1
00:00:00,480 --> 00:00:03,480
foreign

2
00:00:09,300 --> 00:00:12,780
we have one more thing left for this

3
00:00:11,160 --> 00:00:14,639
morning

4
00:00:12,780 --> 00:00:15,719
benno is one of my favorite people at

5
00:00:14,639 --> 00:00:18,960
the conference

6
00:00:15,719 --> 00:00:20,640
um I once described him as the kindest

7
00:00:18,960 --> 00:00:22,199
and most empathetic [ __ ] troll I've

8
00:00:20,640 --> 00:00:23,100
ever met

9
00:00:22,199 --> 00:00:24,840
um

10
00:00:23,100 --> 00:00:27,359
words for that effect that was his that

11
00:00:24,840 --> 00:00:29,939
was your Twitter bio for like a year I

12
00:00:27,359 --> 00:00:32,520
think I wear it as a badge of Pride he

13
00:00:29,939 --> 00:00:35,160
does but he is lovely he has some ideas

14
00:00:32,520 --> 00:00:37,079
about ideas and shockingly he's going to

15
00:00:35,160 --> 00:00:38,600
be able to he's going to talk about them

16
00:00:37,079 --> 00:00:40,920
for 45 minutes

17
00:00:38,600 --> 00:00:44,840
uh please make my band I feel welcome

18
00:00:40,920 --> 00:00:44,840
with the terrifying longevity of ideas

19
00:00:49,020 --> 00:00:52,559
um I think I just got one question

20
00:00:50,879 --> 00:00:54,360
before I started how does this all work

21
00:00:52,559 --> 00:00:55,559
again

22
00:00:54,360 --> 00:00:59,100
um anyway

23
00:00:55,559 --> 00:01:00,660
um so yes my God I I am on a thing again

24
00:00:59,100 --> 00:01:02,579
it's um

25
00:01:00,660 --> 00:01:03,960
I worked out just before it's been about

26
00:01:02,579 --> 00:01:05,939
I think three and a half years since

27
00:01:03,960 --> 00:01:07,500
I've been up here and so uh just before

28
00:01:05,939 --> 00:01:10,560
I get started I just really like to

29
00:01:07,500 --> 00:01:12,720
thank um Daisy and Lily and everyone's

30
00:01:10,560 --> 00:01:15,960
involved for giving me the opportunity

31
00:01:12,720 --> 00:01:19,200
to get back on my [ __ ]

