1
00:00:12,240 --> 00:00:17,199
hello everyone it's lightning

2
00:00:14,799 --> 00:00:18,800
talks time uh finally

3
00:00:17,199 --> 00:00:21,359
uh hey everyone i'm christopher

4
00:00:18,800 --> 00:00:23,519
neugebauer i am coming to you from the

5
00:00:21,359 --> 00:00:26,960
unseated land of the coast miwok people

6
00:00:23,519 --> 00:00:28,800
here in petaluma california uh still uh

7
00:00:26,960 --> 00:00:30,960
annoyingly prevented from being on the

8
00:00:28,800 --> 00:00:33,120
correct time zone for lightning talks

9
00:00:30,960 --> 00:00:34,559
and so as with last year i've adjusted

10
00:00:33,120 --> 00:00:36,880
the timing of lightning talks to my

11
00:00:34,559 --> 00:00:38,399
convenience and not yours thank you all

12
00:00:36,880 --> 00:00:39,920
for joining uh this is going to be

13
00:00:38,399 --> 00:00:41,840
really exciting

14
00:00:39,920 --> 00:00:45,280
today we've got a couple of people

15
00:00:41,840 --> 00:00:47,440
joining me on stage uh libby hello i'm

16
00:00:45,280 --> 00:00:49,840
libby i am trapped in an apartment in

17
00:00:47,440 --> 00:00:51,760
the nunawal country aka the ict in

18
00:00:49,840 --> 00:00:54,239
canberra it is nice to join you and for

19
00:00:51,760 --> 00:00:55,840
a split second you saw my cat bobo uh

20
00:00:54,239 --> 00:00:57,520
prepare for ultimate chaos he will

21
00:00:55,840 --> 00:00:58,800
probably make appearances every now and

22
00:00:57,520 --> 00:01:01,440
then

23
00:00:58,800 --> 00:01:05,199
uh and also joining is our friend

24
00:01:01,440 --> 00:01:05,199
patrick the distant bodied clock

25
00:01:05,840 --> 00:01:10,640
hi patrick thank you thank you

26
00:01:07,680 --> 00:01:10,640
clock the distance

27
00:01:10,720 --> 00:01:17,119
okay uh so as is usually the case we

28
00:01:13,760 --> 00:01:19,840
have a bunch of uh speakers giving

29
00:01:17,119 --> 00:01:21,600
five-minute talks on various topics

30
00:01:19,840 --> 00:01:23,920
which will hopefully be of interest to

31
00:01:21,600 --> 00:01:25,520
all of you uh

32
00:01:23,920 --> 00:01:27,280
we can't really do the finger applause

33
00:01:25,520 --> 00:01:30,000
thing that we do uh when we're all in

34
00:01:27,280 --> 00:01:32,479
person but i will use my trusty yeet

35
00:01:30,000 --> 00:01:34,159
button if people go over five minutes so

36
00:01:32,479 --> 00:01:36,240
uh please keep your talks to five

37
00:01:34,159 --> 00:01:38,240
minutes everybody who's presenting

38
00:01:36,240 --> 00:01:41,360
uh otherwise yeah we won't hear

39
00:01:38,240 --> 00:01:44,720
everything you have to say

40
00:01:41,360 --> 00:01:46,560
our first presenter is tyson klug who is

41
00:01:44,720 --> 00:01:49,920
going to be talking about putting python

42
00:01:46,560 --> 00:01:49,920
in all of the files

43
00:01:50,560 --> 00:01:56,880
hi pycon um in true lightning talk

44
00:01:54,240 --> 00:01:58,560
spirit my tool broke two minutes ago and

45
00:01:56,880 --> 00:02:03,439
i fixed it again so

46
00:01:58,560 --> 00:02:04,719
paid um or yeet all the files this is a

47
00:02:03,439 --> 00:02:06,399
personal project that i started

48
00:02:04,719 --> 00:02:10,560
yesterday lots of fun

49
00:02:06,399 --> 00:02:13,040
um 20 years earlier well we used to use

50
00:02:10,560 --> 00:02:15,200
php and we used to be able to just

51
00:02:13,040 --> 00:02:17,760
insert little bits of code inside a

52
00:02:15,200 --> 00:02:19,680
bigger document it's a really cool trick

53
00:02:17,760 --> 00:02:22,080
because it means that you can just write

54
00:02:19,680 --> 00:02:25,120
for example html or postscript or

55
00:02:22,080 --> 00:02:28,480
whatever you want and then just put code

56
00:02:25,120 --> 00:02:30,480
where you need it which is awesome um

57
00:02:28,480 --> 00:02:32,239
python doesn't really support this i

58
00:02:30,480 --> 00:02:34,319
mean maybe a good thing maybe a bad

59
00:02:32,239 --> 00:02:37,920
thing but in my case i wanted to because

60
00:02:34,319 --> 00:02:39,599
i just wanted to stick in a date into a

61
00:02:37,920 --> 00:02:40,959
postscript file but postscript doesn't

62
00:02:39,599 --> 00:02:41,920
support that

63
00:02:40,959 --> 00:02:44,959
so

64
00:02:41,920 --> 00:02:47,680
yeah php was an option but

65
00:02:44,959 --> 00:02:49,040
i don't want to use php and i certainly

66
00:02:47,680 --> 00:02:50,480
don't want to be writing lots and lots

67
00:02:49,040 --> 00:02:53,440
of python if it's just inserting the

68
00:02:50,480 --> 00:02:56,720
date so uh yeah like i could write lots

69
00:02:53,440 --> 00:02:58,800
of code to do report labing to um inject

70
00:02:56,720 --> 00:03:02,159
a date into my postscript output and a

71
00:02:58,800 --> 00:03:03,120
pdf um or i could just fix this little

72
00:03:02,159 --> 00:03:06,159
thing

73
00:03:03,120 --> 00:03:07,840
so need more curly boy let's see what

74
00:03:06,159 --> 00:03:11,680
can we do

75
00:03:07,840 --> 00:03:13,120
the program um all we need to do is use

76
00:03:11,680 --> 00:03:17,040
paint

77
00:03:13,120 --> 00:03:21,280
uh let's see so it just takes a

78
00:03:17,040 --> 00:03:25,440
file and it turns all of your

79
00:03:21,280 --> 00:03:28,239
source content into compiled python

80
00:03:25,440 --> 00:03:31,360
uh what sort of compiled python well

81
00:03:28,239 --> 00:03:33,200
it starts in like um just like python

82
00:03:31,360 --> 00:03:36,480
it creates this module where it just

83
00:03:33,200 --> 00:03:38,640
prints everything out so nice and simple

84
00:03:36,480 --> 00:03:41,680
um but you then you can put a curly boy

85
00:03:38,640 --> 00:03:43,280
in there woohoo more curly boy uh and

86
00:03:41,680 --> 00:03:46,400
here we go we've got a curly brace and

87
00:03:43,280 --> 00:03:48,400
two snakes that gets us into just plain

88
00:03:46,400 --> 00:03:52,000
python mode spit it out exactly as you

89
00:03:48,400 --> 00:03:54,640
saw it um there's also support for echo

90
00:03:52,000 --> 00:03:57,360
mode which just prints whatever you want

91
00:03:54,640 --> 00:04:00,400
so we can see here that i said oh yeah

92
00:03:57,360 --> 00:04:03,519
assign hello world to some text and then

93
00:04:00,400 --> 00:04:05,280
i told it to use echo mode to output the

94
00:04:03,519 --> 00:04:07,120
text nice and easy

95
00:04:05,280 --> 00:04:09,920
well that means you can do some cool

96
00:04:07,120 --> 00:04:11,760
stuff so you can just use valid python

97
00:04:09,920 --> 00:04:15,200
anywhere in your file

98
00:04:11,760 --> 00:04:17,120
well what is this useful for yeah i was

99
00:04:15,200 --> 00:04:20,479
trying to scratch an itch as i said i

100
00:04:17,120 --> 00:04:22,400
was trying to write some um postscript

101
00:04:20,479 --> 00:04:24,720
yeah i know postscript

102
00:04:22,400 --> 00:04:27,759
so here it was it's basically trying to

103
00:04:24,720 --> 00:04:30,720
put like a watermark onto a file and i

104
00:04:27,759 --> 00:04:31,520
needed to have the date on there

105
00:04:30,720 --> 00:04:34,800
so

106
00:04:31,520 --> 00:04:37,600
here we are i'm using uh let's see so

