160 --> 00:54:00,840
And unfortunately, people then spin off thinking, yeah,
1161
00:54:00,840 --> 00:54:02,600
therefore you can do loads of other things.
1162
00:54:02,600 --> 00:54:06,080
But really, that's such a tight little thing.
1163
00:54:06,080 --> 00:54:08,600
But actually, it doesn't even depend on you.
1164
00:54:08,600 --> 00:54:11,720
I mean, a lot of what the predictive text does
1165
00:54:11,720 --> 00:54:16,720
on a smartphone is just that in terms of percentages.
1166
00:54:16,720 --> 00:54:18,880
This is what large language models do too.
1167
00:54:18,880 --> 00:54:22,120
And basically, it tries to pick the next word.
1168
00:54:22,120 --> 00:54:24,400
And it predicts the next word based on statistics
1169
00:54:24,400 --> 00:54:27,320
to tell it what the next word usually is.
1170
00:54:27,320 --> 00:54:30,240
And again, once it updates its models exactly the same way
1171
00:54:30,240 --> 00:54:32,720
Google does, that's why Google can correct things better
1172
00:54:32,720 --> 00:54:34,400
than anything else, better than a spell checker
1173
00:54:34,400 --> 00:54:38,560
because it just statistically, given the last million things,
1174
00:54:38,560 --> 00:54:39,760
what did you actually mean?
1175
00:54:39,760 --> 00:54:41,600
Because you're no different from the last million people,
1176
00:54:41,600 --> 00:54:42,320
really.
1177
00:54:42,320 --> 00:54:44,320
And yeah, it does that very well.
1178
00:54:44,320 --> 00:54:46,400
The problem is that I think a lot of people think,
1179
00:54:46,400 --> 00:54:48,840
well, if you can do that, then well, what else can it do?
1180
00:54:48,840 --> 00:54:49,760
It's reading my mind.
1181
00:54:49,760 --> 00:54:51,440
And of course, it isn't exactly the same.
1182
00:54:51,440 --> 00:54:53,760
It's almost pointing out that you aren't actually
1183
00:54:53,760 --> 00:54:54,640
doing anything special.
1184
00:54:54,640 --> 00:54:55,360
You think you are.
1185
00:54:55,360 --> 00:54:57,720
But when you're saying, hi, mum, I'll be late,
1186
00:54:57,720 --> 00:55:00,360
you're not really doing anything particularly special.
1187
00:55:00,360 --> 00:55:02,880
And you get CEO to go, don't worry.
1188
00:55:02,880 --> 00:55:06,400
The cancellation will be solved down the stack.
1189
00:55:06,400 --> 00:55:07,160
Down the stack.
1190
00:55:12,080 --> 00:55:14,560
Thank you very much for all of that.
1191
00:55:14,560 --> 00:55:16,240
There's a lot to take away.
1192
00:55:16,240 --> 00:55:17,760
Try not to be so grumpy.
1193
00:55:17,760 --> 00:55:22,240
So before we end, we have the shout outs and AOBs.
1194
00:55:22,240 --> 00:55:25,160
I got the first one, which is a book
1195
00:55:25,160 --> 00:55:30,520
that I just finished reading called Clearbright Future,
1196
00:55:30,520 --> 00:55:35,840
A Radical Defense of Human Beings, or Human Being, sorry,
1197
00:55:35,840 --> 00:55:38,440
by Paul Mason.
1198
00:55:38,440 --> 00:55:40,880
Paul Mason is quite a well-known figure.
1199
00:55:40,880 --> 00:55:42,320
That Paul Mason?
1200
00:55:42,320 --> 00:55:43,680
That Paul Mason, yes.
1201
00:55:43,680 --> 00:55:44,160
Oh, dear.
1202
00:55:44,160 --> 00:55:45,560
I thought it was quite an interesting book.
1203
00:55:45,560 --> 00:55:46,800
I thought it was actually quite good.
1204
00:55:46,800 --> 00:55:49,400
It was written a while ago, but I just listened to it.
1205
00:55:49,400 --> 00:55:51,080
And it's got a lot of things that I
1206
00:55:51,080 --> 00:55:54,720
thought was actually quite relevant for right now.
1207
00:55:54,720 --> 00:55:56,800
So I do recommend it.
1208
00:55:56,800 --> 00:56:01,880
I gave it a reasonable score on Bookworm.