32
00:01:15,960 --> 00:01:21,600
um I will say before I start I do have a

33
00:01:19,200 --> 00:01:23,580
couple of content warnings I will be

34
00:01:21,600 --> 00:01:25,979
discussing the covid-19 pandemic but in

35
00:01:23,580 --> 00:01:28,020
very general terms I will also be

36
00:01:25,979 --> 00:01:30,479
discussing culture War issues

37
00:01:28,020 --> 00:01:32,759
um so including hate speech I'm trying

38
00:01:30,479 --> 00:01:34,860
not to get too deep into it but

39
00:01:32,759 --> 00:01:37,259
yeah if that's too much for you I

40
00:01:34,860 --> 00:01:38,579
absolutely do not mind if you decide

41
00:01:37,259 --> 00:01:39,900
that you need to be somewhere else for a

42
00:01:38,579 --> 00:01:40,619
bit

43
00:01:39,900 --> 00:01:42,720
um

44
00:01:40,619 --> 00:01:45,240
and the other thing that I'll say is I'm

45
00:01:42,720 --> 00:01:46,979
not an expert in a lot of things I'm an

46
00:01:45,240 --> 00:01:48,659
expert in some things I know a lot about

47
00:01:46,979 --> 00:01:51,299
other things

48
00:01:48,659 --> 00:01:53,340
um but I think that it's wise to sort of

49
00:01:51,299 --> 00:01:55,920
take to heart a thing that Tom Eastman

50
00:01:53,340 --> 00:01:57,360
once said about me which is um

51
00:01:55,920 --> 00:01:59,520
that

52
00:01:57,360 --> 00:02:01,320
you know he sees me as someone whose

53
00:01:59,520 --> 00:02:02,700
judgment he trusts but whose words he

54
00:02:01,320 --> 00:02:03,840
does not

55
00:02:02,700 --> 00:02:06,600
um

56
00:02:03,840 --> 00:02:09,720
and that's partly because I am as Daisy

57
00:02:06,600 --> 00:02:12,540
described me a troll to some extent but

58
00:02:09,720 --> 00:02:14,640
also you are free to come up with your

59
00:02:12,540 --> 00:02:16,680
own ideas about ideas and all this kind

60
00:02:14,640 --> 00:02:19,980
of stuff but you know the other thing is

61
00:02:16,680 --> 00:02:21,900
I did tweet or toot

62
00:02:19,980 --> 00:02:24,180
um this slide

63
00:02:21,900 --> 00:02:25,560
um I'm fortunately deleted it but I just

64
00:02:24,180 --> 00:02:27,239
wanted to share it with everyone because

65
00:02:25,560 --> 00:02:28,560
I thought it was really cool

66
00:02:27,239 --> 00:02:30,959
anyway

67
00:02:28,560 --> 00:02:33,660
so this talk started with a phrase that

68
00:02:30,959 --> 00:02:35,099
got stuck in my head once uh when I was

69
00:02:33,660 --> 00:02:38,160
thinking about a different talk idea

70
00:02:35,099 --> 00:02:41,040
which is that ideas have a shape

71
00:02:38,160 --> 00:02:43,920
and the shape lingers

72
00:02:41,040 --> 00:02:45,599
which sounded cool and you know as with

73
00:02:43,920 --> 00:02:47,040
all these things it sort of took up

74
00:02:45,599 --> 00:02:48,000
residence in my head and I sort of

75
00:02:47,040 --> 00:02:49,980
understood what I was getting with it

76
00:02:48,000 --> 00:02:52,739
but I figure I should explain it to

77
00:02:49,980 --> 00:02:54,900
everyone else first so we can start with

78
00:02:52,739 --> 00:02:56,519
the first bit but I think we even need

79
00:02:54,900 --> 00:02:58,260
to start with the first part of that

80
00:02:56,519 --> 00:03:01,620
even

81
00:02:58,260 --> 00:03:03,480
um so I could do the usual stick of

82
00:03:01,620 --> 00:03:05,760
putting up a dictionary definition or a

83
00:03:03,480 --> 00:03:08,580
Wikipedia screenshot or something but

84
00:03:05,760 --> 00:03:10,019
the thing with ideas is that they fall

85
00:03:08,580 --> 00:03:12,420
into one of those you know I knows that

86
00:03:10,019 --> 00:03:14,220
when I seize it kind of categories where

87
00:03:12,420 --> 00:03:16,080
the response to someone trying to put

88
00:03:14,220 --> 00:03:18,120
words around it to Define it is to

89
00:03:16,080 --> 00:03:19,980
squirm and escape and go and gallivant

90
00:03:18,120 --> 00:03:22,319
off somewhere else in abstract

91
00:03:19,980 --> 00:03:24,780
philosophical thought land

92
00:03:22,319 --> 00:03:26,159
um and so for the purposes of this talk

93
00:03:24,780 --> 00:03:28,379
I'm just going to Define an idea as an

94
00:03:26,159 --> 00:03:30,540
abstraction because abstractions are

95
00:03:28,379 --> 00:03:32,580
things that you know as people who are

96
00:03:30,540 --> 00:03:34,319
involved in Tech we understand you know

97
00:03:32,580 --> 00:03:37,019
we create a thing yeah we take a thing

98
00:03:34,319 --> 00:03:38,340
we create a representation of it and

99
00:03:37,019 --> 00:03:41,400
then we come up with a way to realize

100
00:03:38,340 --> 00:03:43,200
that in software or whatever else and

101
00:03:41,400 --> 00:03:46,379
but the thing with abstractions is that

102
00:03:43,200 --> 00:03:48,599
they're definitionally not complete uh

103
00:03:46,379 --> 00:03:50,159
they're simplified or idealized or

104
00:03:48,599 --> 00:03:52,319
whatever you know some or all of the

105
00:03:50,159 --> 00:03:53,819
detail or Nuance is gone because part of

106
00:03:52,319 --> 00:03:55,019
the you know part of the point is we

107
00:03:53,819 --> 00:03:56,760
don't want to have to have them in our

108
00:03:55,019 --> 00:03:58,860
head all the time because they just

109
00:03:56,760 --> 00:04:00,720
don't fit

110
00:03:58,860 --> 00:04:02,879
um you know I can't speak to your brain

111
00:04:00,720 --> 00:04:05,940
but I know that mine can't hold all of

112
00:04:02,879 --> 00:04:07,319
reality all in in it all at once not

113
00:04:05,940 --> 00:04:09,299
that I've tried

114
00:04:07,319 --> 00:04:10,319
um but you know I know that when I'm

115
00:04:09,299 --> 00:04:12,180
trying to hold even a moderately

116
00:04:10,319 --> 00:04:13,620
complicated system in my head it can get

117
00:04:12,180 --> 00:04:15,120
pretty tricky you know ask anyone who's

118
00:04:13,620 --> 00:04:16,919
interrupted me while I'm trying to debug

119
00:04:15,120 --> 00:04:19,560
something

120
00:04:16,919 --> 00:04:22,139
um and so you know you naturally try to

121
00:04:19,560 --> 00:04:23,940
smooth things down and simplify them to

122
00:04:22,139 --> 00:04:26,940
the bits that you need

123
00:04:23,940 --> 00:04:28,620
um the bits that you consider important

124
00:04:26,940 --> 00:04:30,960
um and one of the most obvious things

125
00:04:28,620 --> 00:04:34,440
that we come across as you know techies

126
00:04:30,960 --> 00:04:36,419
software types whatever is an API

127
00:04:34,440 --> 00:04:38,400
um and the most common way you see this

128
00:04:36,419 --> 00:04:40,080
is you know it's an expression of the

129
00:04:38,400 --> 00:04:41,400
important parts of a piece of software

130
00:04:40,080 --> 00:04:43,440
by virtue of it being the way you're

131
00:04:41,400 --> 00:04:44,820
supposed to interact with it you know

132
00:04:43,440 --> 00:04:46,860
some things are Made Easy other things

133
00:04:44,820 --> 00:04:48,780
are made harder other things are

134
00:04:46,860 --> 00:04:50,639
notionally impossible

135
00:04:48,780 --> 00:04:51,780
um some some things are obvious some

136
00:04:50,639 --> 00:04:53,940
things are hidden some things are

137
00:04:51,780 --> 00:04:56,040
documented most aren't

138
00:04:53,940 --> 00:04:58,199
um some abstractions make hard things

139
00:04:56,040 --> 00:04:59,940
easy some make easy things hard

140
00:04:58,199 --> 00:05:02,160
um you know some are vital some whatever

141
00:04:59,940 --> 00:05:05,100
you know but that all of them Express

142
00:05:02,160 --> 00:05:06,300
something about not only the world at

143
00:05:05,100 --> 00:05:08,100
least the bit of it they're concerned

144
00:05:06,300 --> 00:05:09,419
with but often something about the

145
00:05:08,100 --> 00:05:12,300
people that made them or the context

146
00:05:09,419 --> 00:05:13,800
they were made in at the time and the

147
00:05:12,300 --> 00:05:15,060
other thing about apis is that they're

148
00:05:13,800 --> 00:05:18,000
composable

149
00:05:15,060 --> 00:05:19,979
in theory you can put them together to

150
00:05:18,000 --> 00:05:22,380
make bigger things you know things that

151
00:05:19,979 --> 00:05:24,600
then take on their own complexities and

152
00:05:22,380 --> 00:05:27,479
also may need to be simplified

153
00:05:24,600 --> 00:05:29,520
so that you can then reason about

154
00:05:27,479 --> 00:05:31,800
whatever part of the system needs work

155
00:05:29,520 --> 00:05:34,139
at that point and so you build

156
00:05:31,800 --> 00:05:36,180
abstractions around abstractions built

157
00:05:34,139 --> 00:05:38,400
out of abstractions and it all just gets

158
00:05:36,180 --> 00:05:39,960
abstract

159
00:05:38,400 --> 00:05:42,240
um until it doesn't

160
00:05:39,960 --> 00:05:44,100
because somewhere underneath all of that

161
00:05:42,240 --> 00:05:45,780
is a reality

162
00:05:44,100 --> 00:05:48,600
and

163
00:05:45,780 --> 00:05:50,460
you know that reality generally picks

164
00:05:48,600 --> 00:05:52,039
either the worst or the funniest or both

165
00:05:50,460 --> 00:05:54,240
times to stick his head up and point out

166
00:05:52,039 --> 00:05:57,360
all the places where your glorious

167
00:05:54,240 --> 00:05:58,860
abstract Tower of ideas leaks like a

168
00:05:57,360 --> 00:06:02,340
sieve

169
00:05:58,860 --> 00:06:05,039
so with the most tenuous idea of what an

170
00:06:02,340 --> 00:06:08,340
idea is let's talk about why some of

171
00:06:05,039 --> 00:06:10,860
them seem to stick around for so long

172
00:06:08,340 --> 00:06:13,380
there's a famous quote from the

173
00:06:10,860 --> 00:06:15,300
screenwriter John Rogers solidarity with

174
00:06:13,380 --> 00:06:17,160
the Striking workers of the writer's

175
00:06:15,300 --> 00:06:19,680
Guild of America by the way of whom he

176
00:06:17,160 --> 00:06:21,300
was the president at one point there are

177
00:06:19,680 --> 00:06:23,039
two novels that can change a bookish 14

178
00:06:21,300 --> 00:06:24,840
year old's life the Lord of the Rings

179
00:06:23,039 --> 00:06:26,819
and Atlas Shrugged

180
00:06:24,840 --> 00:06:28,440
one is a childish fantasy that often

181
00:06:26,819 --> 00:06:30,060
engenders a lifelong obsession with its

182
00:06:28,440 --> 00:06:31,800
unbelievable Heroes leading to an

183
00:06:30,060 --> 00:06:33,240
emotionally stunted socially crippled

184
00:06:31,800 --> 00:06:34,139
adulthood unable to deal with the real

185
00:06:33,240 --> 00:06:37,139
world

186
00:06:34,139 --> 00:06:39,300
the other of course involves Orcs

187
00:06:37,139 --> 00:06:40,740
I've I've never read Atlas Shrug and I

188
00:06:39,300 --> 00:06:42,840
don't really intend to

189
00:06:40,740 --> 00:06:44,460
um but this was you know in terms of the

190
00:06:42,840 --> 00:06:45,479
books that hit me at the hardest when I

191
00:06:44,460 --> 00:06:48,419
was doing all that world view

192
00:06:45,479 --> 00:06:50,940
development stuff in my younger years is

193
00:06:48,419 --> 00:06:52,620
more of it's like a book length Zine

194
00:06:50,940 --> 00:06:54,780
masquerading as a religious text

195
00:06:52,620 --> 00:06:56,520
masquerading as a joke masquerading as

196
00:06:54,780 --> 00:06:59,160
all of the above it's called um

197
00:06:56,520 --> 00:07:00,660
principia discordia or how I found

198
00:06:59,160 --> 00:07:03,479
goddess and what I did to her when I

199
00:07:00,660 --> 00:07:05,100
found her which professed itself to be

200
00:07:03,479 --> 00:07:07,800
the holy book of religion worshiping

201
00:07:05,100 --> 00:07:10,919
Eris the sort of Greek slash Roman god

202
00:07:07,800 --> 00:07:12,180
of chaos and Discord and this thing blew

203
00:07:10,919 --> 00:07:14,460
my mind mainly because I hadn't

204
00:07:12,180 --> 00:07:16,380
discovered post-modernism yet

205
00:07:14,460 --> 00:07:17,580
um the main thing I liked about it

206
00:07:16,380 --> 00:07:19,560
though was it had this sort of

207
00:07:17,580 --> 00:07:21,840
relentlessly kind of joyful and

208
00:07:19,560 --> 00:07:24,360
irreverent vibe to it while also feeling

209
00:07:21,840 --> 00:07:26,460
very Earnest at times and I can't see

210
00:07:24,360 --> 00:07:28,979
why that's relevant to me now

211
00:07:26,460 --> 00:07:30,479
but anyway one of the passengers that

212
00:07:28,979 --> 00:07:32,940
stuck with me was they talked about the

213
00:07:30,479 --> 00:07:34,919
difference between reality and our ideas

214
00:07:32,940 --> 00:07:37,680
about reality

215
00:07:34,919 --> 00:07:40,800
they Define reality as pure chaos

216
00:07:37,680 --> 00:07:43,139
which our minds can't cope with and so

217
00:07:40,800 --> 00:07:45,900
we kind of construct a set of ideas

218
00:07:43,139 --> 00:07:47,220
about reality that they call a reality

219
00:07:45,900 --> 00:07:50,099
grid

220
00:07:47,220 --> 00:07:51,840
and so we can look through look at the

221
00:07:50,099 --> 00:07:53,520
chaos through that and use it to make

222
00:07:51,840 --> 00:07:55,740
sense we kind of map bits of the chaos

223
00:07:53,520 --> 00:07:57,360
to points on the grid and these grids

224
00:07:55,740 --> 00:08:00,360
are often shared often at a cultural

225
00:07:57,360 --> 00:08:02,819
level they're often also misperceived as

226
00:08:00,360 --> 00:08:05,220
being the actual reality itself

227
00:08:02,819 --> 00:08:06,479
and that's part of their explanation for

228
00:08:05,220 --> 00:08:08,220
the difficulties of cross-cultural

229
00:08:06,479 --> 00:08:10,380
communication and it was the first time

230
00:08:08,220 --> 00:08:12,180
I'd come across this kind of notion and

231
00:08:10,380 --> 00:08:14,400
it just made a lot of sense at the time

232
00:08:12,180 --> 00:08:16,860
it still does and

233
00:08:14,400 --> 00:08:19,740
thinking about that is part of how I get

234
00:08:16,860 --> 00:08:21,960
to this idea that you know grids shapes

235
00:08:19,740 --> 00:08:24,259
we get to Big notion of an idea having a

236
00:08:21,960 --> 00:08:24,259
shape

237
00:08:24,379 --> 00:08:30,300
so some shapes fit together well others

238
00:08:28,319 --> 00:08:32,219
don't sometimes you can overcome the

239
00:08:30,300 --> 00:08:35,459
lack of fit by introducing more shapes

240
00:08:32,219 --> 00:08:36,899
other times you can't uh you you know

241
00:08:35,459 --> 00:08:38,459
your Tower of abstractions has to be

242
00:08:36,899 --> 00:08:40,219
built out of shapes and the shapes have

243
00:08:38,459 --> 00:08:42,539
to fit together well enough

244
00:08:40,219 --> 00:08:45,300
that it does it's stable and doesn't

245
00:08:42,539 --> 00:08:47,700
fall over all the time ideally

246
00:08:45,300 --> 00:08:50,100
um so this kind of construct allowed me

247
00:08:47,700 --> 00:08:52,380
to kind of express some of the ways that

248
00:08:50,100 --> 00:08:53,940
ideas get locked into place

249
00:08:52,380 --> 00:08:55,260
when you're dealing with physical space

250
00:08:53,940 --> 00:08:56,820
you're constrained by things if you're

251
00:08:55,260 --> 00:08:59,220
constructing a building in an urban

252
00:08:56,820 --> 00:09:00,959