107
00:04:34,800 --> 00:04:41,120
i'm from date time import date time

108
00:04:37,600 --> 00:04:43,759
and then um that's imported

109
00:04:41,120 --> 00:04:46,560
and now all i have to do is say

110
00:04:43,759 --> 00:04:48,639
use echo mode and put the date in so

111
00:04:46,560 --> 00:04:50,160
let's see does this do what i want it to

112
00:04:48,639 --> 00:04:52,479
do

113
00:04:50,160 --> 00:04:52,479
okay

114
00:04:54,400 --> 00:05:00,720
i run it

115
00:04:56,639 --> 00:05:04,160
hey date time is not defined

116
00:05:00,720 --> 00:05:07,280
anyway uh yes it did actually run um

117
00:05:04,160 --> 00:05:09,120
but broken as always uh so other cool

118
00:05:07,280 --> 00:05:10,320
things let's actually dive into the code

119
00:05:09,120 --> 00:05:11,840
a little bit we've only got a minute and

120
00:05:10,320 --> 00:05:13,919
a half left

121
00:05:11,840 --> 00:05:16,240
what have we got so it's a little state

122
00:05:13,919 --> 00:05:17,919
machine it is actually an encoder uh

123
00:05:16,240 --> 00:05:21,680
sorry a decoder so you can use it to

124
00:05:17,919 --> 00:05:23,360
decode bytes and it decodes bytes into

125
00:05:21,680 --> 00:05:25,520
python

126
00:05:23,360 --> 00:05:29,600
it registers its codec

127
00:05:25,520 --> 00:05:31,759
it does some cool stuff it has a

128
00:05:29,600 --> 00:05:32,639
loader to do um

129
00:05:31,759 --> 00:05:35,520
to

130
00:05:32,639 --> 00:05:38,400
bring it into python land um

131
00:05:35,520 --> 00:05:40,800
it has asynchronous io so it actually

132
00:05:38,400 --> 00:05:42,960
runs in an async i o loop meaning that

133
00:05:40,800 --> 00:05:44,639
in future maybe this could be an async i

134
00:05:42,960 --> 00:05:46,880
o web service like thing and i should

135
00:05:44,639 --> 00:05:47,919
include it into asgi

136
00:05:46,880 --> 00:05:50,800
um

137
00:05:47,919 --> 00:05:52,639
it has some very simple options so all

138
00:05:50,800 --> 00:05:55,840
it does is

139
00:05:52,639 --> 00:05:55,840
loads up um

140
00:05:56,080 --> 00:06:01,520
python module we've got a minute left

141
00:05:58,720 --> 00:06:04,639
great um and it runs things so

142
00:06:01,520 --> 00:06:06,960
uh yeah lots of fun um

143
00:06:04,639 --> 00:06:09,199
lets you do things that you otherwise

144
00:06:06,960 --> 00:06:10,960
couldn't do and be more

145
00:06:09,199 --> 00:06:13,360
uh immediate in the way that you do

146
00:06:10,960 --> 00:06:16,080
stuff uh is there anything else in my

147
00:06:13,360 --> 00:06:18,160
slide deck yes i think there is

148
00:06:16,080 --> 00:06:21,280
only all the things well

149
00:06:18,160 --> 00:06:22,880
in theory you can actually go and put a

150
00:06:21,280 --> 00:06:26,880
new python file

151
00:06:22,880 --> 00:06:29,600
and say that the encoding is um

152
00:06:26,880 --> 00:06:30,479
let's see so you file

153
00:06:29,600 --> 00:06:32,560
well

154
00:06:30,479 --> 00:06:35,440
you know that you can do a codex at the

155
00:06:32,560 --> 00:06:37,919
top like this

156
00:06:35,440 --> 00:06:37,919
coding

157
00:06:38,080 --> 00:06:42,000
eat which means in theory you could

158
00:06:40,160 --> 00:06:44,080
actually put

159
00:06:42,000 --> 00:06:46,319
python code to generate python code

160
00:06:44,080 --> 00:06:48,880
within your python code

161
00:06:46,319 --> 00:06:51,120
have fun with that idea hit me up in the

162
00:06:48,880 --> 00:06:54,479
chat and uh i'll publish this on github

163
00:06:51,120 --> 00:06:56,639
and give you all some links bye

164
00:06:54,479 --> 00:06:58,960
thank you tyson for putting pythons in

165
00:06:56,639 --> 00:07:03,440
our pythons in our pythons

166
00:06:58,960 --> 00:07:05,919
uh up next we have daniel who is going

167
00:07:03,440 --> 00:07:10,240
to be talking to us about racing network

168
00:07:05,919 --> 00:07:10,240
x and tackling code complexity

169
00:07:11,120 --> 00:07:15,360
hello everyone um firstly i acknowledge

170
00:07:13,759 --> 00:07:16,880
the nanga wall people traditional

171
00:07:15,360 --> 00:07:18,960
custodians of the land of which i'm

172
00:07:16,880 --> 00:07:20,720
presenting and pay my respects to their

173
00:07:18,960 --> 00:07:22,080
elders past and present i extend that

174
00:07:20,720 --> 00:07:24,000
respect to aboriginal and torres strait

175
00:07:22,080 --> 00:07:25,440
islander people here today

176
00:07:24,000 --> 00:07:27,360
in this lightning talk i'm going to be

177
00:07:25,440 --> 00:07:30,000
speeding through a journey i took trying

178
00:07:27,360 --> 00:07:31,520
to optimize algorithms in network x i'm

179
00:07:30,000 --> 00:07:32,639
starting here in my jupyter notebook so

180
00:07:31,520 --> 00:07:35,680
that i can run

181
00:07:32,639 --> 00:07:37,280
this cell here

182
00:07:35,680 --> 00:07:38,560
and

183
00:07:37,280 --> 00:07:41,800
we'll come back to that later and see

184
00:07:38,560 --> 00:07:41,800
how it's doing

185
00:07:44,720 --> 00:07:48,720
first of all um so my talk is racing

186
00:07:47,440 --> 00:07:51,280
network x

187
00:07:48,720 --> 00:07:53,120
um so first of all what is networkx it's

188
00:07:51,280 --> 00:07:55,199
the gateway library to graph analysis in

189
00:07:53,120 --> 00:07:56,720
python but it doesn't do everything time

190
00:07:55,199 --> 00:07:58,960
and space limitations make it not very

191
00:07:56,720 --> 00:08:00,400
useful sometimes when playing with large

192
00:07:58,960 --> 00:08:02,400
data sets i still want to be able to

193
00:08:00,400 --> 00:08:04,080
calculate graph metrics but network x

194
00:08:02,400 --> 00:08:06,080
takes too long runs out of memory or

195
00:08:04,080 --> 00:08:08,160
fails to work for other reasons as this

196
00:08:06,080 --> 00:08:10,160
happens i always ask myself or whether i

197
00:08:08,160 --> 00:08:11,919
can figure out how to do it

198
00:08:10,160 --> 00:08:13,199
let's consider a graph metric in this

199
00:08:11,919 --> 00:08:14,960
case we're going to look at something

200
00:08:13,199 --> 00:08:16,960
called eccentricity

201
00:08:14,960 --> 00:08:19,039
what is eccentricity the simplest

202
00:08:16,960 --> 00:08:20,639
explanation is the shortest distance to

203
00:08:19,039 --> 00:08:22,319
the furthest node

204
00:08:20,639 --> 00:08:24,400
for example if i wanted to calculate the

205
00:08:22,319 --> 00:08:26,080
eccentricity of the highlighted circled

206
00:08:24,400 --> 00:08:28,080
node in this picture we need to figure

207
00:08:26,080 --> 00:08:29,599
out how far away all the other nodes are

208
00:08:28,080 --> 00:08:31,280
let's do that by jumping out from that

209
00:08:29,599 --> 00:08:32,880
node until we've reached them all here

210
00:08:31,280 --> 00:08:34,399
we've already started having jumped out

211
00:08:32,880 --> 00:08:36,560
one hop

212
00:08:34,399 --> 00:08:39,360
now we've jumped out two hops

213
00:08:36,560 --> 00:08:41,039
and now three hops

214
00:08:39,360 --> 00:08:42,320
and finally after jumping out four hops

215
00:08:41,039 --> 00:08:44,159
we've reached every other node in the

216
00:08:42,320 --> 00:08:46,160