1209
00:56:01,880 --> 00:56:04,920
Did you read How to Stop Fascism Out of Interest?
1210
00:56:04,920 --> 00:56:09,160
I have not, but I'm going to read that pretty soon.
1211
00:56:09,160 --> 00:56:09,840
It's good.
1212
00:56:09,840 --> 00:56:12,920
I should point out that Paul Mason has stepped
1213
00:56:12,920 --> 00:56:15,200
into some problems recently, which we're not
1214
00:56:15,200 --> 00:56:17,600
going to go over through here, because this is not
1215
00:56:17,600 --> 00:56:18,440
the place to do it.
1216
00:56:18,440 --> 00:56:20,760
But beware.
1217
00:56:20,760 --> 00:56:21,400
Oh, right.
1218
00:56:21,400 --> 00:56:23,080
OK.
1219
00:56:23,080 --> 00:56:26,920
On that bombshell, Wuthering Bites.
1220
00:56:26,920 --> 00:56:28,760
Yes.
1221
00:56:28,760 --> 00:56:31,760
So I should just drop this as a natural thing.
1222
00:56:31,760 --> 00:56:33,760
I'm going to talk at Wuthering Bites.
1223
00:56:33,760 --> 00:56:37,200
Now, actually, I have never talked in a conference.
1224
00:56:37,200 --> 00:56:40,080
I've never been asked to talk in a conference before.
1225
00:56:40,080 --> 00:56:41,400
So this is the first time for me.
1226
00:56:41,400 --> 00:56:42,920
That might seem a little bit, are you serious?
1227
00:56:42,920 --> 00:56:43,720
I'm like, no.
1228
00:56:43,720 --> 00:56:43,960
Really?
1229
00:56:43,960 --> 00:56:44,440
I haven't.
1230
00:56:44,440 --> 00:56:44,760
I haven't.
1231
00:56:44,760 --> 00:56:48,480
Yeah, I have spoken, but I haven't been put on a speaker
1232
00:56:48,480 --> 00:56:49,880
schedule ever before.
1233
00:56:49,880 --> 00:56:51,320
So it's the first time.
1234
00:56:51,320 --> 00:56:53,920
I try not to fuck it up too badly.
1235
00:56:53,920 --> 00:56:57,240
I think I'm going to be talking about, what did Andrew actually
1236
00:56:57,240 --> 00:56:58,200
ask me to talk about?
1237
00:56:58,200 --> 00:56:58,920
Oh, yeah.
1238
00:56:58,920 --> 00:57:02,880
Sort of a game I was developing over during lockdown.
1239
00:57:02,880 --> 00:57:06,360
That was Floor 13, which is a reboot of a game
1240
00:57:06,360 --> 00:57:07,400
I did in the 90s.
1241
00:57:07,400 --> 00:57:11,200
So no doubt, I'll talk about lots of other things as well.
1242
00:57:11,200 --> 00:57:13,760
But it'll be games, games development.
1243
00:57:13,760 --> 00:57:15,560
It could have been AI, so I'm very glad, actually.
1244
00:57:15,560 --> 00:57:17,560
Andrew didn't choose that, so I'm definitely happy.
1245
00:57:17,560 --> 00:57:20,440
Anyway, that's on the, what is it?
1246
00:57:20,440 --> 00:57:21,560
27th.
1247
00:57:21,560 --> 00:57:22,040
Is that it?
1248
00:57:22,040 --> 00:57:22,840
Yeah, end of August.
1249
00:57:22,840 --> 00:57:24,040
Yeah, 27th of August.
1250
00:57:24,040 --> 00:57:26,680
But somewhere up north, which is the only problem.
1251
00:57:26,680 --> 00:57:29,920
I still don't quite understand where it is, but I'll try to get there.
1252
00:57:29,920 --> 00:57:30,800
It's outside Manchester?
1253
00:57:30,800 --> 00:57:31,360
Somewhere up north?
1254
00:57:31,360 --> 00:57:32,280
Yeah.
1255
00:57:32,280 --> 00:57:34,480
It's outside Manchester somewhere in the Peak District,
1256
00:57:34,480 --> 00:57:35,000
isn't it?
1257
00:57:35,000 --> 00:57:36,560
It's outside of Manchester.
1258
00:57:36,560 --> 00:57:37,600
Yeah.
1259
00:57:37,600 --> 00:57:40,840
And I've forgotten the name of it now, but it's called, anyway.