environment you're constrained by the

253
00:08:59,220 --> 00:09:02,640
shape and the size of the the lot that

254
00:09:00,959 --> 00:09:04,019
you're building on or by various

255
00:09:02,640 --> 00:09:05,220
regulations or laws or other

256
00:09:04,019 --> 00:09:07,860
requirements that pertain to the

257
00:09:05,220 --> 00:09:09,540
jurisdiction you're in similarly if you

258
00:09:07,860 --> 00:09:11,040
want to change or replace an idea you

259
00:09:09,540 --> 00:09:12,899
can be constrained by the other ideas

260
00:09:11,040 --> 00:09:14,940
that have been placed around it

261
00:09:12,899 --> 00:09:17,940
which forces the new idea to have to

262
00:09:14,940 --> 00:09:20,160
take on aspects of the old one either in

263
00:09:17,940 --> 00:09:22,440
itself or by the affordances that you've

264
00:09:20,160 --> 00:09:26,959
had to make around it and a great

265
00:09:22,440 --> 00:09:26,959
example of this involves CPU design

266
00:09:28,560 --> 00:09:31,620
um

267
00:09:29,519 --> 00:09:33,000
digital equipment corporations PDP range

268
00:09:31,620 --> 00:09:36,540
of mini computers weren't the first

269
00:09:33,000 --> 00:09:38,459
computers ever made but the PDP 7 was

270
00:09:36,540 --> 00:09:39,600
the system on which Unix was developed

271
00:09:38,459 --> 00:09:42,240
originally

272
00:09:39,600 --> 00:09:45,120
and then it got ported to the pdp-11

273
00:09:42,240 --> 00:09:47,160
and on these machines is where we get

274
00:09:45,120 --> 00:09:49,920
the beginnings of a little language

275
00:09:47,160 --> 00:09:53,339
called C

276
00:09:49,920 --> 00:09:55,560
um and the the PDP systems are also

277
00:09:53,339 --> 00:09:57,300
where you get really good examples of

278
00:09:55,560 --> 00:09:59,940
you know the beginnings of what we have

279
00:09:57,300 --> 00:10:02,300
as computer CPU architecture today at

280
00:09:59,940 --> 00:10:05,459
least in part

281
00:10:02,300 --> 00:10:07,500
they have a CPU although although and

282
00:10:05,459 --> 00:10:08,940
oftentimes it wasn't integrated into a

283
00:10:07,500 --> 00:10:10,380
chip at that point you've got some

284
00:10:08,940 --> 00:10:12,959
memory you've got various Communications

285
00:10:10,380 --> 00:10:14,880
buses out to other things cpu's got some

286
00:10:12,959 --> 00:10:17,640
registers and whatnot

287
00:10:14,880 --> 00:10:19,320
um but the primary storage is in Ram

288
00:10:17,640 --> 00:10:20,640
rather than in the registers so on so

289
00:10:19,320 --> 00:10:23,399
forth we're talking significantly

290
00:10:20,640 --> 00:10:25,320
simpler designs though this critically

291
00:10:23,399 --> 00:10:28,080
they're strictly in order they do one

292
00:10:25,320 --> 00:10:30,480
instruction at a time

293
00:10:28,080 --> 00:10:33,779
um and they've got no like the ram is

294
00:10:30,480 --> 00:10:36,420
flat there's no caching there's no

295
00:10:33,779 --> 00:10:37,680
um like Ram that's here or Ram that you

296
00:10:36,420 --> 00:10:39,660
have to talk through someone else to get

297
00:10:37,680 --> 00:10:41,640
to or anything like that

298
00:10:39,660 --> 00:10:43,380
um it's all just that and the thing is

299
00:10:41,640 --> 00:10:46,200
that this is the environment that c was

300
00:10:43,380 --> 00:10:47,640
developed in and it works well there

301
00:10:46,200 --> 00:10:49,380
and it works great there if you're

302
00:10:47,640 --> 00:10:53,220
talking about c as a low-level language

303
00:10:49,380 --> 00:10:56,760
this is where it is a low-level language

304
00:10:53,220 --> 00:10:59,700
um but the thing is it encodes all of

305
00:10:56,760 --> 00:11:02,160
the assumptions and the realities that

306
00:10:59,700 --> 00:11:05,579
were present in this system

307
00:11:02,160 --> 00:11:07,800
and namely that instructions are run one

308
00:11:05,579 --> 00:11:09,980
after the other and that memory is flat

309
00:11:07,800 --> 00:11:13,440
and updates immediately

310
00:11:09,980 --> 00:11:15,420
and those don't hold when your CPU looks

311
00:11:13,440 --> 00:11:18,180
like that

312
00:11:15,420 --> 00:11:21,240
um that's not even the most recent Intel

313
00:11:18,180 --> 00:11:25,320
Core that's I think Sunny Cove which is

314
00:11:21,240 --> 00:11:26,820
about 2018 2019 but

315
00:11:25,320 --> 00:11:29,160
the thing is that's not even the only

316
00:11:26,820 --> 00:11:31,800
core in that system there's multiples of

317
00:11:29,160 --> 00:11:33,660
them like you can't find systems these

318
00:11:31,800 --> 00:11:35,100
odds unless you're going very small you

319
00:11:33,660 --> 00:11:37,260
can't find systems that only have one

320
00:11:35,100 --> 00:11:39,000
core in them anymore and even within

321
00:11:37,260 --> 00:11:40,620
those cores there's just this Cavalcade

322
00:11:39,000 --> 00:11:43,079
of execution units that can run things

323
00:11:40,620 --> 00:11:44,279
in parallel to make things go faster and

324
00:11:43,079 --> 00:11:46,260
to make use of all the tiny little

325
00:11:44,279 --> 00:11:48,360
transistors we can make now and we have

326
00:11:46,260 --> 00:11:50,040
multiple levels of cache to cover for

327
00:11:48,360 --> 00:11:52,200
the fact that CPUs got faster but the

328
00:11:50,040 --> 00:11:54,839
RAM interfaces didn't keep up and all of

329
00:11:52,200 --> 00:11:58,220
this has to present the abstraction

330
00:11:54,839 --> 00:11:58,220
that it's a PDP

331
00:11:58,320 --> 00:12:03,240
um it has to present the the illusion

332
00:12:00,480 --> 00:12:05,100
that it's an in-order flat memory system

333
00:12:03,240 --> 00:12:06,779
so that c doesn't fail

334
00:12:05,100 --> 00:12:08,779
and so you get a huge amount of really

335
00:12:06,779 --> 00:12:11,160
impressive magic going on under the hood

336
00:12:08,779 --> 00:12:12,899
managing the cache hierarchy dynamically

337
00:12:11,160 --> 00:12:15,360
reordering instructions so as to keep as

338
00:12:12,899 --> 00:12:17,220
much of its resources busy as it can uh

339
00:12:15,360 --> 00:12:18,480
guessing Which Way upcoming branches are

340
00:12:17,220 --> 00:12:20,000
going to go so there's not going to be

341
00:12:18,480 --> 00:12:23,100
sitting there twiddling its registers

342
00:12:20,000 --> 00:12:24,899
and then reality turns up to point out

343
00:12:23,100 --> 00:12:27,060
that how all of these layers of

344
00:12:24,899 --> 00:12:30,300
abstraction leak like a sieve in the

345
00:12:27,060 --> 00:12:31,980
form of Specter and meltdown this is a

346
00:12:30,300 --> 00:12:35,240
logo that someone came up with respected

347
00:12:31,980 --> 00:12:35,240
because it's holding a branch

348
00:12:36,000 --> 00:12:38,480
uh anyway

349
00:12:37,860 --> 00:12:40,920
um

350
00:12:38,480 --> 00:12:42,480
thing is that wasn't the intended

351
00:12:40,920 --> 00:12:45,180
outcome

352
00:12:42,480 --> 00:12:47,160
um these things are Crete over time the

353
00:12:45,180 --> 00:12:49,079
designers of the PDP were making the

354
00:12:47,160 --> 00:12:51,540
best system they knew how to for the

355
00:12:49,079 --> 00:12:53,100
time I mean they were also trying to

356
00:12:51,540 --> 00:12:54,600
come in as being very simple they didn't

357
00:12:53,100 --> 00:12:56,040
even call it a computer at first because

358
00:12:54,600 --> 00:12:58,980
their backers didn't want them to

359
00:12:56,040 --> 00:13:00,959
because computer made people think big

360
00:12:58,980 --> 00:13:03,300
and expensive and room sized whereas

361
00:13:00,959 --> 00:13:04,920
they were making simple and like desk

362
00:13:03,300 --> 00:13:05,760
sized

363
00:13:04,920 --> 00:13:07,320
um

364
00:13:05,760 --> 00:13:08,880
and

365
00:13:07,320 --> 00:13:10,680
the trouble is that when you start out

366
00:13:08,880 --> 00:13:12,360
on something like this your ideas your

367
00:13:10,680 --> 00:13:15,000
ideas may be unfettered and free and

368
00:13:12,360 --> 00:13:17,579
able to change at will at the start

369
00:13:15,000 --> 00:13:19,380
but the more people who share your ideas

370
00:13:17,579 --> 00:13:20,760
and the more fixed they become in the

371
00:13:19,380 --> 00:13:22,620
form of you know physical objects or

372
00:13:20,760 --> 00:13:25,740
stuff like that the harder they are to

373
00:13:22,620 --> 00:13:27,839
change and the further down underneath

374
00:13:25,740 --> 00:13:29,760
all of the the top level ideas the

375
00:13:27,839 --> 00:13:32,519
harder they are to move it just gets

376
00:13:29,760 --> 00:13:34,260
worse over time and all of this could be

377
00:13:32,519 --> 00:13:36,480
fixed

378
00:13:34,260 --> 00:13:38,100
if you are willing to change a bunch of

379
00:13:36,480 --> 00:13:39,540
stuff

380
00:13:38,100 --> 00:13:40,800
um you know you could unlock to unlock

381
00:13:39,540 --> 00:13:43,139
the full potential of the way modern

382
00:13:40,800 --> 00:13:44,279
CPUs work if they weren't spending so

383
00:13:43,139 --> 00:13:46,440
much of their time trying to look like a

384
00:13:44,279 --> 00:13:48,180
PDP you would have to have a completely

385
00:13:46,440 --> 00:13:49,560
different language or a completely

386
00:13:48,180 --> 00:13:52,440
different way of expressing your

387
00:13:49,560 --> 00:13:53,459
programs to tackle the fact that they

388
00:13:52,440 --> 00:13:57,480
were going to be trying to run

389
00:13:53,459 --> 00:13:59,760
sequentially and Intel tried this

390
00:13:57,480 --> 00:14:01,380
um the the titanium one of the features

391
00:13:59,760 --> 00:14:03,180
of it was that it had an instruction set

392
00:14:01,380 --> 00:14:05,160
that was explicitly parallel you told it

393
00:14:03,180 --> 00:14:07,560
which instructions could run at the same

394
00:14:05,160 --> 00:14:10,200
time the trouble was that Intel said and

395
00:14:07,560 --> 00:14:13,560
then the compiler will handle that bit

396
00:14:10,200 --> 00:14:16,320
and the compiler said no I won't

397
00:14:13,560 --> 00:14:18,120
um but you know you can also look at

398
00:14:16,320 --> 00:14:19,620
other like architectures out there

399
00:14:18,120 --> 00:14:21,139
Digital Signal processes for example

400
00:14:19,620 --> 00:14:22,860
have a completely different way of

401
00:14:21,139 --> 00:14:24,420
programming them although they do have

402
00:14:22,860 --> 00:14:26,880
an assembly language but you generally

403
00:14:24,420 --> 00:14:27,839
can't do as much looping or branching in

404
00:14:26,880 --> 00:14:30,899
them

405
00:14:27,839 --> 00:14:33,540
um and uh even recently the Raspberry Pi

406
00:14:30,899 --> 00:14:35,339
RP 2040 microcontroller has a

407
00:14:33,540 --> 00:14:37,440
programmable i o unit that has its own

408
00:14:35,339 --> 00:14:39,540
little instructions set and these don't

409
00:14:37,440 --> 00:14:42,240
generally use C for programming because

410
00:14:39,540 --> 00:14:45,540
it just doesn't make sense there modern

411
00:14:42,240 --> 00:14:47,639
GPU programming however does use C but

412
00:14:45,540 --> 00:14:49,380
it's still very weird and has more in

413
00:14:47,639 --> 00:14:51,300
common with almost the older sign of

414
00:14:49,380 --> 00:14:52,920
cray style

415
00:14:51,300 --> 00:14:55,079
um super computer kinds of things but

416
00:14:52,920 --> 00:14:57,240
you know either way when you change

417
00:14:55,079 --> 00:14:59,100
things you have to change a lot of

418
00:14:57,240 --> 00:15:01,680
things and so the things don't tend to

419
00:14:59,100 --> 00:15:03,660
change very much

420
00:15:01,680 --> 00:15:05,600
for example python started out as an

421
00:15:03,660 --> 00:15:08,880
untyped language

422
00:15:05,600 --> 00:15:11,160
objects knew what type they were but

423
00:15:08,880 --> 00:15:13,380
that type wasn't advertised as part of

424
00:15:11,160 --> 00:15:16,380
the variable they were bound to

425
00:15:13,380 --> 00:15:19,860
or variables and this was almost durager

426
00:15:16,380 --> 00:15:20,639
at the time that python came about

427
00:15:19,860 --> 00:15:22,800
um

428
00:15:20,639 --> 00:15:24,360
a whole bunch of areas of languages

429
00:15:22,800 --> 00:15:26,339
design were moving away from the

430
00:15:24,360 --> 00:15:29,339
strictness of compiled languages into

431
00:15:26,339 --> 00:15:31,560
this glorious Dynamic elegant expressive

432
00:15:29,339 --> 00:15:32,880
future that was permitted by the fact

433
00:15:31,560 --> 00:15:34,260
that computers have gotten fast enough

434
00:15:32,880 --> 00:15:36,899
that you could run all your code inside

435
00:15:34,260 --> 00:15:40,139
an interpreter or a VM

436
00:15:36,899 --> 00:15:41,399
um and you know so you know Java came

437
00:15:40,139 --> 00:15:43,019
out only a couple of years after the

438
00:15:41,399 --> 00:15:44,639
first release of python for example and

439
00:15:43,019 --> 00:15:46,019
even and python itself was only about a

440
00:15:44,639 --> 00:15:47,820
year after Pearl five and so all of

441
00:15:46,019 --> 00:15:50,579
these kind of dynamic languages were

442
00:15:47,820 --> 00:15:52,440
showing up and a lot of them well Python

443
00:15:50,579 --> 00:15:54,240
and Pearl at least kind of rejected the

444
00:15:52,440 --> 00:15:57,480
notion of types

445
00:15:54,240 --> 00:15:58,680
um as being kind of slightly irrelevant

446
00:15:57,480 --> 00:16:00,660
um

447
00:15:58,680 --> 00:16:02,699
and you know as with all hot new ideas

448
00:16:00,660 --> 00:16:03,959
it was just everywhere and until it

449
00:16:02,699 --> 00:16:06,120
wasn't

450
00:16:03,959 --> 00:16:09,420
um but being able to say conclusively

451
00:16:06,120 --> 00:16:12,300
what type of thing a variable points at

452
00:16:09,420 --> 00:16:15,660
can be useful

453
00:16:12,300 --> 00:16:17,160
um but abandoning the way that you did

454
00:16:15,660 --> 00:16:19,320
that made it less fiddly and you're able

455
00:16:17,160 --> 00:16:21,540
to write your code faster it just felt

456
00:16:19,320 --> 00:16:23,279
nicer in some ways but then later on we

457
00:16:21,540 --> 00:16:25,980
discovered those kind of middle paths

458
00:16:23,279 --> 00:16:28,320
you can take through you know types and

459
00:16:25,980 --> 00:16:29,519
no types mainly things like type

460
00:16:28,320 --> 00:16:32,519
inference

461
00:16:29,519 --> 00:16:34,500
and just sort of more expressive types

462
00:16:32,519 --> 00:16:36,300
um I mean if you wanted a truly sort of

463
00:16:34,500 --> 00:16:39,540
bewildering example of that rust's type

464
00:16:36,300 --> 00:16:41,459
system is incredible and terrifying at

465
00:16:39,540 --> 00:16:42,240
the same time

466
00:16:41,459 --> 00:16:44,399
um

467
00:16:42,240 --> 00:16:46,920
and so it became somewhat cool to start

468
00:16:44,399 --> 00:16:49,320
adding types back into things and so you

469
00:16:46,920 --> 00:16:50,759
see that with things like typescript and

470
00:16:49,320 --> 00:16:53,940
with the typing mechanisms that got

471
00:16:50,759 --> 00:16:56,880
retrofitted into python 3.