network so the circled node would have

217
00:08:44,159 --> 00:08:47,920
an eccentricity of four

218
00:08:46,160 --> 00:08:49,200
this is useful because nodes with low

219
00:08:47,920 --> 00:08:51,360
eccentricity will be positioned

220
00:08:49,200 --> 00:08:53,360
centrally on the network nodes that fall

221
00:08:51,360 --> 00:08:55,200
on the outskirts of the net of a network

222
00:08:53,360 --> 00:08:56,720
would have a much higher eccentricity

223
00:08:55,200 --> 00:08:58,160
because they needed to traverse the

224
00:08:56,720 --> 00:08:59,519
entire network in order to reach

225
00:08:58,160 --> 00:09:01,920
everyone

226
00:08:59,519 --> 00:09:04,080
by the way this method of hopping out

227
00:09:01,920 --> 00:09:05,839
being demonstrated is an example of brep

228
00:09:04,080 --> 00:09:07,360
first searching which is great to use if

229
00:09:05,839 --> 00:09:08,640
you don't have a specific destination in

230
00:09:07,360 --> 00:09:10,160
mind

231
00:09:08,640 --> 00:09:11,920
if we think about how most graph query

232
00:09:10,160 --> 00:09:13,360
languages work you typically need to

233
00:09:11,920 --> 00:09:15,120
start somewhere

234
00:09:13,360 --> 00:09:16,560
from somewhere in the network and nodes

235
00:09:15,120 --> 00:09:18,240
with low eccentricity are typically a

236
00:09:16,560 --> 00:09:19,519
good place to start they're more likely

237
00:09:18,240 --> 00:09:20,800
to be highly connected or have

238
00:09:19,519 --> 00:09:22,640
information passing through them as

239
00:09:20,800 --> 00:09:26,640
opposed to our highly eccentric mr burns

240
00:09:22,640 --> 00:09:28,000
node over in the uh edge

241
00:09:26,640 --> 00:09:30,160
so how does network x go with

242
00:09:28,000 --> 00:09:32,240
calculating eccentricity well for one

243
00:09:30,160 --> 00:09:34,000
thing network x's eccentricity algorithm

244
00:09:32,240 --> 00:09:35,519
doesn't work on digraphs or directed

245
00:09:34,000 --> 00:09:37,360
graphs you can convert it to an

246
00:09:35,519 --> 00:09:38,800
undirected graph but it's not a true

247
00:09:37,360 --> 00:09:40,480
reflection of how information is flowing

248
00:09:38,800 --> 00:09:42,480
in the network

249
00:09:40,480 --> 00:09:44,160
network x's eccentricity algorithm also

250
00:09:42,480 --> 00:09:46,000
doesn't work on disconnected graphs you

251
00:09:44,160 --> 00:09:48,000
can cycle through connected components

252
00:09:46,000 --> 00:09:49,680
but as a personal pet peeve network x

253
00:09:48,000 --> 00:09:52,240
moves this package around every single

254
00:09:49,680 --> 00:09:54,640
time it updates

255
00:09:52,240 --> 00:09:56,480
i'm even accounting for these issues if

256
00:09:54,640 --> 00:09:58,800
the data set is too big network x will

257
00:09:56,480 --> 00:10:00,880
just run out of memory or take too long

258
00:09:58,800 --> 00:10:02,240
unfortunately i'm as stubborn as a mule

259
00:10:00,880 --> 00:10:05,040
approximations weren't going to cut it

260
00:10:02,240 --> 00:10:06,240
either i want to calculate uh the real

261
00:10:05,040 --> 00:10:07,920
thing

262
00:10:06,240 --> 00:10:09,360
the stanford network a large network

263
00:10:07,920 --> 00:10:10,399
data set collection is a great place to

264
00:10:09,360 --> 00:10:12,240
pull some large networks for

265
00:10:10,399 --> 00:10:14,160
benchmarking so i downloaded a nice big

266
00:10:12,240 --> 00:10:15,440
graph and began experimenting

267
00:10:14,160 --> 00:10:16,880
i wrote some code to mimic how

268
00:10:15,440 --> 00:10:18,880
information was propagating through the

269
00:10:16,880 --> 00:10:20,720
network in a parallel breadth first

270
00:10:18,880 --> 00:10:22,320
search style you'll notice here that i'm

271
00:10:20,720 --> 00:10:24,480
working with the numeric representation

272
00:10:22,320 --> 00:10:26,399
of the graph the adjacency matrix the

273
00:10:24,480 --> 00:10:28,160
rows and columns of the adjacency matrix

274
00:10:26,399 --> 00:10:29,839
reflect each node and typically the

275
00:10:28,160 --> 00:10:31,360
values of the matrix we won if there's a

276
00:10:29,839 --> 00:10:33,519
direct connection between the row node

277
00:10:31,360 --> 00:10:35,440
and column node or zero if there isn't

278
00:10:33,519 --> 00:10:38,160
sometimes this uh value will represent

279
00:10:35,440 --> 00:10:40,000
the weight of the connection

280
00:10:38,160 --> 00:10:42,160
the ipython magic function time it is

281
00:10:40,000 --> 00:10:43,920
great for overall benchmarking but p run

282
00:10:42,160 --> 00:10:46,320
was invaluable for finding where in the

283
00:10:43,920 --> 00:10:48,480
where the in the code my the bottlenecks

284
00:10:46,320 --> 00:10:49,839
were p run shows you the number of times

285
00:10:48,480 --> 00:10:52,480
functions are called and the times taken

286
00:10:49,839 --> 00:10:54,640
per call and overall um how long things

287
00:10:52,480 --> 00:10:56,399
take which is super handy

288
00:10:54,640 --> 00:10:58,480
other useful tricks include either tools

289
00:10:56,399 --> 00:11:00,399
and more edit tools i found that if i'm

290
00:10:58,480 --> 00:11:01,839
ever looping through anything i have a

291
00:11:00,399 --> 00:11:03,519
look at those libraries and see if what

292
00:11:01,839 --> 00:11:04,959
i'm trying to do is there

293
00:11:03,519 --> 00:11:06,720
another useful thing i learned was a

294
00:11:04,959 --> 00:11:09,120
stitch between c contiguous arrays and

295
00:11:06,720 --> 00:11:11,360
fortran contiguous arrays in numpy

296
00:11:09,120 --> 00:11:14,880
order c is better for row operations and

297
00:11:11,360 --> 00:11:16,720
order f is better for column operations

298
00:11:14,880 --> 00:11:18,720
eventually my approach was faster but

299
00:11:16,720 --> 00:11:20,800
still limited by the space complexity of

300
00:11:18,720 --> 00:11:22,720
the big snap graphs

301
00:11:20,800 --> 00:11:24,320
so i got rid of the dense matrix numpy

302
00:11:22,720 --> 00:11:26,160
stuff because that was taking far too

303
00:11:24,320 --> 00:11:28,480
much memory and started working

304
00:11:26,160 --> 00:11:30,240
exclusively with sci-fi sparse matrices

305
00:11:28,480 --> 00:11:31,760
that helped but the biggest bottleneck

306
00:11:30,240 --> 00:11:34,000
was the matrix dot product

307
00:11:31,760 --> 00:11:36,000
multiplication which is always

308
00:11:34,000 --> 00:11:37,440
computationally very expensive

309
00:11:36,000 --> 00:11:39,200
ultimately i had to do with matrix

310
00:11:37,440 --> 00:11:41,600
algebra altogether and tried replicating

311
00:11:39,200 --> 00:11:43,200
it using binary set operations initially

312
00:11:41,600 --> 00:11:44,959
i used python sets but they still

313
00:11:43,200 --> 00:11:46,160
suffered from space limitations because

314
00:11:44,959 --> 00:11:48,000
you need to store an additional hash

315
00:11:46,160 --> 00:11:50,320
table for lookups the work around there

316
00:11:48,000 --> 00:11:52,320
was to use bit arrows bit arrays are a

317
00:11:50,320 --> 00:11:53,680
very space efficient boolean collection

318