1260
00:57:40,840 --> 00:57:41,680
Hepdenbridge.
1261
00:57:41,680 --> 00:57:43,440
Hepdenbridge, that's it.
1262
00:57:43,440 --> 00:57:48,600
I haven't been north of what for quite some time, so wish me luck.
1263
00:57:48,600 --> 00:57:52,200
There used to be quite a good folk club in Hepdenbridge.
1264
00:57:52,200 --> 00:57:53,520
I don't know if it's still there.
1265
00:57:53,520 --> 00:57:55,960
I think Hepdenbridge is quite a cool place.
1266
00:57:55,960 --> 00:57:56,480
Or it was.
1267
00:57:56,480 --> 00:57:57,280
It is a very cool place.
1268
00:57:57,280 --> 00:57:58,840
I'm sure it still is, yes.
1269
00:57:58,840 --> 00:58:00,440
It's quite pricey now.
1270
00:58:00,440 --> 00:58:01,240
Away you go.
1271
00:58:01,240 --> 00:58:04,320
Yeah, I didn't look too closely at the hotel booking I made,
1272
00:58:04,320 --> 00:58:07,360
because gosh, yeah.
1273
00:58:07,360 --> 00:58:08,760
Anyway, yeah, it's me.
1274
00:58:08,760 --> 00:58:10,000
Woo.
1275
00:58:10,000 --> 00:58:13,440
MNT Pocket Reform.
1276
00:58:13,440 --> 00:58:15,560
Yeah, so I was just going to mention,
1277
00:58:15,560 --> 00:58:21,520
I've been playing with this open source hardware laptop.
1278
00:58:21,520 --> 00:58:26,440
So there's a company called MNT that I think are based in Germany.
1279
00:58:26,440 --> 00:58:28,960
They crowdfunded a couple of open source laptops
1280
00:58:28,960 --> 00:58:34,360
that have got fully sustainable, replaceable parts and so on,
1281
00:58:34,360 --> 00:58:35,840
and upgradable.
1282
00:58:35,840 --> 00:58:40,520
I took delivery of my Pocket Reform a couple of weeks ago.
1283
00:58:40,520 --> 00:58:42,160
I did a tiny, tiny unboxing video.
1284
00:58:42,160 --> 00:58:43,400
I haven't written a review of it yet,
1285
00:58:43,400 --> 00:58:46,320
because I've been too busy doing things like taking it apart
1286
00:58:46,320 --> 00:58:47,520
and fiddling with the insides.
1287
00:58:47,520 --> 00:58:49,160
It's really fun.
1288
00:58:49,160 --> 00:58:51,560
It does remind me somewhat of a netbook,
1289
00:58:51,560 --> 00:58:53,000
but it's much better powered.
1290
00:58:53,000 --> 00:58:55,160
So what's the form factor for that?
1291
00:58:55,160 --> 00:58:58,920
It's about, gosh, I can't even tell you the size of it.
1292
00:58:58,920 --> 00:59:00,720
It's quite thick, so it's probably.
1293
00:59:00,720 --> 00:59:01,680
You have it in front of you.
1294
00:59:01,680 --> 00:59:03,280
Why don't you just hold it up?
1295
00:59:03,280 --> 00:59:03,760
I don't.
1296
00:59:03,760 --> 00:59:04,680
It's downstairs.
1297
00:59:04,680 --> 00:59:07,160
Sorry, otherwise I would show you.
1298
00:59:07,160 --> 00:59:11,080
It's kind of thick, but very small.
1299
00:59:11,080 --> 00:59:15,400
It's less than half the size of a MacBook Pro 13 inch.
1300
00:59:15,400 --> 00:59:20,120
So it's tiny.
1301
00:59:20,120 --> 00:59:21,640
And it's got a little track ball.
1302
00:59:21,640 --> 00:59:26,600
It's got an orthographic keyboard.
1303
00:59:26,600 --> 00:59:31,520
Instead of being a stepped one, it's all laid out as a grid.
1304
00:59:31,520 --> 00:59:34,240
And yeah, I've been enjoying it.
1305
00:59:34,240 --> 00:59:38,240
I would encourage folks to at least have a look at the company.
1306
00:59:38,240 --> 00:59:41,040
They seem to be really interesting and really
1307
00:59:41,040 --> 00:59:45,000
on board with doing things in an open way, which I approve of.
1308
00:59:45,000 --> 00:59:47,640
So yeah.