472
00:16:53,940 --> 00:16:58,620
um but adding types to python is not

473
00:16:56,880 --> 00:17:01,800
just a straightforward thing because

474
00:16:58,620 --> 00:17:03,300
they weren't there before typescript had

475
00:17:01,800 --> 00:17:04,620
the benefit of starting as effectively a

476
00:17:03,300 --> 00:17:06,299
whole new language even if it was

477
00:17:04,620 --> 00:17:08,339
getting transpiled back to a language it

478
00:17:06,299 --> 00:17:10,319
looked very similar

479
00:17:08,339 --> 00:17:13,040
um rust is also a whole new language and

480
00:17:10,319 --> 00:17:14,819
so it can do its own thing in contrast

481
00:17:13,040 --> 00:17:17,220
Python's type hinting is getting

482
00:17:14,819 --> 00:17:19,559
retrofitted and got retrofitted in 3.5

483
00:17:17,220 --> 00:17:21,059
and by that point you know backwards

484
00:17:19,559 --> 00:17:22,220
compatibility was something we did not

485
00:17:21,059 --> 00:17:25,380
break

486
00:17:22,220 --> 00:17:26,880
and they were purely optional and then

487
00:17:25,380 --> 00:17:28,319
you've got to contend with the fact that

488
00:17:26,880 --> 00:17:30,179
a huge amount of python code out there

489
00:17:28,319 --> 00:17:31,559
just didn't have any of those Tire pins

490
00:17:30,179 --> 00:17:33,780
and if you want more sort of discussion

491
00:17:31,559 --> 00:17:37,160
of that I can recommend this talk by

492
00:17:33,780 --> 00:17:37,160
some guy called Chris nog about

493
00:17:37,440 --> 00:17:41,640
um

494
00:17:38,700 --> 00:17:44,820
but that's not the biggest problem that

495
00:17:41,640 --> 00:17:47,220
Python 3 in itself had let's talk about

496
00:17:44,820 --> 00:17:50,400
text

497
00:17:47,220 --> 00:17:52,559
you see python along with its peers at

498
00:17:50,400 --> 00:17:54,660
the time was developed in C

499
00:17:52,559 --> 00:17:56,940
mostly on things that were sufficiently

500
00:17:54,660 --> 00:17:58,620
Unix that we can call them that and as

501
00:17:56,940 --> 00:18:00,600
such a Stringer is a sequence of

502
00:17:58,620 --> 00:18:01,380
characters hopefully terminated by a

503
00:18:00,600 --> 00:18:03,720
null

504
00:18:01,380 --> 00:18:05,460
a character is a byte and a byte was

505
00:18:03,720 --> 00:18:07,679
well and truly eight bits by that point

506
00:18:05,460 --> 00:18:11,100
eight bits is plenty you see because

507
00:18:07,679 --> 00:18:13,080
there are only 26 characters

508
00:18:11,100 --> 00:18:15,179
um oh sorry that if you had capitals is

509
00:18:13,080 --> 00:18:17,760
52 I guess and you know then there's

510
00:18:15,179 --> 00:18:19,020
numerals and punctuation and you know

511
00:18:17,760 --> 00:18:21,059
all that kind of stuff but even when you

512
00:18:19,020 --> 00:18:22,440
add those that's that's like 100 or so

513
00:18:21,059 --> 00:18:24,179
and you've still got some room for magic

514
00:18:22,440 --> 00:18:27,120
control sequences and stuff so eight

515
00:18:24,179 --> 00:18:28,100
bits is tons right

516
00:18:27,120 --> 00:18:30,840
um

517
00:18:28,100 --> 00:18:32,460
but by the time we realized this was

518
00:18:30,840 --> 00:18:34,020
rubbish

519
00:18:32,460 --> 00:18:36,419
um which we should have a lot earlier

520
00:18:34,020 --> 00:18:39,120
than we did but we didn't it was all a

521
00:18:36,419 --> 00:18:41,220
bit late Unix and C's relationships with

522
00:18:39,120 --> 00:18:42,960
alternate encodings especially ones that

523
00:18:41,220 --> 00:18:44,220
require more than one byte to represent

524
00:18:42,960 --> 00:18:45,360
a character has always been a bit

525
00:18:44,220 --> 00:18:47,280
fraught

526
00:18:45,360 --> 00:18:49,020
and a large part of this comes down to

527
00:18:47,280 --> 00:18:51,059
an assumed fungibility between a

528
00:18:49,020 --> 00:18:52,200
sequence of characters and a sequence of

529
00:18:51,059 --> 00:18:54,840
bytes

530
00:18:52,200 --> 00:18:57,360
the only real difference between those

531
00:18:54,840 --> 00:18:58,919
two kinds of types in C is that the C

532
00:18:57,360 --> 00:19:00,720
library routines that handle strings

533
00:18:58,919 --> 00:19:02,280
expect a null at the end whereas those

534
00:19:00,720 --> 00:19:05,100
that handle bites expect you to tell you

535
00:19:02,280 --> 00:19:07,860
them tell say how big the buffer is

536
00:19:05,100 --> 00:19:08,820
so many bugs

537
00:19:07,860 --> 00:19:10,380
um

538
00:19:08,820 --> 00:19:12,960
the problem for python is that it made

539
00:19:10,380 --> 00:19:14,280
this assumption too python 2 treats a

540
00:19:12,960 --> 00:19:16,500
sequence of characters in a sequence of

541
00:19:14,280 --> 00:19:18,539
bytes as the same thing

542
00:19:16,500 --> 00:19:21,179
and while you can convert from one of

543
00:19:18,539 --> 00:19:22,740
these to a Unicode string eventually

544
00:19:21,179 --> 00:19:25,320
um if you want to that's entirely

545
00:19:22,740 --> 00:19:27,240
optional and when the decision was made

546
00:19:25,320 --> 00:19:29,820
to allow compatibility breaks between

547
00:19:27,240 --> 00:19:31,500
Python 2 and Python 3 changing that

548
00:19:29,820 --> 00:19:33,600
model from a type that represented both

549
00:19:31,500 --> 00:19:36,240
text and bytes with an optional extra

550
00:19:33,600 --> 00:19:37,980
type supporting Unicode to a type that

551
00:19:36,240 --> 00:19:40,620
only supported bytes that could be

552
00:19:37,980 --> 00:19:43,140
converted via an encoding to a type that

553
00:19:40,620 --> 00:19:45,660
only represented text this was one of

554
00:19:43,140 --> 00:19:48,059
the reasons why the the comparative

555
00:19:45,660 --> 00:19:50,460
compatibility break was made

556
00:19:48,059 --> 00:19:52,640
because it wasn't clear that that kind

557
00:19:50,460 --> 00:19:55,500
of shift could be made any other way

558
00:19:52,640 --> 00:19:58,320
incidentally Python 3 was released in

559
00:19:55,500 --> 00:20:01,440
December 2008 which is closer to the

560
00:19:58,320 --> 00:20:04,940
founding of Amazon than to today also to

561
00:20:01,440 --> 00:20:08,880
the theatrical release of The Lion King

562
00:20:04,940 --> 00:20:10,500
though not to python 1.0

563
00:20:08,880 --> 00:20:12,900
but not by much

564
00:20:10,500 --> 00:20:14,940
anyway as you can imagine this whole

565
00:20:12,900 --> 00:20:17,580
change caused a bunch of issues some of

566
00:20:14,940 --> 00:20:19,980
those were simply because everything had

567
00:20:17,580 --> 00:20:23,640
changed and now everyone's cheese had

568
00:20:19,980 --> 00:20:25,020
been moved and um and but you know some

569
00:20:23,640 --> 00:20:27,299
people relied on that equivalent

570
00:20:25,020 --> 00:20:29,700
equivalence they baked it into the way

571
00:20:27,299 --> 00:20:32,940
that they did it notably web developers

572
00:20:29,700 --> 00:20:34,740
because while HTTP at least up to

573
00:20:32,940 --> 00:20:37,679
version 1.1 is made up of things that

574
00:20:34,740 --> 00:20:39,480
look like strings it's those are defined

575
00:20:37,679 --> 00:20:43,020
to be bites

576
00:20:39,480 --> 00:20:44,760
and on initial release Python 3 didn't

577
00:20:43,020 --> 00:20:47,880
have a bunch of the affordances that

578
00:20:44,760 --> 00:20:51,179
made dealing with byte strings easier

579
00:20:47,880 --> 00:20:52,380
and so it just made it really tricky but

580
00:20:51,179 --> 00:20:53,760
I mean a whole bunch of the other stuff

581
00:20:52,380 --> 00:20:55,320
was just a whole bunch of people going

582
00:20:53,760 --> 00:20:57,720
why

583
00:20:55,320 --> 00:21:00,720
why did you need to do this

584
00:20:57,720 --> 00:21:02,520
everything worked fine before

585
00:21:00,720 --> 00:21:05,280
me

586
00:21:02,520 --> 00:21:07,140
you know why did you need to do that and

587
00:21:05,280 --> 00:21:08,760
I mean most of those people probably had

588
00:21:07,140 --> 00:21:10,799
English as a first language or at least

589
00:21:08,760 --> 00:21:13,140
a close one because

590
00:21:10,799 --> 00:21:15,480
the reason why you needed to do that is

591
00:21:13,140 --> 00:21:18,960
the alternative is giving a huge amount

592
00:21:15,480 --> 00:21:22,679
of preference to languages whose

593
00:21:18,960 --> 00:21:25,200
character set fits in ASCII I.E English

594
00:21:22,679 --> 00:21:28,200
and making all text by definition a

595
00:21:25,200 --> 00:21:31,740
Unicode string meant that now a string

596
00:21:28,200 --> 00:21:33,840
in Python meant any string of any text

597
00:21:31,740 --> 00:21:37,260
if nothing else it makes an explicit

598
00:21:33,840 --> 00:21:39,360
distinction between bytes and text and

599
00:21:37,260 --> 00:21:41,580
that kind of explicitness is you know

600
00:21:39,360 --> 00:21:44,400
far more pythonic than the implicit

601
00:21:41,580 --> 00:21:47,880
distinction that was there before

602
00:21:44,400 --> 00:21:49,980
and the push back against this is one of

603
00:21:47,880 --> 00:21:53,460
the reasons that Python 3 took a long

604
00:21:49,980 --> 00:21:56,220
time to get moving like python 3.5 I

605
00:21:53,460 --> 00:21:58,679
think is when when the the motion really

606
00:21:56,220 --> 00:22:01,020
started to take hold in terms of uh

607
00:21:58,679 --> 00:22:03,659
moving away from python 2.