00:11:52,320 --> 00:11:55,200
that could easily represent the graph

319
00:11:53,680 --> 00:11:57,120
adjacency matrix

320
00:11:55,200 --> 00:11:58,880
this created some additional overhead in

321
00:11:57,120 --> 00:12:00,800
the conversion but once that was done my

322
00:11:58,880 --> 00:12:02,320
approach could not only run on the large

323
00:12:00,800 --> 00:12:03,920
snap data set but was faster than other

324
00:12:02,320 --> 00:12:05,600
optimized graph algorithm libraries as

325
00:12:03,920 --> 00:12:07,680
well

326
00:12:05,600 --> 00:12:09,440
go back to here is it finished running

327
00:12:07,680 --> 00:12:11,040
no it hasn't it usually takes

328
00:12:09,440 --> 00:12:14,880
under five minutes but

329
00:12:11,040 --> 00:12:14,880
i'm doing streaming stuff as well thanks

330
00:12:15,519 --> 00:12:17,760
thanks daniel

331
00:12:16,959 --> 00:12:21,839
uh

332
00:12:17,760 --> 00:12:23,360
up next we have jeff who is going to be

333
00:12:21,839 --> 00:12:24,720
taking us

334
00:12:23,360 --> 00:12:27,120
well through something slightly more

335
00:12:24,720 --> 00:12:29,200
relaxing than networks

336
00:12:27,120 --> 00:12:30,639
go ahead jeff thanks so much christopher

337
00:12:29,200 --> 00:12:33,040
uh yeah it is something completely

338
00:12:30,639 --> 00:12:35,040
different uh i want to thank everybody

339
00:12:33,040 --> 00:12:36,480
very quickly for a fantastic conference

340
00:12:35,040 --> 00:12:38,800
but i would also like to invite

341
00:12:36,480 --> 00:12:42,240
everybody for a moment of high

342
00:12:38,800 --> 00:12:44,639
contemplation and apologies for the pun

343
00:12:42,240 --> 00:12:45,920
and uh really what i want to do today in

344
00:12:44,639 --> 00:12:47,279
five minutes is i just want to share

345
00:12:45,920 --> 00:12:49,440
something that's helping me through

346
00:12:47,279 --> 00:12:51,200
these really unusual times that we've

347
00:12:49,440 --> 00:12:54,480
all been sharing

348
00:12:51,200 --> 00:12:56,320
and i will illustrate this through a

349
00:12:54,480 --> 00:12:58,639
really really terrible watercolor and a

350
00:12:56,320 --> 00:13:01,200
story uh so apologies in advance for the

351
00:12:58,639 --> 00:13:02,800
terrible watercolor uh but yeah this

352
00:13:01,200 --> 00:13:05,360
this little fella here we're gonna call

353
00:13:02,800 --> 00:13:08,880
them uh susan the schneck

354
00:13:05,360 --> 00:13:10,880
and uh susan the schneck uh circa 2019

355
00:13:08,880 --> 00:13:12,639
was having a pretty happy life i was

356
00:13:10,880 --> 00:13:14,160
meeting all sorts of other schnecks

357
00:13:12,639 --> 00:13:17,600
potentially in person at conferences

358
00:13:14,160 --> 00:13:18,800
like this uh and then things changed uh

359
00:13:17,600 --> 00:13:22,000
i'm going to draw your attention to the

360
00:13:18,800 --> 00:13:24,560
left to this uh well called the y axis

361
00:13:22,000 --> 00:13:26,480
uh you know vibes were pretty good uh

362
00:13:24,560 --> 00:13:30,079
maybe bushfire is not withstanding vibes

363
00:13:26,480 --> 00:13:32,240
are okay in 2019 and uh we can

364
00:13:30,079 --> 00:13:35,839
potentially argue that vibes are

365
00:13:32,240 --> 00:13:38,320
pretty awful in uh uh

366
00:13:35,839 --> 00:13:40,079
now so so essentially like susan got

367
00:13:38,320 --> 00:13:42,399
thrown into a little box

368
00:13:40,079 --> 00:13:43,519
and over time susan got so caught up

369
00:13:42,399 --> 00:13:45,600
with all the things that susan couldn't

370
00:13:43,519 --> 00:13:48,160
do uh that susan sort of tied herself

371
00:13:45,600 --> 00:13:50,639
into a knot and you could say that this

372
00:13:48,160 --> 00:13:52,399
little axis here the the yellow axis

373
00:13:50,639 --> 00:13:54,560
which is like happiness you can say that

374
00:13:52,399 --> 00:13:56,000
susan probably was a bit miserable at

375
00:13:54,560 --> 00:13:57,519
that point

376
00:13:56,000 --> 00:13:59,600
uh so the method that i'm going to show

377
00:13:57,519 --> 00:14:01,199
you today is just it's a bit to sort of

378
00:13:59,600 --> 00:14:03,760
be aware that you know maybe you're

379
00:14:01,199 --> 00:14:05,440
tying yourself into knots like susan and

380
00:14:03,760 --> 00:14:07,839
even if your external vibes your

381
00:14:05,440 --> 00:14:09,760
external factors don't change uh if you

382
00:14:07,839 --> 00:14:12,480
practice it fairly diligently you might

383
00:14:09,760 --> 00:14:14,240
find yourself getting a lot happier

384
00:14:12,480 --> 00:14:15,920
so let's go away from watercolors and

385
00:14:14,240 --> 00:14:17,199
actually get into the exercise so for

386
00:14:15,920 --> 00:14:20,720
the next two minutes we're going to

387
00:14:17,199 --> 00:14:22,399
close our eyes and focus on our breath

388
00:14:20,720 --> 00:14:24,639
we're not going to try to control our

389
00:14:22,399 --> 00:14:26,079
breath we're just going to remain aware

390
00:14:24,639 --> 00:14:27,760
of our breath

391
00:14:26,079 --> 00:14:30,240
and that means we're going to know when

392
00:14:27,760 --> 00:14:31,920
it comes in and out we might know

393
00:14:30,240 --> 00:14:34,079
whether it's hard or whether it's soft

394
00:14:31,920 --> 00:14:35,279
whether it's fast or whether it's slow

395
00:14:34,079 --> 00:14:36,959
whether it's coming through our nose

396
00:14:35,279 --> 00:14:38,480
whether it's coming through our mouth

397
00:14:36,959 --> 00:14:40,079
really doesn't matter if it's any of

398
00:14:38,480 --> 00:14:41,839
these things it just matters that we're

399
00:14:40,079 --> 00:14:43,279
aware of it

400
00:14:41,839 --> 00:14:45,279
and sometimes placing a hand on your

401
00:14:43,279 --> 00:14:46,399
stomach so you can feel the rise or fall

402
00:14:45,279 --> 00:14:48,880
that helps

403
00:14:46,399 --> 00:14:50,160
or even just trying to focus on

404
00:14:48,880 --> 00:14:52,000
the area around your nostril that can

405
00:14:50,160 --> 00:14:53,760
sometimes also help as well

406
00:14:52,000 --> 00:14:57,120
and even though two minutes is a short

407
00:14:53,760 --> 00:14:59,040
time i guarantee 99 of us including me

408
00:14:57,120 --> 00:15:01,680
will get distracted

409
00:14:59,040 --> 00:15:03,360
so those are the instructions i wish you

410
00:15:01,680 --> 00:15:05,920
the best of success

411
00:15:03,360 --> 00:15:08,160
and yes it doesn't matter if the mind

412
00:15:05,920 --> 00:15:09,760
will wonder this is natural and expected

413
00:15:08,160 --> 00:15:12,800
what matters is that you bring it back

414
00:15:09,760 --> 00:15:14,079
gently without self-reproach

415
00:15:12,800 --> 00:15:15,920
so we're going to do these two minutes

416
00:15:14,079 --> 00:15:17,440
now uh if you want to go and have a look

417
00:15:15,920 --> 00:15:19,680
at the slides later

418
00:15:17,440 --> 00:15:22,880
they've got references right at the end

419
00:15:19,680 --> 00:15:22,880
let's give it a go right now

420
00:15:25,279 --> 00:15:30,120
so two minutes starting now

421
00:16:29,120 --> 00:16:32,440
one minute

422
00:17:26,050 --> 00:17:30,080
[Music]