1309
00:59:47,640 --> 00:59:49,320
I love Cory Doctorow, but because of him,
1310
00:59:49,320 --> 00:59:51,000
all I've heard of is the framework.
1311
00:59:51,000 --> 00:59:52,320
It's framework.
1312
00:59:52,320 --> 00:59:54,000
I may also have a framework.
1313
00:59:54,000 --> 00:59:55,000
Yeah, I'm sure you do.
1314
00:59:55,000 --> 00:59:57,880
I would even question that.
1315
00:59:57,880 --> 01:00:01,720
Which is also a very nice piece of hardware.
1316
01:00:01,720 --> 01:00:06,040
I will likely go for a framework when my current family of laptops
1317
01:00:06,040 --> 01:00:07,040
dies.
1318
01:00:07,040 --> 01:00:08,360
Yeah, they're really good.
1319
01:00:08,360 --> 01:00:11,560
I'm still on my 2015 Mac.
1320
01:00:11,560 --> 01:00:15,800
It would be nice if they made, like I
1321
01:00:15,800 --> 01:00:17,840
like to travel with a backup.
1322
01:00:17,840 --> 01:00:20,720
And I want the backup to be as small and light as possible.
1323
01:00:20,720 --> 01:00:22,120
So what I'd really like them to do
1324
01:00:22,120 --> 01:00:25,680
is to add to their lineup with a really small light thing
1325
01:00:25,680 --> 01:00:28,360
that you could travel with and use as a tablet
1326
01:00:28,360 --> 01:00:33,800
as well as a laptop that I could.
1327
01:00:33,800 --> 01:00:37,920
I like having things match each other so that they can use
1328
01:00:37,920 --> 01:00:42,360
the same AC adapters and things.
1329
01:00:42,360 --> 01:00:45,040
I just want to make sure everything can play Marvel Snap.
1330
01:00:45,040 --> 01:00:46,560
And then I'm happy with whatever I have.
1331
01:00:46,560 --> 01:00:47,080
It's great.
1332
01:00:47,080 --> 01:00:49,080
Yeah, I don't care about that.
1333
01:00:49,080 --> 01:00:54,440
But I do care about the laptops I had.
1334
01:00:54,440 --> 01:00:58,400
I currently have a large surface and a small surface.
1335
01:00:58,400 --> 01:01:02,400
And the laptops, it's not ideal because they're not repairable
1336
01:01:02,400 --> 01:01:04,320
and that's against my principles.
1337
01:01:04,320 --> 01:01:10,280
But at the same time, it is nice that if something happens
1338
01:01:10,280 --> 01:01:12,840
to the AC adapter and it gets lost somewhere,
1339
01:01:12,840 --> 01:01:17,520
you can get one in any Best Buy or any, you know.
1340
01:01:17,520 --> 01:01:19,920
Because the ones I had before were really obscure Macs.
1341
01:01:19,920 --> 01:01:23,520
It was really, really hard if something got lost.
1342
01:01:23,520 --> 01:01:27,840
Yes, I've had to buy a new cable for a MacBook,
1343
01:01:27,840 --> 01:01:30,360
one of those Maglev things, what the hell they're called.
1344
01:01:30,360 --> 01:01:31,840
Yeah, that's impressive when you see
1345
01:01:31,840 --> 01:01:33,400
how much they charge for those.
1346
01:01:33,400 --> 01:01:38,640
Yeah, well, the surface things are not terribly expensive.
1347
01:01:38,640 --> 01:01:42,360
Are you not always just using USB-C to charge your laptops?
1348
01:01:42,360 --> 01:01:48,240
No, the surface has a particular type of plug.
1349
01:01:48,240 --> 01:01:50,000
Microsoft proprietary one.
1350
01:01:50,000 --> 01:01:50,920
I think it's possible.
1351
01:01:50,920 --> 01:01:54,360
I think in some cases it is possible to do it by a USB-C
1352
01:01:54,360 --> 01:01:57,560
if you have the right USB-C. But in practice,
1353
01:01:57,560 --> 01:01:58,320
it's easier for me.
1354
01:01:58,320 --> 01:02:03,840
I have a cable that goes to the surface plug from a USB-C
1355
01:02:03,840 --> 01:02:08,840
and then I have this adapter that it's a standard USB-C
1356
01:02:08,840 --> 01:02:10,440
adapter that has it.
1357
01:02:10,440 --> 01:02:12,720
You have to have the right wattage and stuff.