608
00:22:01,020 --> 00:22:05,400
and for a lot of time there was a lot of

609
00:22:03,659 --> 00:22:07,799
talk that python wouldn't survive the

610
00:22:05,400 --> 00:22:10,080
transition and I think that if it wasn't

611
00:22:07,799 --> 00:22:14,100
for the community that we have around

612
00:22:10,080 --> 00:22:16,260
python like conferences like this and uh

613
00:22:14,100 --> 00:22:17,760
and also you know the sort of the happy

614
00:22:16,260 --> 00:22:20,520
accident of the scientific Community

615
00:22:17,760 --> 00:22:23,520
picking up Python and running with it

616
00:22:20,520 --> 00:22:26,100
um python may be may have been where

617
00:22:23,520 --> 00:22:27,780
Pearl and Ruby are now which is while

618
00:22:26,100 --> 00:22:29,760
they're still used and there's they've

619
00:22:27,780 --> 00:22:31,500
still got communities around them you

620
00:22:29,760 --> 00:22:33,179
don't see a lot of new projects or at

621
00:22:31,500 --> 00:22:36,620
least a lot of not a lot of visible new

622
00:22:33,179 --> 00:22:36,620
projects getting started in them

623
00:22:37,440 --> 00:22:41,100
um you know but I'm really glad that

624
00:22:39,240 --> 00:22:43,020
python did survive primarily because of

625
00:22:41,100 --> 00:22:45,000
the community around it like this

626
00:22:43,020 --> 00:22:45,960
community

627
00:22:45,000 --> 00:22:49,799
um

628
00:22:45,960 --> 00:22:52,080
but as this kind of points out

629
00:22:49,799 --> 00:22:53,700
when you've got you know for the C and

630
00:22:52,080 --> 00:22:55,260
CPU examples you've got these two things

631
00:22:53,700 --> 00:22:57,840
that sort of lock together and make

632
00:22:55,260 --> 00:22:59,940
either of them hard to move and then the

633
00:22:57,840 --> 00:23:03,659
Python 3 example is very much showing

634
00:22:59,940 --> 00:23:05,840
you that when you do try to move them

635
00:23:03,659 --> 00:23:08,100
it all gets a bit fraught

636
00:23:05,840 --> 00:23:10,140
and so the shapes of these things

637
00:23:08,100 --> 00:23:12,960
locking together just makes everything

638
00:23:10,140 --> 00:23:14,159
hard to change

639
00:23:12,960 --> 00:23:19,020
um

640
00:23:14,159 --> 00:23:22,679
but with that now for a brief tangent

641
00:23:19,020 --> 00:23:23,520
in 1994 I joined a choir

642
00:23:22,679 --> 00:23:25,140
um

643
00:23:23,520 --> 00:23:27,059
what you need to what you need to know

644
00:23:25,140 --> 00:23:29,880
about me at the time sorry this is the

645
00:23:27,059 --> 00:23:34,380
Flinders University Choral Society

646
00:23:29,880 --> 00:23:35,159
um we wore orange t-shirts by choice

647
00:23:34,380 --> 00:23:38,760
um

648
00:23:35,159 --> 00:23:41,220
and our acronym is deliberate I'll let

649
00:23:38,760 --> 00:23:42,900
you work that one out yourself

650
00:23:41,220 --> 00:23:44,280
um but what you needed to know about me

651
00:23:42,900 --> 00:23:47,600
at the time is that I was socially

652
00:23:44,280 --> 00:23:49,799
awkward as I was very socially awkward

653
00:23:47,600 --> 00:23:51,600
what I probably needed to know about me

654
00:23:49,799 --> 00:23:53,940
at the time is that I was neurodivergent

655
00:23:51,600 --> 00:23:56,700
but that's in the past

656
00:23:53,940 --> 00:23:58,260
um anyway I was confronted with this

657
00:23:56,700 --> 00:24:00,419
group of people who were just

658
00:23:58,260 --> 00:24:03,840
wonderfully accepting

659
00:24:00,419 --> 00:24:05,880
um very noisy very fun very open about a

660
00:24:03,840 --> 00:24:08,580
bunch of things lots of fun to be around

661
00:24:05,880 --> 00:24:11,820
but I had no idea at all how to

662
00:24:08,580 --> 00:24:14,760
integrate with them or even how to start

663
00:24:11,820 --> 00:24:16,799
and so I approached it in a way that now

664
00:24:14,760 --> 00:24:18,179
I realize is exactly the same way that I

665
00:24:16,799 --> 00:24:19,679
approached learning new systems or

666
00:24:18,179 --> 00:24:23,280
debugging them

667
00:24:19,679 --> 00:24:23,940
um I just sat back and I watched it

668
00:24:23,280 --> 00:24:26,039
um

669
00:24:23,940 --> 00:24:28,500
and I watched how they interacted with

670
00:24:26,039 --> 00:24:31,140
each other I learned the songs that we

671
00:24:28,500 --> 00:24:33,480
sang at the pub after rehearsal

672
00:24:31,140 --> 00:24:35,280
um I watched how in-jokes played out

673
00:24:33,480 --> 00:24:36,600
even though I didn't understand it I

674
00:24:35,280 --> 00:24:40,260
just worked out sort of the calls and

675
00:24:36,600 --> 00:24:41,760
the responses uh so I memorized bits of

676
00:24:40,260 --> 00:24:43,740
Law and history that I heard people

677
00:24:41,760 --> 00:24:45,900
talking about and eventually I had

678
00:24:43,740 --> 00:24:47,780
enough of that in my head that I felt

679
00:24:45,900 --> 00:24:50,580
confident taking part in it myself

680
00:24:47,780 --> 00:24:53,100
riffing on it at times combining bits

681
00:24:50,580 --> 00:24:56,400
and pieces in ways that seemed funny and

682
00:24:53,100 --> 00:24:57,900
I did that so effectively that people

683
00:24:56,400 --> 00:25:00,020
have told me that they thought I'd been

684
00:24:57,900 --> 00:25:04,380
around for even longer than I had been

685
00:25:00,020 --> 00:25:06,600
because I just seemed to fit in so well

686
00:25:04,380 --> 00:25:08,520
um and this whole experience was just

687
00:25:06,600 --> 00:25:10,919
vital to how I developed as a social

688
00:25:08,520 --> 00:25:12,659
person I wouldn't be up here talking

689
00:25:10,919 --> 00:25:14,940
like this in this kind of environment

690
00:25:12,659 --> 00:25:16,260
without that and so much of my sense of

691
00:25:14,940 --> 00:25:18,539
humor and my way of interacting with

692
00:25:16,260 --> 00:25:21,419
people is based on having far too much

693
00:25:18,539 --> 00:25:23,100
vaguely relevant social context stored

694
00:25:21,419 --> 00:25:24,480
in my head that I could make quick

695
00:25:23,100 --> 00:25:25,980
connections between things and find

696
00:25:24,480 --> 00:25:28,320
something funny or interesting in there

697
00:25:25,980 --> 00:25:30,779
and just get it out there quickly

698
00:25:28,320 --> 00:25:33,679
so it pisses me off to no end that AI is

699
00:25:30,779 --> 00:25:33,679
going to steal my job

700
00:25:35,039 --> 00:25:39,419
AI is interesting

701
00:25:37,500 --> 00:25:41,279
um interesting in the sense that I'm

702
00:25:39,419 --> 00:25:42,840
fascinated by the technology and the

703
00:25:41,279 --> 00:25:45,600
mechanics and the thought going into it

704
00:25:42,840 --> 00:25:47,820
but appalled at the way that a lot of

705
00:25:45,600 --> 00:25:48,600
people seem to want to use it

706
00:25:47,820 --> 00:25:51,299
um

707
00:25:48,600 --> 00:25:53,100
it also ties back into the general sort

708
00:25:51,299 --> 00:25:54,779
of gist of the sort of the shapes of

709
00:25:53,100 --> 00:25:56,460
ideas I was talking to but in a slightly

710
00:25:54,779 --> 00:25:58,980
different way

711
00:25:56,460 --> 00:26:00,260
the main type of AI that we refer to at

712
00:25:58,980 --> 00:26:03,600
the moment

713
00:26:00,260 --> 00:26:07,380
is some form of trained neural network

714
00:26:03,600 --> 00:26:09,360
model I will flag here I do not know

715
00:26:07,380 --> 00:26:10,860
what I'm talking about

716
00:26:09,360 --> 00:26:12,419
um this means that what it's doing is

717
00:26:10,860 --> 00:26:14,220
using a prompt to generate a starting

718
00:26:12,419 --> 00:26:17,700
Edition and then repeatedly selecting a

719
00:26:14,220 --> 00:26:19,620
next token in chat gpt's case a word to

720
00:26:17,700 --> 00:26:21,539
put next and this is where the snarky

721
00:26:19,620 --> 00:26:23,820
description of spicy autocorrect comes

722
00:26:21,539 --> 00:26:26,400
in it's very much a growth of the same

723
00:26:23,820 --> 00:26:27,720
idea of as the Markov chain Bots that we

724
00:26:26,400 --> 00:26:29,659
used to have all over Twitter all of the

725
00:26:27,720 --> 00:26:32,039
ebooks accounts

726
00:26:29,659 --> 00:26:35,580
you know back when Bots were a thing on

727
00:26:32,039 --> 00:26:37,860
Twitter and back when Twitter was a rest

728
00:26:35,580 --> 00:26:38,520
in peace hellish bird site

729
00:26:37,860 --> 00:26:41,400
um

730
00:26:38,520 --> 00:26:44,400
but to generate this they need a lot of

731
00:26:41,400 --> 00:26:46,440
content and so you just have to shovel

732
00:26:44,400 --> 00:26:48,299
as much content as you can into them and

733
00:26:46,440 --> 00:26:49,860
then they build up a model of what words

734
00:26:48,299 --> 00:26:51,900
follow what other words in a given

735
00:26:49,860 --> 00:26:54,419
context and the main distinction is that

736
00:26:51,900 --> 00:26:57,360
the Markov model Works purely on on

737
00:26:54,419 --> 00:26:59,700
working out the frequency of which word

738
00:26:57,360 --> 00:27:00,679
follows which other word whereas these

739
00:26:59,700 --> 00:27:03,140
are

740
00:27:00,679 --> 00:27:05,220
significantly more complex

741
00:27:03,140 --> 00:27:06,720
the bits that I find interesting

742
00:27:05,220 --> 00:27:08,760
personally are sort of the design and

743
00:27:06,720 --> 00:27:10,020
the mechanics of training the models and

744
00:27:08,760 --> 00:27:13,020
the technical aspects of making it

745
00:27:10,020 --> 00:27:16,380
usable you know the API aspect of it but

746
00:27:13,020 --> 00:27:19,440
you know the less cool bits like the the

747
00:27:16,380 --> 00:27:21,240
Gathering of the content to feed them

748
00:27:19,440 --> 00:27:22,919
um and sort of the ways in which a lot

749
00:27:21,240 --> 00:27:24,299
of companies are saying oh all that

750
00:27:22,919 --> 00:27:25,860
content that you are generating that's

751
00:27:24,299 --> 00:27:29,220
ours now and we have a Perpetual license

752
00:27:25,860 --> 00:27:30,960
to it or they just go out and take it

753
00:27:29,220 --> 00:27:32,220
um that's stuff that other people have

754
00:27:30,960 --> 00:27:34,820
covered really well and so I'm not going

755
00:27:32,220 --> 00:27:37,380
to go too far into that in this one

756
00:27:34,820 --> 00:27:39,059
but I will say that you know if your

757
00:27:37,380 --> 00:27:41,820
business idea requires people giving you

758
00:27:39,059 --> 00:27:43,740
things for free uh or worse you just

759
00:27:41,820 --> 00:27:46,200
taking them whether that be content or

760
00:27:43,740 --> 00:27:48,059
labor or anything so that you can then

761
00:27:46,200 --> 00:27:49,980
profit from it your business idea kind

762
00:27:48,059 --> 00:27:51,120
of sucks

763
00:27:49,980 --> 00:27:54,539
um

764
00:27:51,120 --> 00:27:56,220
but I do have one excellent AI business

765
00:27:54,539 --> 00:27:58,320
idea

766
00:27:56,220 --> 00:27:59,640
um there's a story that did the round

767
00:27:58,320 --> 00:28:02,460
several years ago where a company

768
00:27:59,640 --> 00:28:04,380
decided to try and improve hiring and

769
00:28:02,460 --> 00:28:06,960
what they did is they fed all of the CVS

770
00:28:04,380 --> 00:28:08,779
into a model along with the the hiring

771
00:28:06,960 --> 00:28:11,820
outcomes

772
00:28:08,779 --> 00:28:14,159
and you know design and this would make

773
00:28:11,820 --> 00:28:14,760
them hire better people right

774
00:28:14,159 --> 00:28:16,860
um

775
00:28:14,760 --> 00:28:19,880
is anyone shocked that it turned out to

776
00:28:16,860 --> 00:28:19,880
be sexist and racist

777
00:28:20,159 --> 00:28:24,600
um my pitch is to lean into that though

778
00:28:22,679 --> 00:28:26,940
if you think that there's bias in your

779
00:28:24,600 --> 00:28:29,220
system we'll train a model on it and

780
00:28:26,940 --> 00:28:31,799
show you how biased it is

781
00:28:29,220 --> 00:28:34,400
so that you can then go oh where that's

782
00:28:31,799 --> 00:28:34,400
sexist

783
00:28:35,220 --> 00:28:40,799
um if anyone wants to do this talk to me

784
00:28:37,440 --> 00:28:42,179
after my idea can't have it

785
00:28:40,799 --> 00:28:43,020
um but anyway

786
00:28:42,179 --> 00:28:44,580
um

787