423
00:17:27,360 --> 00:17:31,120
way to ruin the vibes there with the one

424
00:17:30,080 --> 00:17:33,600
minute

425
00:17:31,120 --> 00:17:36,640
i mean i appreciate that it was a little

426
00:17:33,600 --> 00:17:38,880
bit toned down but i was i was so

427
00:17:36,640 --> 00:17:41,520
relaxed i was so trying to match the

428
00:17:38,880 --> 00:17:44,720
energy in the room

429
00:17:41,520 --> 00:17:47,120
i promise one minute

430
00:17:44,720 --> 00:17:50,799
righto okay next we have

431
00:17:47,120 --> 00:17:54,000
tim who is doing some things with the

432
00:17:50,799 --> 00:17:54,000
pycon au logo

433
00:17:58,160 --> 00:18:03,760
all right so um this is sort of

434
00:18:00,480 --> 00:18:05,919
something that started uh at last pycon

435
00:18:03,760 --> 00:18:07,840
and i uh put up a sort of a

436
00:18:05,919 --> 00:18:09,600
image of this um onto twitter a short

437
00:18:07,840 --> 00:18:11,120
video um and a lot of people have been

438
00:18:09,600 --> 00:18:12,720
asking about it which is sitting over

439
00:18:11,120 --> 00:18:14,000
here in the background

440
00:18:12,720 --> 00:18:16,400
um so there's a bit of elements of

441
00:18:14,000 --> 00:18:18,559
hardware software and woodwork involved

442
00:18:16,400 --> 00:18:20,400
in this little project um so like i said

443
00:18:18,559 --> 00:18:23,280
i started this at last pycon and i

444
00:18:20,400 --> 00:18:25,840
managed to get this far

445
00:18:23,280 --> 00:18:26,960
um it's been sitting on my desk uh sort

446
00:18:25,840 --> 00:18:29,280
of

447
00:18:26,960 --> 00:18:30,960
annoying me ever since and a couple of

448
00:18:29,280 --> 00:18:33,200
weeks ago i sort of looked at this and

449
00:18:30,960 --> 00:18:34,720
thought it's about time i actually

450
00:18:33,200 --> 00:18:36,240
got on to and actually made this into

451
00:18:34,720 --> 00:18:38,080
something for real

452
00:18:36,240 --> 00:18:40,720
so first step is a little bit of

453
00:18:38,080 --> 00:18:40,720
prototyping

454
00:18:41,120 --> 00:18:45,919
i caught my first test that the uh snit

455
00:18:43,600 --> 00:18:48,160
egg uh this is sort of where i tried out

456
00:18:45,919 --> 00:18:50,000
the concept um it's a

457
00:18:48,160 --> 00:18:53,039
potentially drill a hole fill it with

458
00:18:50,000 --> 00:18:55,440
hot glue and and put a led inside and

459
00:18:53,039 --> 00:18:56,480
you get this sort of nice output here

460
00:18:55,440 --> 00:18:58,400
um

461
00:18:56,480 --> 00:19:00,320
this is my second sort of test where we

462
00:18:58,400 --> 00:19:01,840
actually made it into a full uh what i

463
00:19:00,320 --> 00:19:02,880
call a baby snack

464
00:19:01,840 --> 00:19:04,320
and

465
00:19:02,880 --> 00:19:05,520
it started to see you could see the

466
00:19:04,320 --> 00:19:07,280
colors sort of starting to blend and it

467
00:19:05,520 --> 00:19:10,400
looks kind of kind of cool

468
00:19:07,280 --> 00:19:13,360
so moving on to the final build first

469
00:19:10,400 --> 00:19:14,880
stage was uh drilling out or laying out

470
00:19:13,360 --> 00:19:17,440
you can see my sort of marks there where

471
00:19:14,880 --> 00:19:18,559
i've laid it out uh drilling in all the

472
00:19:17,440 --> 00:19:19,919
positioning

473
00:19:18,559 --> 00:19:21,520
and then putting this through onto a

474
00:19:19,919 --> 00:19:23,039
router table to route out the rest of

475
00:19:21,520 --> 00:19:25,200
the

476
00:19:23,039 --> 00:19:27,360
rest of the shape uh putting in the leds

477
00:19:25,200 --> 00:19:29,120
and hot glue uh this is where i ran to

478
00:19:27,360 --> 00:19:33,360
my first problem uh i ran out of hot

479
00:19:29,120 --> 00:19:34,559
glue which is not a useful thing uh

480
00:19:33,360 --> 00:19:37,360
it's sort of a bit of a rush there to

481
00:19:34,559 --> 00:19:39,039
try and get another lot of it in uh

482
00:19:37,360 --> 00:19:40,880
in the middle of a lockdown but i

483
00:19:39,039 --> 00:19:42,320
managed to do that but it does have some

484
00:19:40,880 --> 00:19:44,400
side effects and that you get bubbles

485
00:19:42,320 --> 00:19:47,280
and that starting to come through

486
00:19:44,400 --> 00:19:48,960
from that we move on to wiring

487
00:19:47,280 --> 00:19:50,640
more wiring

488
00:19:48,960 --> 00:19:53,360
crimping

489
00:19:50,640 --> 00:19:54,960
before we get to our first test

490
00:19:53,360 --> 00:19:56,799
and uh we start to see it sort of how

491
00:19:54,960 --> 00:19:58,720
it's going to actually look uh there's a

492
00:19:56,799 --> 00:20:00,240
few sort of uh other points that start

493
00:19:58,720 --> 00:20:01,919
to come out here those the green and red

494
00:20:00,240 --> 00:20:03,760
is very bright um and we're gonna have

495
00:20:01,919 --> 00:20:06,400
to tone those down and we'll comment

496
00:20:03,760 --> 00:20:09,280
exactly how that works next so the

497
00:20:06,400 --> 00:20:11,760
actual hardware the hardware um

498
00:20:09,280 --> 00:20:13,200
yeah there's a lot of wiring and uh as

499
00:20:11,760 --> 00:20:15,840
you expect for these kind of things

500
00:20:13,200 --> 00:20:18,960
where we control every led

501
00:20:15,840 --> 00:20:21,120
most of the control was all done with a

502
00:20:18,960 --> 00:20:22,480
bunch of shift registers

503
00:20:21,120 --> 00:20:24,640
these ones are particularly nice they

504
00:20:22,480 --> 00:20:28,159
can sync 150 milliamps which is perfect

505
00:20:24,640 --> 00:20:30,320
for running leds uh they work by having

506
00:20:28,159 --> 00:20:31,280
three wires coming into them a data a

507
00:20:30,320 --> 00:20:32,720
clock

508
00:20:31,280 --> 00:20:35,039
and a latch wire

509
00:20:32,720 --> 00:20:37,919
you set a value on the data polls the

510
00:20:35,039 --> 00:20:40,640
clock and it will uh shift the the value

511
00:20:37,919 --> 00:20:42,000
in do that repetitively and you'll get

512
00:20:40,640 --> 00:20:44,880
all of your data in

513
00:20:42,000 --> 00:20:46,640
and finally pulse the latch it switches

514
00:20:44,880 --> 00:20:48,320
on to the outputs and you just you know

515
00:20:46,640 --> 00:20:49,919
when you what you want to see comes up

516
00:20:48,320 --> 00:20:52,240
on the display

517
00:20:49,919 --> 00:20:53,919
uh at the moment it's running with a

518
00:20:52,240 --> 00:20:55,440
raspberry pi pico which is sort of

519
00:20:53,919 --> 00:20:58,480
probably my favorite controller at this

520
00:20:55,440 --> 00:20:58,480
particular point in time

521
00:20:58,559 --> 00:21:02,159
how we get into the actual

522