1358
01:02:12,720 --> 01:02:13,240
Yeah.
1359
01:02:13,240 --> 01:02:13,960
OK.
1360
01:02:13,960 --> 01:02:18,560
I mean, not only do I have to remember I have a 2015 Mac,
1361
01:02:18,560 --> 01:02:22,040
I have to remember it's a late 2015 because they have
1362
01:02:22,040 --> 01:02:24,720
the early and the late have different plugs.
1363
01:02:24,720 --> 01:02:25,720
Oh, Jesus.
1364
01:02:25,720 --> 01:02:27,240
Yeah.
1365
01:02:27,240 --> 01:02:30,480
But look, I'm still using it from like 2015.
1366
01:02:30,480 --> 01:02:33,520
So in that respect, I have to say, well done.
1367
01:02:33,520 --> 01:02:36,800
If there was one thing I wish Apple had never done,
1368
01:02:36,800 --> 01:02:40,240
it's that idiotic magnetic connection.
1369
01:02:40,240 --> 01:02:42,880
I don't know.
1370
01:02:42,880 --> 01:02:44,080
I mean, I kick things about.
1371
01:02:44,080 --> 01:02:46,440
Having the thing fall out and not charge
1372
01:02:46,440 --> 01:02:48,960
is a bigger issue for me than tripping on the thing
1373
01:02:48,960 --> 01:02:50,360
and having it land on the floor.
1374
01:02:50,360 --> 01:02:52,680
I'm one of these people, I have no vision
1375
01:02:52,680 --> 01:02:54,160
of anything below three feet.
1376
01:02:54,160 --> 01:02:55,560
So I kick things all the time.
1377
01:02:55,560 --> 01:02:58,080
So I regularly kick things out.
1378
01:02:58,080 --> 01:02:59,840
So I'm sort of thankful.
1379
01:02:59,840 --> 01:03:01,440
But then I paid a couple of times
1380
01:03:01,440 --> 01:03:02,960
and then I'm sort of not so thankful.
1381
01:03:02,960 --> 01:03:05,560
So yeah, swings and roundabouts.
1382
01:03:05,560 --> 01:03:07,960
And that was our quick shout outs.
1383
01:03:10,920 --> 01:03:11,480
Thank you.
1384
01:03:11,480 --> 01:03:13,160
Thank you, Andy.
1385
01:03:13,160 --> 01:03:17,040
And on that note, I'm going to wrap up the show and say,
1386
01:03:17,040 --> 01:03:23,200
thank you very much to Wendy and to Andy and to David.
1387
01:03:23,200 --> 01:03:25,240
Wendy, would you like to say bye?
1388
01:03:25,240 --> 01:03:26,240
Bye.
1389
01:03:26,240 --> 01:03:28,160
Andy, would you like to say bye?
1390
01:03:28,160 --> 01:03:29,560
Goodbye.
1391
01:03:29,560 --> 01:03:31,200
And David?
1392
01:03:31,200 --> 01:03:31,720
Bye bye.
1393
01:03:31,720 --> 01:03:33,800
And thank you, Ian.
1394
01:03:33,800 --> 01:03:34,800
Thank you very much.
1395
01:03:34,800 --> 01:03:37,560
And if you want to join us on Tech Grumps,
1396
01:03:37,560 --> 01:03:44,200
then get in touch with us via the hashtag on most platforms.
1397
01:03:44,200 --> 01:03:45,840
You can find us.
1398
01:03:45,840 --> 01:03:46,760
We'll get talking.
1399
01:03:46,760 --> 01:03:49,640
And potentially, you could be here.
1400
01:03:49,640 --> 01:03:51,440
Don't be shy.
1401
01:03:51,440 --> 01:03:53,400
Don't be shy, indeed.
1402
01:03:53,400 --> 01:03:56,240
But if you want to use Twitter, it'll only be me.
1403
01:03:56,240 --> 01:03:58,640
Because no one else here is on Twitter.
1404
01:03:58,640 --> 01:04:02,160
Yeah, I'm not good on Twitter to get involved in that.
1405
01:04:02,160 --> 01:04:04,440
Yeah, I'll stay on for a bit.
1406
01:04:04,440 --> 01:04:05,960
But fuck me.
1407
01:04:05,960 --> 01:04:07,760
I have to read some things on Twitter,
1408
01:04:07,760 --> 01:04:35,760
but as little as possible.