00:28:43,020 --> 00:28:45,380
and it brings me back to the topic at

788
00:28:44,580 --> 00:28:48,840
hand

789
00:28:45,380 --> 00:28:50,400
bias is a representation of one of these

790
00:28:48,840 --> 00:28:53,100
kinds of ideas it's been lodged

791
00:28:50,400 --> 00:28:54,900
somewhere and dealing with that requires

792
00:28:53,100 --> 00:28:57,240
self-evaluation

793
00:28:54,900 --> 00:28:59,820
we're familiar with the notion of things

794
00:28:57,240 --> 00:29:01,320
being you know familiar we get used to

795
00:28:59,820 --> 00:29:03,659
things they become part of the scenery

796
00:29:01,320 --> 00:29:06,000
part of a routine even annoying things

797
00:29:03,659 --> 00:29:08,179
can become just something we deal with

798
00:29:06,000 --> 00:29:11,039
like init systems

799
00:29:08,179 --> 00:29:12,900
ideas do the same kind of thing if

800
00:29:11,039 --> 00:29:16,140
they're left alone for too long they can

801
00:29:12,900 --> 00:29:18,000
become assumptions or even beliefs and

802
00:29:16,140 --> 00:29:20,220
those things can be dangerous especially

803
00:29:18,000 --> 00:29:22,200
if you forget or you don't realize that

804
00:29:20,220 --> 00:29:24,539
they're there

805
00:29:22,200 --> 00:29:25,740
um one of one of my issues with these AI

806
00:29:24,539 --> 00:29:27,659
models is they can't they can't

807
00:29:25,740 --> 00:29:30,720
introspect they can't evaluate

808
00:29:27,659 --> 00:29:32,399
themselves content goes in gets turned

809
00:29:30,720 --> 00:29:34,980
into a thing from which content can

810
00:29:32,399 --> 00:29:37,140
emerge but the ability of that system to

811
00:29:34,980 --> 00:29:39,120
understand why it made that content is

812
00:29:37,140 --> 00:29:41,700
not there

813
00:29:39,120 --> 00:29:44,159
um not only that it doesn't know if it

814
00:29:41,700 --> 00:29:45,480
was if the content either the content

815
00:29:44,159 --> 00:29:47,700
that it generated or the content that

816
00:29:45,480 --> 00:29:51,899
it's using is good

817
00:29:47,700 --> 00:29:53,340
for any of multiple definitions of good

818
00:29:51,899 --> 00:29:54,960
um hence all the effort that they have

819
00:29:53,340 --> 00:29:57,299
to put around put into it making guard

820
00:29:54,960 --> 00:29:59,460
rails around what they'll output and all

821
00:29:57,299 --> 00:30:01,440
of the hilarious you know my grandmother

822
00:29:59,460 --> 00:30:04,260
worked in a Napalm Factory ways that

823
00:30:01,440 --> 00:30:07,980
they used to get around it

824
00:30:04,260 --> 00:30:10,620
um being able to evaluate the Worth or

825
00:30:07,980 --> 00:30:13,860
the value or the goodness of our ideas

826
00:30:10,620 --> 00:30:15,840
is critical and our assumptions because

827
00:30:13,860 --> 00:30:16,980
uh

828
00:30:15,840 --> 00:30:19,200
and we even need to be able to

829
00:30:16,980 --> 00:30:21,600
reevaluate what idea our idea of worth

830
00:30:19,200 --> 00:30:24,240
is you know or goodness

831
00:30:21,600 --> 00:30:25,799
because those ideas are forming part of

832
00:30:24,240 --> 00:30:27,899
the broader sort of geometry and

833
00:30:25,799 --> 00:30:30,840
topography that influences what other

834
00:30:27,899 --> 00:30:33,120
ideas that we'll accept or find Value in

835
00:30:30,840 --> 00:30:35,640
lacking that ability is what makes these

836
00:30:33,120 --> 00:30:38,340
AI systems unable to produce anything

837
00:30:35,640 --> 00:30:41,399
truly new except in as much as they're

838
00:30:38,340 --> 00:30:44,059
constructing new things from the

839
00:30:41,399 --> 00:30:47,760
dismembered parts of other things

840
00:30:44,059 --> 00:30:49,320
but I think I can sense someone thinking

841
00:30:47,760 --> 00:30:52,080
very loudly that they've come up with

842
00:30:49,320 --> 00:30:54,120
the perfect well actually for later on

843
00:30:52,080 --> 00:30:55,500
didn't AI just make a new sorting

844
00:30:54,120 --> 00:30:59,039
algorithm

845
00:30:55,500 --> 00:31:01,220
and it did but not that kind of model

846
00:30:59,039 --> 00:31:03,899
this kind of model

847
00:31:01,220 --> 00:31:05,460
is there might be some of that large

848
00:31:03,899 --> 00:31:08,399
language model generative stuff going on

849
00:31:05,460 --> 00:31:09,899
under the hood but it's got the

850
00:31:08,399 --> 00:31:11,100
particular model that they're using for

851
00:31:09,899 --> 00:31:13,380
this one which is the same one that they

852
00:31:11,100 --> 00:31:14,399
use to play go and Starcraft and several

853
00:31:13,380 --> 00:31:18,120
other things

854
00:31:14,399 --> 00:31:20,700
it has a notion of what good is for it

855
00:31:18,120 --> 00:31:22,679
it knows in this case that it's been

856
00:31:20,700 --> 00:31:24,059
given a list of unsorted items and the

857
00:31:22,679 --> 00:31:26,039
thing that it makes has to turn that

858
00:31:24,059 --> 00:31:29,580
into a sorted list

859
00:31:26,039 --> 00:31:32,700
and it'll end up taking less time or

860
00:31:29,580 --> 00:31:34,260
fewer instructions to do that is better

861
00:31:32,700 --> 00:31:35,880
and so

862
00:31:34,260 --> 00:31:38,399
you know it's a very simplistic

863
00:31:35,880 --> 00:31:41,159
definition of what good is though and

864
00:31:38,399 --> 00:31:44,340
the thing that really worries me is that

865
00:31:41,159 --> 00:31:47,940
if when or if AI gets used in the

866
00:31:44,340 --> 00:31:50,480
development of policy or law or hell

867
00:31:47,940 --> 00:31:54,799
even the next Marvel movie

868
00:31:50,480 --> 00:31:54,799
who gets to Define what's good

869
00:31:55,740 --> 00:31:59,580
um

870
00:31:56,940 --> 00:32:01,860
because you and I can re-examine our

871
00:31:59,580 --> 00:32:03,120
ideas we can look at these things and

872
00:32:01,860 --> 00:32:06,120
say hang on this one isn't working for

873
00:32:03,120 --> 00:32:07,620
me anymore and if we can if not outright

874
00:32:06,120 --> 00:32:09,419
discard it we can at least sort of flag

875
00:32:07,620 --> 00:32:11,340
it as saying this not this one anymore

876
00:32:09,419 --> 00:32:12,960
or something like that

877
00:32:11,340 --> 00:32:16,080
and

878
00:32:12,960 --> 00:32:18,240
you know that we get

879
00:32:16,080 --> 00:32:20,460
um sorry

880
00:32:18,240 --> 00:32:21,539
and if we don't do this as a it's

881
00:32:20,460 --> 00:32:23,039
dangerous because we can end up with

882
00:32:21,539 --> 00:32:25,860
these ideas that stick around for too

883
00:32:23,039 --> 00:32:28,679
long and those ones can be at a personal

884
00:32:25,860 --> 00:32:30,120
level or they can be really they can be

885
00:32:28,679 --> 00:32:31,380
at a societal level which can be even

886
00:32:30,120 --> 00:32:33,779
more pernicious

887
00:32:31,380 --> 00:32:35,460
and just ha

888
00:32:33,779 --> 00:32:38,460
so many people can end up assuming that

889
00:32:35,460 --> 00:32:41,760
these ideas are immutable that they are

890
00:32:38,460 --> 00:32:44,340
for lack of anything else they are

891
00:32:41,760 --> 00:32:46,440
um and as with all the ideas these ideas

892
00:32:44,340 --> 00:32:48,179
have a shape and that shape has an

893
00:32:46,440 --> 00:32:49,620
effect on the kinds of ideas that we can

894
00:32:48,179 --> 00:32:51,179
put with them

895
00:32:49,620 --> 00:32:53,100
I mean going back and through and

896
00:32:51,179 --> 00:32:55,620
digging through all of the things in

897
00:32:53,100 --> 00:32:57,539
your head can be confronting as hell

898
00:32:55,620 --> 00:32:58,980
but it can be really important in my

899
00:32:57,539 --> 00:33:00,779
case I was looking at the stories of a

900
00:32:58,980 --> 00:33:02,760
bunch of my friends

901
00:33:00,779 --> 00:33:04,380
um and that they were telling about

902
00:33:02,760 --> 00:33:07,260
their experience of discovering that

903
00:33:04,380 --> 00:33:09,899
they were or they weren't neurotypical

904
00:33:07,260 --> 00:33:13,440
and realizing that they mapped far too

905
00:33:09,899 --> 00:33:16,860
well onto what I thought was just a form

906
00:33:13,440 --> 00:33:19,200
of laziness or distractedness on my part

907
00:33:16,860 --> 00:33:21,120
um you know for other people it might be

908
00:33:19,200 --> 00:33:22,559
a realization that no matter how hard

909
00:33:21,120 --> 00:33:24,179
they try they just don't fit into the

910
00:33:22,559 --> 00:33:26,880
gender they were born with

911
00:33:24,179 --> 00:33:28,799
or even just a realization that they're

912
00:33:26,880 --> 00:33:30,840
in the wrong career

913
00:33:28,799 --> 00:33:32,399
um all of these things are ideas with

914
00:33:30,840 --> 00:33:34,019
shapes that have been built up around

915
00:33:32,399 --> 00:33:36,600
them and changing them it can be

916
00:33:34,019 --> 00:33:39,360
massively disruptive

917
00:33:36,600 --> 00:33:41,640
um but that's not the only thing that

918
00:33:39,360 --> 00:33:44,279
can go wrong if we fail to sort of

919
00:33:41,640 --> 00:33:46,519
examine the foundational ideas that are

920
00:33:44,279 --> 00:33:48,779
around us

921
00:33:46,519 --> 00:33:50,940
and because you can end up taking them

922
00:33:48,779 --> 00:33:53,760
for granted

923
00:33:50,940 --> 00:33:55,679
um we get used to things we get used to

924
00:33:53,760 --> 00:33:57,659
being able to get what we need when we

925
00:33:55,679 --> 00:33:59,460
need it from the supermarket

926
00:33:57,659 --> 00:34:01,679
we get used to being able to get

927
00:33:59,460 --> 00:34:03,960
together every year for conferences

928
00:34:01,679 --> 00:34:04,799
we get used to a whole lot of these

929
00:34:03,960 --> 00:34:07,940
things

930
00:34:04,799 --> 00:34:07,940
and then it stops

931
00:34:09,000 --> 00:34:11,119
um

932
00:34:12,000 --> 00:34:15,780
one of the most

933
00:34:13,619 --> 00:34:19,200
fascinating things about how the

934
00:34:15,780 --> 00:34:21,480
pandemic affected Society was just the

935
00:34:19,200 --> 00:34:25,020
exposure of how much of it had come to

936
00:34:21,480 --> 00:34:27,359
rely on global logistics

937
00:34:25,020 --> 00:34:29,820
um not only had a lot of manufacturing

938
00:34:27,359 --> 00:34:32,339
moved offshore

939
00:34:29,820 --> 00:34:34,200
but companies have become so allergic to

940
00:34:32,339 --> 00:34:35,280
keeping stock on hand both for reasons

941
00:34:34,200 --> 00:34:36,960
of not wanting to be exposed to

942
00:34:35,280 --> 00:34:38,220
inventory that might become hard to sell

943
00:34:36,960 --> 00:34:39,960
or because they didn't want to pay for

944
00:34:38,220 --> 00:34:42,720
the space to store it that they were

945
00:34:39,960 --> 00:34:44,399
just existentially dependent on the

946
00:34:42,720 --> 00:34:46,679
ability to move things around and when

947
00:34:44,399 --> 00:34:48,960
that stopped working they were stuffed

948
00:34:46,679 --> 00:34:50,820
you couldn't get tinned tomatoes not

949
00:34:48,960 --> 00:34:53,960
because of a lack of tomatoes but

950
00:34:50,820 --> 00:34:53,960
because of a lack of tins

951
00:34:54,359 --> 00:34:57,300
industry had assumed that Global

952
00:34:56,159 --> 00:35:00,119
Logistics was something that couldn't

953
00:34:57,300 --> 00:35:02,280
fail because up till now it never had

954
00:35:00,119 --> 00:35:04,920
and then like a container ship getting

955
00:35:02,280 --> 00:35:06,119
wedged in a canal reality came along and

956
00:35:04,920 --> 00:35:07,740
pantsed them

957
00:35:06,119 --> 00:35:09,660
but

958
00:35:07,740 --> 00:35:11,760
there was an odd positive to this kind

959
00:35:09,660 --> 00:35:13,200
of things changing as well

960
00:35:11,760 --> 00:35:14,579
we suddenly discovered that social

961
00:35:13,200 --> 00:35:16,440
welfare was something we could actually

962
00:35:14,579 --> 00:35:18,540
do

963
00:35:16,440 --> 00:35:20,760
um I mean we've probably forgot it again

964
00:35:18,540 --> 00:35:23,880
or at least anyone involved in it tried

965
00:35:20,760 --> 00:35:25,400
to and absolutely don't remind the labor

966
00:35:23,880 --> 00:35:28,200
party that we could do that

967
00:35:25,400 --> 00:35:30,660
but in general

968
00:35:28,200 --> 00:35:33,420