00:21:00,400 --> 00:21:03,280
brightness i was using a technique

523
00:21:02,159 --> 00:21:05,200
called bam

524
00:21:03,280 --> 00:21:08,159
the angle modulation it's a sort of an

525
00:21:05,200 --> 00:21:10,880
alternate to pwm that can be implemented

526
00:21:08,159 --> 00:21:12,640
efficiently in software

527
00:21:10,880 --> 00:21:14,799
in this particular

528
00:21:12,640 --> 00:21:16,000
picture i have the this of the green

529
00:21:14,799 --> 00:21:18,159
which is what the waveforms will look

530
00:21:16,000 --> 00:21:21,760
like from bam and below that is what a

531
00:21:18,159 --> 00:21:22,640
traditional pwm waveform would look like

532
00:21:21,760 --> 00:21:25,120
the

533
00:21:22,640 --> 00:21:26,559
ratio between the on and off states

534
00:21:25,120 --> 00:21:29,200
and runs at a high enough frequency and

535
00:21:26,559 --> 00:21:32,159
you can control brightness

536
00:21:29,200 --> 00:21:34,159
um so micro python turned out really

537
00:21:32,159 --> 00:21:36,720
really good in this particular scenario

538
00:21:34,159 --> 00:21:40,320
here uh the uh

539
00:21:36,720 --> 00:21:42,159
this threading model on on the micro uh

540
00:21:40,320 --> 00:21:43,760
on the pico sorry it was great in that

541
00:21:42,159 --> 00:21:45,200
it would actually use both

542
00:21:43,760 --> 00:21:46,400
cores which mean you can run the

543
00:21:45,200 --> 00:21:48,720
animations

544
00:21:46,400 --> 00:21:49,919
and output on separate uh threads in

545
00:21:48,720 --> 00:21:51,679
parallel

546
00:21:49,919 --> 00:21:54,159
although i did run into some limits here

547
00:21:51,679 --> 00:21:55,120
uh the speed of the of the of micro

548
00:21:54,159 --> 00:21:56,480
python

549
00:21:55,120 --> 00:22:00,080
it just isn't quite up to it so i could

550
00:21:56,480 --> 00:22:02,320
only go run it minutes using 4-bit bam

551
00:22:00,080 --> 00:22:05,679
um meaning you only got 16 levels of

552
00:22:02,320 --> 00:22:08,799
brightness but that's more than enough

553
00:22:05,679 --> 00:22:10,840
so the final result

554
00:22:08,799 --> 00:22:14,159
is this and hopefully it'll

555
00:22:10,840 --> 00:22:15,840
run um which is a bit of an up close uh

556
00:22:14,159 --> 00:22:17,039
someone a nicer effect that came out of

557
00:22:15,840 --> 00:22:19,200
this is that you can kind of see that

558
00:22:17,039 --> 00:22:21,120
the colors blending as it comes out

559
00:22:19,200 --> 00:22:22,799
which uh which was actually quite a nice

560
00:22:21,120 --> 00:22:25,760
sort of side effect of running it

561
00:22:22,799 --> 00:22:29,039
through this glue which was

562
00:22:25,760 --> 00:22:29,039
causing it to diffuse nicely

563
00:22:29,360 --> 00:22:32,000
um

564
00:22:30,240 --> 00:22:34,480
so

565
00:22:32,000 --> 00:22:36,320
what's next um i wouldn't mind porting

566
00:22:34,480 --> 00:22:37,840
this to a different language just to get

567
00:22:36,320 --> 00:22:41,120
a bit more speed out of it probably

568
00:22:37,840 --> 00:22:43,200
laser cut we can reuse the controller

569
00:22:41,120 --> 00:22:44,799
don't run out of glue would be a handy

570
00:22:43,200 --> 00:22:45,600
anything to do here

571
00:22:44,799 --> 00:22:47,600
um

572
00:22:45,600 --> 00:22:49,679
slide some code if you want to read more

573
00:22:47,600 --> 00:22:52,159
about what bam is

574
00:22:49,679 --> 00:22:53,600
there i'd have to also quickly shout out

575
00:22:52,159 --> 00:22:55,840
to jim massad who jumped in from

576
00:22:53,600 --> 00:22:57,679
micropython to help me we'll actually

577
00:22:55,840 --> 00:22:58,880
just listen to me solve my own problems

578
00:22:57,679 --> 00:23:01,679
thank you

579
00:22:58,880 --> 00:23:04,480
oh thank you for that tim it is cool to

580
00:23:01,679 --> 00:23:06,320
see people doing such pretty things with

581
00:23:04,480 --> 00:23:07,520
uh with the mascot of this year

582
00:23:06,320 --> 00:23:11,679
conference

583
00:23:07,520 --> 00:23:14,240
uh up next we have matt sendia

584
00:23:11,679 --> 00:23:16,000
uh who is going to be talking about

585
00:23:14,240 --> 00:23:17,520
what it says on your screen making

586
00:23:16,000 --> 00:23:19,440
welcome

587
00:23:17,520 --> 00:23:21,440
then hi everyone my name is matt um my

588
00:23:19,440 --> 00:23:22,799
pronouns are they and them i'm going to

589
00:23:21,440 --> 00:23:24,559
be talking to you about neurodiversity

590
00:23:22,799 --> 00:23:26,400
today

591
00:23:24,559 --> 00:23:27,919
um i'd like to start by acknowledging

592
00:23:26,400 --> 00:23:29,840
the traditional custodians of the land

593
00:23:27,919 --> 00:23:32,080
from which i'm presenting the waranjari

594
00:23:29,840 --> 00:23:33,520
people of the cooler nation and pay my

595
00:23:32,080 --> 00:23:36,320
respects to their elders past present

596
00:23:33,520 --> 00:23:38,640
and emerging

597
00:23:36,320 --> 00:23:41,520
so i have autism

598
00:23:38,640 --> 00:23:41,520
and adhd

599
00:23:42,000 --> 00:23:45,520
these are both two types of

600
00:23:44,000 --> 00:23:47,200
neurodiversity

601
00:23:45,520 --> 00:23:49,440
which refers to the variation in the

602
00:23:47,200 --> 00:23:51,919
human brain regarding sociability

603
00:23:49,440 --> 00:23:54,080
learning attention and mood and other

604
00:23:51,919 --> 00:23:56,240
mental functions

605
00:23:54,080 --> 00:23:58,320
but they're by no means the only two

606
00:23:56,240 --> 00:24:00,159
types of neurodiversity

607
00:23:58,320 --> 00:24:02,080
i'm recently diagnosed i'm still

608
00:24:00,159 --> 00:24:02,840
learning about my brain

609
00:24:02,080 --> 00:24:05,200
um

610
00:24:02,840 --> 00:24:07,200
i my intent for this talk is to raise

611
00:24:05,200 --> 00:24:09,279
awareness not to claim that i have all

612
00:24:07,200 --> 00:24:11,440
the answers um we have a saying that if

613
00:24:09,279 --> 00:24:13,840
you've met one person with autism you've

614
00:24:11,440 --> 00:24:16,840
met one person with autism we're all

615
00:24:13,840 --> 00:24:19,840
different and these are specifically my

616
00:24:16,840 --> 00:24:21,679
experiences i found a long time ago that

617
00:24:19,840 --> 00:24:22,880
i always struggled with timekeeping and

618
00:24:21,679 --> 00:24:25,200
timesheets

619
00:24:22,880 --> 00:24:27,039
i'd seemingly lose hours of productivity

620
00:24:25,200 --> 00:24:28,159
some days and not know where they went

621
00:24:27,039 --> 00:24:30,720
um

622
00:24:28,159 --> 00:24:32,880
and it made it really challenging to to

623
00:24:30,720 --> 00:24:34,960
build against clients uh and so i ended

624
00:24:32,880 --> 00:24:36,240
up having to use this um

625
00:24:34,960 --> 00:24:38,159
app that a colleague wrote which would

626
00:24:36,240 --> 00:24:39,600
literally pop up every 15 minutes to

627
00:24:38,159 --> 00:24:41,360
help me remember

628
00:24:39,600 --> 00:24:43,520
to timesheet my work and keep track of

629
00:24:41,360 --> 00:24:45,120
what i was doing um

630
00:24:43,520 --> 00:24:47,360
timesheets don't really work for me and

631
00:24:45,120 --> 00:24:49,760
i'm trying to find ways to have jobs

632
00:24:47,360 --> 00:24:51,200
that don't don't involve them so that's

633
00:24:49,760 --> 00:24:53,200
one thing to think about for people who

634
00:24:51,200 --> 00:24:55,120
are neurodiverse

635
00:24:53,200 --> 00:24:57,360
and so now i know why

636
00:24:55,120 --> 00:24:59,120
it's because i hyper focus and sometimes

637
00:24:57,360 --> 00:25:00,720
experience time blindness

638
00:24:59,120 --> 00:25:01,520
i can focus on things for hours which is

639
00:25:00,720 --> 00:25:04,799
like a

640
00:25:01,520 --> 00:25:06,960
next level flow state um

641
00:25:04,799 --> 00:25:09,200
where i can get quite upset if i get

642
00:25:06,960 --> 00:25:11,039
interrupted or i have to context switch

643
00:25:09,200 --> 00:25:13,279
it means i can get a lot done but it

644
00:25:11,039 --> 00:25:15,919
also means that

645
00:25:13,279 --> 00:25:17,200
there is nothing else outside of this um

646
00:25:15,919 --> 00:25:19,840
and it means i can completely lose track

647
00:25:17,200 --> 00:25:22,320
of time and forget important things like

648
00:25:19,840 --> 00:25:24,240
eating

649
00:25:22,320 --> 00:25:27,440
i also experience executive dysfunction

650
00:25:24,240 --> 00:25:29,440
um if things are not interesting

651
00:25:27,440 --> 00:25:30,720
novel or urgent

652
00:25:29,440 --> 00:25:32,000
they don't get done

653
00:25:30,720 --> 00:25:34,240
this talk this talk was written in the

654
00:25:32,000 --> 00:25:36,159
last two hours it hit the urgent the

655
00:25:34,240 --> 00:25:37,520
urgent checkbox

656
00:25:36,159 --> 00:25:39,279
um

657
00:25:37,520 --> 00:25:41,840
i struggle with engagement

658
00:25:39,279 --> 00:25:44,240
sometimes i find that i have to watch

659
00:25:41,840 --> 00:25:46,240
talks at double time occasionally just

660
00:25:44,240 --> 00:25:47,840
to keep keep engaged with them if i can

661
00:25:46,240 --> 00:25:49,200
unless they're written by libra necky or

662
00:25:47,840 --> 00:25:51,360
russell t mcgee in which case they work

663
00:25:49,200 --> 00:25:53,440
just fine

664
00:25:51,360 --> 00:25:55,200
um i find that i'm occasionally

665
00:25:53,440 --> 00:25:57,440
distracted by environmental factors such

666
00:25:55,200 --> 00:25:59,279
as background noises like elevated dings

667
00:25:57,440 --> 00:26:00,480
or fluorescent lighting or movements in

668
00:25:59,279 --> 00:26:02,159
the periphery

669
00:26:00,480 --> 00:26:04,080
so i often need to use noise cancelling

670
00:26:02,159 --> 00:26:05,679
headphones or like work at my home

671
00:26:04,080 --> 00:26:06,960
office where i can

672
00:26:05,679 --> 00:26:09,200
you know

673
00:26:06,960 --> 00:26:11,840
close out those distractions and

674
00:26:09,200 --> 00:26:13,200
focus properly on my work

675
00:26:11,840 --> 00:26:15,039
i find that different communication

676
00:26:13,200 --> 00:26:16,559
styles work differently for me instant

677
00:26:15,039 --> 00:26:19,039
messages

678
00:26:16,559 --> 00:26:21,279
are great for

679
00:26:19,039 --> 00:26:23,760
not putting people on the spot but you

680
00:26:21,279 --> 00:26:25,919
can often lose tone with them um if you

681
00:26:23,760 --> 00:26:28,559
don't take care video

682
00:26:25,919 --> 00:26:30,880
and audio are great um

683
00:26:28,559 --> 00:26:32,559
but because you've got the body language

684
00:26:30,880 --> 00:26:34,960
and the tone but they can be stressful

685
00:26:32,559 --> 00:26:36,240
from a like a timeliness and response

686
00:26:34,960 --> 00:26:38,720
perspective if you feel like you're

687
00:26:36,240 --> 00:26:40,640
getting put on the spot um audio can be

688
00:26:38,720 --> 00:26:43,840
the worst of both worlds sometimes so i

689
00:26:40,640 --> 00:26:46,240
try to avoid phone calls for example

690
00:26:43,840 --> 00:26:48,720
um one thing that i found worked really

691
00:26:46,240 --> 00:26:50,400
well are personal user manuals which are

692
00:26:48,720 --> 00:26:52,080
ways that people can literally write a

693
00:26:50,400 --> 00:26:53,840
document that says here is how i work

694
00:26:52,080 --> 00:26:55,039
and here's how my brain works

695
00:26:53,840 --> 00:26:56,880
um

696
00:26:55,039 --> 00:26:58,559
and i've you know read about workplaces

697
00:26:56,880 --> 00:27:00,400
that have implemented these

698
00:26:58,559 --> 00:27:02,000
as a as an optional thing where you know

699
00:27:00,400 --> 00:27:03,440
they're not required but it's a great

700
00:27:02,000 --> 00:27:04,799
way for people to

701
00:27:03,440 --> 00:27:06,240
communicate about what works best for

702
00:27:04,799 --> 00:27:08,320
them and find accommodations and

703
00:27:06,240 --> 00:27:09,520
compromise

704
00:27:08,320 --> 00:27:11,279
um

705
00:27:09,520 --> 00:27:13,440
so the main takeaways that i've got from

706
00:27:11,279 --> 00:27:16,080
this talk are that

707
00:27:13,440 --> 00:27:19,600
all of our brains work differently and

708
00:27:16,080 --> 00:27:20,880
even people who aren't neurodiverse

709
00:27:19,600 --> 00:27:22,080
require different ways of thinking

710
00:27:20,880 --> 00:27:24,080
sometimes

711
00:27:22,080 --> 00:27:25,919
um so i'd say

712
00:27:24,080 --> 00:27:28,159
be explicit and

713
00:27:25,919 --> 00:27:30,320
open and proactive about working with

714
00:27:28,159 --> 00:27:31,760
people to accommodate the way they work

715
00:27:30,320 --> 00:27:34,399
um as i said this helps both

716
00:27:31,760 --> 00:27:36,799
neurotypical and neurodivergent folks

717
00:27:34,399 --> 00:27:38,559
stay curious and have empathy for you

718
00:27:36,799 --> 00:27:40,960
know the way people think

719
00:27:38,559 --> 00:27:43,360
um and that's all i've got here are some

720
00:27:40,960 --> 00:27:45,600
references which i will put in the um in

721
00:27:43,360 --> 00:27:48,720
the venulis chat in a second thank you

722
00:27:45,600 --> 00:27:51,520
very much

723
00:27:48,720 --> 00:27:53,440
thank you matt that was a uh that was a

724
00:27:51,520 --> 00:27:54,640