you know there's this nasty sense in the

969
00:35:30,660 --> 00:35:35,400
capital classes that were in the end

970
00:35:33,420 --> 00:35:38,099
game of a particularly nasty round of

971
00:35:35,400 --> 00:35:40,440
Jenga where they keep trying and often

972
00:35:38,099 --> 00:35:42,079
succeeding and knocking out parts that

973
00:35:40,440 --> 00:35:44,339
they don't think that we need anymore

974
00:35:42,079 --> 00:35:46,140
because they prevent them from hoarding

975
00:35:44,339 --> 00:35:47,400
more and more resources and assuming and

976
00:35:46,140 --> 00:35:49,020
they just assumed that the whole edifice

977
00:35:47,400 --> 00:35:50,880
isn't going to collapse or that if it

978
00:35:49,020 --> 00:35:52,380
does that they've got enough resources

979
00:35:50,880 --> 00:35:53,700
that they that'll insulate them from a

980
00:35:52,380 --> 00:35:56,099
result

981
00:35:53,700 --> 00:35:57,020
I'd like to hope that we can reverse

982
00:35:56,099 --> 00:36:00,740
this

983
00:35:57,020 --> 00:36:06,020
but we have to face our own assumptions

984
00:36:00,740 --> 00:36:06,020
that these things can't or won't change

985
00:36:06,180 --> 00:36:09,660
um and I'm thinking very loudly at the

986
00:36:08,220 --> 00:36:11,460
Australian labor party but we'll get

987
00:36:09,660 --> 00:36:13,740
back to them in a sec

988
00:36:11,460 --> 00:36:15,920
um because I've got one more bit to go

989
00:36:13,740 --> 00:36:15,920
through

990
00:36:17,099 --> 00:36:22,320
so another thing that we can show using

991
00:36:20,760 --> 00:36:23,640
the whole kind of notion of ideas is

992
00:36:22,320 --> 00:36:26,880
shapes

993
00:36:23,640 --> 00:36:28,560
is the way that the same idea or similar

994
00:36:26,880 --> 00:36:30,960
ideas can show up in multiple places

995
00:36:28,560 --> 00:36:33,240
independently there are some shapes that

996
00:36:30,960 --> 00:36:34,560
are very similar and you can also have

997
00:36:33,240 --> 00:36:36,000
shapes that end up constructing

998
00:36:34,560 --> 00:36:38,880
themselves and leaving a similarly

999
00:36:36,000 --> 00:36:41,480
shaped Gap and at that point a similarly

1000
00:36:38,880 --> 00:36:44,940
shaped idea can slot in there

1001
00:36:41,480 --> 00:36:46,200
and you know that's somewhat an obvious

1002
00:36:44,940 --> 00:36:47,820
thought but I think there's a couple of

1003
00:36:46,200 --> 00:36:49,560
things that can be drawn out of it the

1004
00:36:47,820 --> 00:36:52,380
first one that is talking about when you

1005
00:36:49,560 --> 00:36:53,339
decide to burn everything down

1006
00:36:52,380 --> 00:36:55,980
um

1007
00:36:53,339 --> 00:36:57,720
deciding that something is not or is no

1008
00:36:55,980 --> 00:36:59,760
longer fit for purpose and just having

1009
00:36:57,720 --> 00:37:02,480
the ability or the resources is to Yeet

1010
00:36:59,760 --> 00:37:04,800
the thing feels great

1011
00:37:02,480 --> 00:37:06,540
until you're standing among the rubble

1012
00:37:04,800 --> 00:37:07,560
of the thing you just demolished and

1013
00:37:06,540 --> 00:37:09,599
then you realize you have to start

1014
00:37:07,560 --> 00:37:12,599
building its replacement

1015
00:37:09,599 --> 00:37:15,599
because all of the ideas that influence

1016
00:37:12,599 --> 00:37:16,740
the thing you just you know torched are

1017
00:37:15,599 --> 00:37:20,099
still there

1018
00:37:16,740 --> 00:37:22,800
and they have the potential to influence

1019
00:37:20,099 --> 00:37:24,660
what you build to replace it that may

1020
00:37:22,800 --> 00:37:26,099
not be a problem sometimes just you know

1021
00:37:24,660 --> 00:37:27,839
taking something out and fixing the

1022
00:37:26,099 --> 00:37:31,140
internal plumbing and putting it back is

1023
00:37:27,839 --> 00:37:34,740
you know worthwhile in and of itself but

1024
00:37:31,140 --> 00:37:36,540
there's a lot of cases where you know in

1025
00:37:34,740 --> 00:37:39,480
like socio-political revolutions have

1026
00:37:36,540 --> 00:37:41,040
happened and entire you know systems

1027
00:37:39,480 --> 00:37:42,599
have been knocked over and then replaced

1028
00:37:41,040 --> 00:37:45,000
by the exact same thing but with

1029
00:37:42,599 --> 00:37:47,760
different people at the top

1030
00:37:45,000 --> 00:37:49,740
um if you want more info more examples

1031
00:37:47,760 --> 00:37:50,940
of that there's an entire podcast about

1032
00:37:49,740 --> 00:37:53,460
it

1033
00:37:50,940 --> 00:37:55,260
um really worth listening to

1034
00:37:53,460 --> 00:37:57,300
um even if you don't listen to all of it

1035
00:37:55,260 --> 00:37:59,280
the first part of the last season gives

1036
00:37:57,300 --> 00:38:01,859
one of the best explainers of Marx and

1037
00:37:59,280 --> 00:38:05,220
Marxism I've ever come across

1038
00:38:01,859 --> 00:38:06,359
um but there's a more pernicious form of

1039
00:38:05,220 --> 00:38:08,400
this kind of parallel construction

1040
00:38:06,359 --> 00:38:11,579
though and it's where someone helps you

1041
00:38:08,400 --> 00:38:13,820
construct or indeed obstruct the space

1042
00:38:11,579 --> 00:38:16,920
for an idea to form

1043
00:38:13,820 --> 00:38:18,720
often Without You realizing it and both

1044
00:38:16,920 --> 00:38:19,640
forms are endemic especially in the

1045
00:38:18,720 --> 00:38:22,560
media

1046
00:38:19,640 --> 00:38:24,420
but let I'm going to talk about the

1047
00:38:22,560 --> 00:38:26,400
obstructive form of it first using the

1048
00:38:24,420 --> 00:38:29,339
language of information security

1049
00:38:26,400 --> 00:38:32,520
uh let's say you have an idea

1050
00:38:29,339 --> 00:38:34,440
an idea can be a complicated thing and

1051
00:38:32,520 --> 00:38:36,780
especially communicating that idea can

1052
00:38:34,440 --> 00:38:39,119
be tricky the way we tend to deal with

1053
00:38:36,780 --> 00:38:41,579
this is that we give it a name

1054
00:38:39,119 --> 00:38:42,900
and after we've explained it once we

1055
00:38:41,579 --> 00:38:43,980
don't have to explain it again we can

1056
00:38:42,900 --> 00:38:45,960
just give you the name and you'll know

1057
00:38:43,980 --> 00:38:49,079
what I'm talking about

1058
00:38:45,960 --> 00:38:50,940
um that the trouble is that when the

1059
00:38:49,079 --> 00:38:53,339
idea becomes more widespread especially

1060
00:38:50,940 --> 00:38:55,680
within a group people start using the

1061
00:38:53,339 --> 00:38:59,339
name assuming that people know the

1062
00:38:55,680 --> 00:39:00,780
detail behind it and I suspected anyone

1063
00:38:59,339 --> 00:39:02,339
with any kind of exposure to offensive

1064
00:39:00,780 --> 00:39:04,079
infosec practice is starting to see

1065
00:39:02,339 --> 00:39:06,599
where this might be going because yes we

1066
00:39:04,079 --> 00:39:08,160
just made a pointer

1067
00:39:06,599 --> 00:39:09,900
and in the head of anyone who hasn't

1068
00:39:08,160 --> 00:39:11,760
been given the full detail of what we're

1069
00:39:09,900 --> 00:39:13,079
talking about we've got a dangling

1070
00:39:11,760 --> 00:39:15,380
pointer

1071
00:39:13,079 --> 00:39:17,400
and Dems be dangerous

1072
00:39:15,380 --> 00:39:18,599
especially an environment where you're

1073
00:39:17,400 --> 00:39:21,780
constantly having to evaluate

1074
00:39:18,599 --> 00:39:22,980
potentially malicious inputs

1075
00:39:21,780 --> 00:39:25,740
um

1076
00:39:22,980 --> 00:39:27,260
this also tends to look a bit like DNS

1077
00:39:25,740 --> 00:39:29,460
cache poisoning attack

1078
00:39:27,260 --> 00:39:32,099
where an attacker tries to plant

1079
00:39:29,460 --> 00:39:35,640
incorrect records into for example a DNS

1080
00:39:32,099 --> 00:39:37,200
resolver so that when you go to a given

1081
00:39:35,640 --> 00:39:40,079
website you end up going to the attacker

1082
00:39:37,200 --> 00:39:43,440
controlled version instead and the human

1083
00:39:40,079 --> 00:39:45,540
space equivalent goes a bit like this

1084
00:39:43,440 --> 00:39:47,700
um socio-political ideas are both a

1085
00:39:45,540 --> 00:39:49,619
nuanced detailed and often contentious

1086
00:39:47,700 --> 00:39:50,880
and let's say for example that you feel

1087
00:39:49,619 --> 00:39:52,500
that successive ways of a story

1088
00:39:50,880 --> 00:39:54,420
austerity-driven cuts to Social Services

1089
00:39:52,500 --> 00:39:56,460
have engendered an over-reliance on

1090
00:39:54,420 --> 00:39:58,980
police forces as a kind of universal

1091
00:39:56,460 --> 00:40:00,420
First Responders despite the fact that

1092
00:39:58,980 --> 00:40:01,920
not only do they often not have the

1093
00:40:00,420 --> 00:40:03,420
right training to deal with a lot of the

1094
00:40:01,920 --> 00:40:05,579
situations in which you need a first

1095
00:40:03,420 --> 00:40:07,680
responder but they also have

1096
00:40:05,579 --> 00:40:10,200
well-documented biases against certain

1097
00:40:07,680 --> 00:40:12,599
sectors of the population that lead to

1098
00:40:10,200 --> 00:40:15,660
not great outcomes let's say you're

1099
00:40:12,599 --> 00:40:17,820
incensed that the police budgets at best

1100
00:40:15,660 --> 00:40:19,800
seem to be quarantined from these cuts

1101
00:40:17,820 --> 00:40:21,660
and at worst seem to be hoovering up all

1102
00:40:19,800 --> 00:40:23,579
the money that used to support a much

1103
00:40:21,660 --> 00:40:26,280
broader set of social supports you might

1104
00:40:23,579 --> 00:40:28,920
Advocate to for argument's sake divert

1105
00:40:26,280 --> 00:40:30,720
police funding to Social Services or you

1106
00:40:28,920 --> 00:40:33,359
might decide to make it even shorter and

1107
00:40:30,720 --> 00:40:35,160
just say defund the police

1108
00:40:33,359 --> 00:40:37,859
um

1109
00:40:35,160 --> 00:40:40,740
which is a great pithy name for a thing

1110
00:40:37,859 --> 00:40:43,560
it charts well you know good on signs

1111
00:40:40,740 --> 00:40:46,079
but the trouble is that

1112
00:40:43,560 --> 00:40:47,720
a whole bunch of um

1113
00:40:46,079 --> 00:40:50,760
chuckle heads

1114
00:40:47,720 --> 00:40:53,180
with columns or TV shows or YouTube

1115
00:40:50,760 --> 00:40:56,400
channels or Twitter accounts or whatever

1116
00:40:53,180 --> 00:40:58,680
X account

1117
00:40:56,400 --> 00:40:59,760
um or say what you really meant is that

1118
00:40:58,680 --> 00:41:01,619
we should just get rid of the police

1119
00:40:59,760 --> 00:41:03,420
entirely we should not have that kind of

1120
00:41:01,619 --> 00:41:05,220
law enforcement arm that must mean

1121
00:41:03,420 --> 00:41:07,140
you're a criminal or you love criminals

1122
00:41:05,220 --> 00:41:09,599
or something like that but the critical

1123
00:41:07,140 --> 00:41:11,940
idea here is that that idea with that

1124
00:41:09,599 --> 00:41:14,579
name is now in a whole bunch of people's

1125
00:41:11,940 --> 00:41:17,280
heads and so you can't use that name

1126
00:41:14,579 --> 00:41:20,820
anymore without not only having to

1127
00:41:17,280 --> 00:41:22,680
explain all of the detail behind it but

1128
00:41:20,820 --> 00:41:24,359
also defending yourself from but doesn't

1129
00:41:22,680 --> 00:41:28,320
that make you a criminal

1130
00:41:24,359 --> 00:41:30,660
and this gets used over and over again

1131
00:41:28,320 --> 00:41:31,859
black lives matter is racist don't you

1132
00:41:30,660 --> 00:41:34,500
know

1133
00:41:31,859 --> 00:41:37,140
trans want trans people want more rights

1134
00:41:34,500 --> 00:41:39,720
than everyone else don't you know

1135
00:41:37,140 --> 00:41:41,760
and unfortunately

1136
00:41:39,720 --> 00:41:43,800
the only way that I know to get around

1137
00:41:41,760 --> 00:41:45,800
this is to actually give people the

1138
00:41:43,800 --> 00:41:48,839
detailed version of what's behind this

1139
00:41:45,800 --> 00:41:51,920
and and in a lot of cases these people

1140
00:41:48,839 --> 00:41:54,720
just don't have enough of the context

1141
00:41:51,920 --> 00:41:56,880
and now I'm going to do the bit of this

1142
00:41:54,720 --> 00:41:58,980
talk I really don't like very much

1143
00:41:56,880 --> 00:42:00,300
let's say you've not met many people

1144
00:41:58,980 --> 00:42:03,480
from Adelaide

1145
00:42:00,300 --> 00:42:05,099
you've got no idea what adeladians are

1146
00:42:03,480 --> 00:42:06,540
um you're a bit suspicious of them

1147
00:42:05,099 --> 00:42:08,339
though maybe you're from Sydney or

1148
00:42:06,540 --> 00:42:09,660
something I don't know you hear they

1149
00:42:08,339 --> 00:42:12,480
drink a lot of wine for some reason

1150
00:42:09,660 --> 00:42:14,880
they've got way too many churches they

1151
00:42:12,480 --> 00:42:17,700
worship these weird balls

1152
00:42:14,880 --> 00:42:19,740
um and then you're on YouTube one day

1153
00:42:17,700 --> 00:42:23,160
and the algorithm coughs up a video to

1154
00:42:19,740 --> 00:42:25,440
you on the add a lady and agenda

1155
00:42:23,160 --> 00:42:27,660
um some person in a well-lit studio with

1156
00:42:25,440 --> 00:42:28,980
a very expensive looking microphone

1157
00:42:27,660 --> 00:42:30,960
um and but a very kind of I'm just a

1158
00:42:28,980 --> 00:42:33,060
regular person kind of vibe tells you to

1159
00:42:30,960 --> 00:42:34,560
like share subscribe don't forget to hit

1160
00:42:33,060 --> 00:42:35,700
the Bell to receive notifications of new

1161
00:42:34,560 --> 00:42:38,099
videos but did you know that all

1162
00:42:35,700 --> 00:42:40,560
adelaideans are serial killers

1163
00:42:38,099 --> 00:42:42,119
they love the stuff they just do it all

1164
00:42:40,560 --> 00:42:44,220
the time they want to make it normal

1165
00:42:42,119 --> 00:42:46,500
they want everyone to do it the vote

1166
00:42:44,220 --> 00:42:49,020
left is just won't let us say this

1167
00:42:46,500 --> 00:42:51,000
anymore we have to put a stop to it I

1168
00:42:49,020 --> 00:42:52,619
used to be a lefty but now I vote for

1169
00:42:51,000 --> 00:42:53,880
Peter Dutton it's not that I like him

1170
00:42:52,619 --> 00:42:55,920
you see it's just that it's worth

1171
00:42:53,880 --> 00:42:58,020
accepting some of this other stuff that

1172
00:42:55,920 --> 00:43:00,980
he does as long as we can deal with the

1173
00:42:58,020 --> 00:43:00,980
adelaidean Menace

1174
00:43:02,339 --> 00:43:07,980
I'm really sorry to anyone who's been on

1175
00:43:05,040 --> 00:43:10,980
the receiving end of anything like that

1176
00:43:07,980 --> 00:43:13,020
um like that is the culture of War stuff

1177
00:43:10,980 --> 00:43:13,859
I was warning about

1178
00:43:13,020 --> 00:43:16,079
um

1179
00:43:13,859 --> 00:43:17,579
and it's superficially funny when you do

1180
00:43:16,079 --> 00:43:20,640
it like that

1181
00:43:17,579 --> 00:43:23,280
but it's the exact Playbook run by so

1182
00:43:20,640 --> 00:43:25,980
many people with platforms from

1183
00:43:23,280 --> 00:43:28,740
colonists and editors at newspapers to

1184
00:43:25,980 --> 00:43:30,300
cross media paleoliths like most of Sky

1185
00:43:28,740 --> 00:43:30,960
News

1186
00:43:30,300 --> 00:43:33,540
um

1187
00:43:30,960 --> 00:43:35,460
to YouTube grifters and everyone in

1188
00:43:33,540 --> 00:43:37,200
between it's designed to encourage

1189
00:43:35,460 --> 00:43:39,420
people especially those with

1190
00:43:37,200 --> 00:43:41,880
pre-existing concerns whether they're

1191
00:43:39,420 --> 00:43:44,460
reasonable or not to start their Journey

1192
00:43:41,880 --> 00:43:46,200
down the path to radicalization by

1193
00:43:44,460 --> 00:43:49,980
taking those concerns and building on

1194
00:43:46,200 --> 00:43:52,740
them with lies and half truths

1195
00:43:49,980 --> 00:43:54,240
um there's multiple tragedies here one

1196
00:43:52,740 --> 00:43:56,579
is that these people don't need to go

1197
00:43:54,240 --> 00:43:58,560
too far down that path to get to the

1198
00:43:56,579 --> 00:44:00,540
point where they will at the very least

1199
00:43:58,560 --> 00:44:02,339
stand by

1200
00:44:00,540 --> 00:44:04,859
while those who are far more committed

1201
00:44:02,339 --> 00:44:08,040
to this plan

1202
00:44:04,859 --> 00:44:10,200
um do the Nastia work another is that if

1203
00:44:08,040 --> 00:44:12,720
they talk to someone else about their

1204
00:44:10,200 --> 00:44:15,000
concerns first then those concerns might

1205
00:44:12,720 --> 00:44:17,220
have been delayed instead they get used

1206
00:44:15,000 --> 00:44:19,260
as a foundation upon which other ideas

1207
00:44:17,220 --> 00:44:21,599
are built and it can be really hard to

1208
00:44:19,260 --> 00:44:23,579
get at those foundational ideas again to

1209
00:44:21,599 --> 00:44:25,859
take to get them out

1210
00:44:23,579 --> 00:44:27,780
and the biggest tragedy here is that it

1211
00:44:25,859 --> 00:44:29,520
means they either have or have been

1212
00:44:27,780 --> 00:44:32,240
taught the tendency to view certain

1213
00:44:29,520 --> 00:44:35,760
groups as less worthy or less human

1214
00:44:32,240 --> 00:44:37,619
based on who they are

1215
00:44:35,760 --> 00:44:39,599
it may be a bit simplistic but there's

1216
00:44:37,619 --> 00:44:41,220
at least an axis along which a given

1217
00:44:39,599 --> 00:44:44,760
aspect of someone is who they are versus

1218
00:44:41,220 --> 00:44:48,359
what they do they are a man but they do

1219
00:44:44,760 --> 00:44:52,220
programming they are cisgender but they

1220
00:44:48,359 --> 00:44:55,500
do big thinky opinion piece France

1221
00:44:52,220 --> 00:44:56,880
at least in my world view it's okay to

1222
00:44:55,500 --> 00:44:58,920
judge people for their actions

1223
00:44:56,880 --> 00:45:01,079
especially when it's clear that those

1224
00:44:58,920 --> 00:45:03,720
actions were by their own choice

1225
00:45:01,079 --> 00:45:06,540
it's not fair to judge people based

1226
00:45:03,720 --> 00:45:08,940
purely on who they are or are not

1227
00:45:06,540 --> 00:45:11,220
and all of these groups are encouraging

1228
00:45:08,940 --> 00:45:13,500
you to judge people based on who they

1229
00:45:11,220 --> 00:45:15,660
are they'll find one or two stories that

1230
00:45:13,500 --> 00:45:17,579
show someone who they claim and can

1231
00:45:15,660 --> 00:45:19,859
claim as a member of that group doing

1232
00:45:17,579 --> 00:45:21,900
something bad and try to use that to

1233
00:45:19,859 --> 00:45:23,520
construct ideas in people's heads that

1234
00:45:21,900 --> 00:45:26,700
encourage them to see everyone in that

1235
00:45:23,520 --> 00:45:29,099
group are somehow bad or wrong

1236
00:45:26,700 --> 00:45:31,140
and the despondency that this generates

1237
00:45:29,099 --> 00:45:33,540
can lead to lead us to forget that

1238
00:45:31,140 --> 00:45:35,880
progress has been made

1239
00:45:33,540 --> 00:45:37,380
things now at least when it comes to the

1240
00:45:35,880 --> 00:45:40,440
rights of people who aren't at the

1241
00:45:37,380 --> 00:45:43,140
center of white patriarchy are better

1242
00:45:40,440 --> 00:45:45,900
than they were when I was a kid but all

1243
00:45:43,140 --> 00:45:48,599
of that progress is one of those things

1244
00:45:45,900 --> 00:45:50,640
that we can get used to and forget that

1245
00:45:48,599 --> 00:45:53,099
it's there and that it needs to be

1246
00:45:50,640 --> 00:45:54,540
defended and pushed further

1247
00:45:53,099 --> 00:45:57,420
and that means we need to bring as many

1248
00:45:54,540 --> 00:45:58,800
of these people on side as we can even

1249
00:45:57,420 --> 00:46:00,240
if that means trying to figure out how

1250
00:45:58,800 --> 00:46:02,460
to break them away from the bad ideas

1251
00:46:00,240 --> 00:46:04,140
that they've become stuck in

1252
00:46:02,460 --> 00:46:06,240
which leads me to another one of these

1253
00:46:04,140 --> 00:46:09,420
ideas with a neat name but a malleable

1254
00:46:06,240 --> 00:46:11,460
concept meet them where they live

1255
00:46:09,420 --> 00:46:15,540
most of the people who use this phrase

1256
00:46:11,460 --> 00:46:17,819
use it to mean more centrism please

1257
00:46:15,540 --> 00:46:19,200
um there existed that they did they say

1258
00:46:17,819 --> 00:46:21,119
that there exists a body of people who

1259
00:46:19,200 --> 00:46:22,740
are turned off by the constant debate

1260
00:46:21,119 --> 00:46:24,119
and especially by the implied strident

1261
00:46:22,740 --> 00:46:25,920
demands of the woke left or anything

1262
00:46:24,119 --> 00:46:27,780
like that and their formulation the way

1263
00:46:25,920 --> 00:46:29,579
to re-engage them is that we should stop

1264
00:46:27,780 --> 00:46:30,960
demanding respect and rights for these

1265
00:46:29,579 --> 00:46:32,880
people and focus purely on things like

1266
00:46:30,960 --> 00:46:35,040
economics

1267
00:46:32,880 --> 00:46:36,540
um but I say we steal that name

1268
00:46:35,040 --> 00:46:37,859
because I think the way that we need to

1269
00:46:36,540 --> 00:46:39,240
meet people where they live is to go to

1270
00:46:37,859 --> 00:46:41,220
them and treat them as humans rather

1271
00:46:39,240 --> 00:46:43,260
than demographics

1272
00:46:41,220 --> 00:46:44,819
these people have their own concerns and

1273
00:46:43,260 --> 00:46:47,339
may indeed be turned off by the strident

1274
00:46:44,819 --> 00:46:50,400
nature of socio-political debate when

1275
00:46:47,339 --> 00:46:52,079
all it is to them is abstract things

1276
00:46:50,400 --> 00:46:53,700
I have to believe that these people can

1277
00:46:52,079 --> 00:46:55,680
be reached though and in the same way

1278
00:46:53,700 --> 00:46:57,780
that it's a win for the regressive types

1279
00:46:55,680 --> 00:47:00,359
to merely get people to tune out and sit

1280
00:46:57,780 --> 00:47:02,339
things out it's a win for us to purely

1281
00:47:00,359 --> 00:47:04,920
get them to care

1282
00:47:02,339 --> 00:47:06,359
to see all humans as humans even the

1283
00:47:04,920 --> 00:47:07,260
ones they might find challenging to

1284
00:47:06,359 --> 00:47:08,940
handle

1285
00:47:07,260 --> 00:47:11,160
and this to me is one of the great

1286
00:47:08,940 --> 00:47:12,540
failures of the emotionally true the

1287
00:47:11,160 --> 00:47:14,640
traditional notionally Progressive

1288
00:47:12,540 --> 00:47:17,160
political parties

1289
00:47:14,640 --> 00:47:18,420
labor the Democrats in the US so on so

1290
00:47:17,160 --> 00:47:20,460
forth they've got so used to campaigning

1291
00:47:18,420 --> 00:47:22,140
by demographics and the tools that

1292
00:47:20,460 --> 00:47:24,599
they're used in that that they've

1293
00:47:22,140 --> 00:47:26,819
mistaken the abstraction the idea for

1294
00:47:24,599 --> 00:47:27,780
reality they're so locked into not only

1295
00:47:26,819 --> 00:47:29,400
the shapes within their own

1296
00:47:27,780 --> 00:47:30,780
organizations but they've locked

1297
00:47:29,400 --> 00:47:32,700
themselves into a way of thinking about

1298
00:47:30,780 --> 00:47:35,520
their role in society and their ability

1299
00:47:32,700 --> 00:47:37,680
to affect change that's drifting further

1300
00:47:35,520 --> 00:47:39,720
and further from the reality of what

1301
00:47:37,680 --> 00:47:42,540
Society actually wants

1302
00:47:39,720 --> 00:47:44,760
and at some point reality is

1303
00:47:42,540 --> 00:47:45,900
figuratively going to punch him in the

1304
00:47:44,760 --> 00:47:47,020
face

1305
00:47:45,900 --> 00:48:01,409
thank you

1306
00:47:47,020 --> 00:48:01,409
[Applause]

1307
00:48:02,760 --> 00:48:08,579
you know that was fantastic

1308
00:48:06,000 --> 00:48:11,640
thank you so so much for uh coming today

1309
00:48:08,579 --> 00:48:13,920
and sharing all of that with us

1310
00:48:11,640 --> 00:48:15,420
um and I look forward to

1311
00:48:13,920 --> 00:48:17,220
watching that again and sharing it with

1312
00:48:15,420 --> 00:48:19,500
everybody else as well because I think

1313
00:48:17,220 --> 00:48:20,760
it was very needed so thank you and

1314
00:48:19,500 --> 00:48:22,560
thank you on behalf of the event thank

1315
00:48:20,760 --> 00:48:25,460
you very much

1316
00:48:22,560 --> 00:48:25,460
better rice

1317
00:48:27,550 --> 00:48:30,499
[Applause]