really important talk and i i thank you

725
00:27:53,440 --> 00:27:58,000
for giving it

726
00:27:54,640 --> 00:28:01,840
uh up next we have jens

727
00:27:58,000 --> 00:28:05,440
who is going to be talking

728
00:28:01,840 --> 00:28:05,440
grassland in lower saxony

729
00:28:05,679 --> 00:28:10,880
yeah thank you very much um yeah why i

730
00:28:08,320 --> 00:28:13,600
want to give this talk um of course i'd

731
00:28:10,880 --> 00:28:15,840
like to get accutaned with geodata in

732
00:28:13,600 --> 00:28:18,000
agriculture and also from those data

733
00:28:15,840 --> 00:28:20,480
that are used from administrative bodies

734
00:28:18,000 --> 00:28:22,720
for at least in lower saxony

735
00:28:20,480 --> 00:28:24,640
then i'd like to show you a few basic

736
00:28:22,720 --> 00:28:27,440
commands and geopandas you need to know

737
00:28:24,640 --> 00:28:30,799
if you want to process such data

738
00:28:27,440 --> 00:28:33,360
and of course you will be eventually

739
00:28:30,799 --> 00:28:35,440
able to validate administrative land use

740
00:28:33,360 --> 00:28:37,679
statistics so you can check if they are

741
00:28:35,440 --> 00:28:40,159
doing good job or not

742
00:28:37,679 --> 00:28:42,960
first of all i like to show where the

743
00:28:40,159 --> 00:28:45,440
lower section is and by the way yes it

744
00:28:42,960 --> 00:28:47,039
has a horse inside so it's a running gag

745
00:28:45,440 --> 00:28:49,120
of the chat today

746
00:28:47,039 --> 00:28:51,039
and lower sex me is located in northern

747
00:28:49,120 --> 00:28:52,640
western region of germany so it's

748
00:28:51,039 --> 00:28:55,279
between

749
00:28:52,640 --> 00:28:57,440
bremen hamburg there are um

750
00:28:55,279 --> 00:29:00,640
just the neighbors of lower saxony and

751
00:28:57,440 --> 00:29:02,399
yes it's germany and yes it's 3

752
00:29:00,640 --> 00:29:04,000
30 a.m so

753
00:29:02,399 --> 00:29:06,320
if you present those times you should

754
00:29:04,000 --> 00:29:08,960
really know your stuff so

755
00:29:06,320 --> 00:29:10,000
first of all we um importantly both

756
00:29:08,960 --> 00:29:12,320
needed

757
00:29:10,000 --> 00:29:14,080
libraries panners and geopandas i have

758
00:29:12,320 --> 00:29:15,679
marked the versions over here

759
00:29:14,080 --> 00:29:18,480
and then we have to get the data of

760
00:29:15,679 --> 00:29:21,279
course because we want to work with data

761
00:29:18,480 --> 00:29:23,600
and first of all you have to

762
00:29:21,279 --> 00:29:25,679
understand some conditions to use this

763
00:29:23,600 --> 00:29:27,919
data and if you click on this link

764
00:29:25,679 --> 00:29:31,520
you will be able to download it then you

765
00:29:27,919 --> 00:29:34,080
can unzip it and you can import them

766
00:29:31,520 --> 00:29:36,080
be sure to enter this kind of projection

767
00:29:34,080 --> 00:29:38,640
because it's the known projection for

768
00:29:36,080 --> 00:29:40,399
all administrative

769
00:29:38,640 --> 00:29:43,360
um

770
00:29:40,399 --> 00:29:44,240
data sets containing gps coordinates and

771
00:29:43,360 --> 00:29:46,320
those

772
00:29:44,240 --> 00:29:48,240
five uh

773
00:29:46,320 --> 00:29:50,559
lines are the first five lines in the

774
00:29:48,240 --> 00:29:52,640
anti-data set you can see what kind of

775
00:29:50,559 --> 00:29:55,120
uh landed this this is grassland and

776
00:29:52,640 --> 00:29:57,360
this is cropland this is just a shortcut

777
00:29:55,120 --> 00:30:00,159
and this is the name of the year where

778
00:29:57,360 --> 00:30:01,520
this polygon has been input into a data

779
00:30:00,159 --> 00:30:04,480
set

780
00:30:01,520 --> 00:30:06,960
um the first shape is the just a sanity

781
00:30:04,480 --> 00:30:08,720
check how how this

782
00:30:06,960 --> 00:30:11,360
shape will look like and let's look not

783
00:30:08,720 --> 00:30:14,480
too far away from lower saxony

784
00:30:11,360 --> 00:30:16,480
and now i just added the group the kind

785
00:30:14,480 --> 00:30:19,120
of agriculture

786
00:30:16,480 --> 00:30:21,279
indicator and you can see that there

787
00:30:19,120 --> 00:30:23,200
something happens in the northern region

788
00:30:21,279 --> 00:30:26,159
and what is happening in the region you

789
00:30:23,200 --> 00:30:28,000
can see over here there is the grassland

790
00:30:26,159 --> 00:30:29,919
the greenland over here that we are

791
00:30:28,000 --> 00:30:30,880
interested in

792
00:30:29,919 --> 00:30:33,679
and

793
00:30:30,880 --> 00:30:35,919
now we know where the grassland is and

794
00:30:33,679 --> 00:30:38,159
next question is we want to see the

795
00:30:35,919 --> 00:30:39,600
distribution of land size

796
00:30:38,159 --> 00:30:42,320
grouped by the management system

797
00:30:39,600 --> 00:30:44,399
therefore pandas does a good job with

798
00:30:42,320 --> 00:30:47,039
the description function and you can see

799
00:30:44,399 --> 00:30:49,760
here that the grassland over here

800
00:30:47,039 --> 00:30:51,840
average 2.6 hectares it's a very they're

801
00:30:49,760 --> 00:30:54,240
also very tiny patches with less than

802
00:30:51,840 --> 00:30:57,360
thousand square meter but also very big

803
00:30:54,240 --> 00:31:02,720
one and this number is equivalent to

804
00:30:57,360 --> 00:31:04,960
300 soccer pitches in fifa regulation so

805
00:31:02,720 --> 00:31:07,600
it's a very huge one the last thing i'd

806
00:31:04,960 --> 00:31:10,000
like to show of course is the sum

807
00:31:07,600 --> 00:31:10,000
uh the

808
00:31:10,640 --> 00:31:17,360
oh i'm sorry uh the share of uh um

809
00:31:14,399 --> 00:31:20,720
each different agricultural systems the

810
00:31:17,360 --> 00:31:22,159
cropland has more than two-thirds of all

811
00:31:20,720 --> 00:31:23,039
of the oil

812
00:31:22,159 --> 00:31:25,120
uh

813
00:31:23,039 --> 00:31:27,440
imported shapes then grassland is coming

814
00:31:25,120 --> 00:31:30,559
with a bit more than a quarter and the

815
00:31:27,440 --> 00:31:32,960
rest is not that relevant because this

816
00:31:30,559 --> 00:31:34,640
sister does only means you don't not get

817
00:31:32,960 --> 00:31:36,880
any money out of the

818
00:31:34,640 --> 00:31:36,880
um

819
00:31:37,120 --> 00:31:44,080
field and that's for my talk today

820
00:31:41,360 --> 00:31:46,559
i try to not go to sleep thank you very

821
00:31:44,080 --> 00:31:49,200
much for uh your attention and happy

822
00:31:46,559 --> 00:31:51,840
price and happy farming thank you

823
00:31:49,200 --> 00:31:55,760
uh thank you jens for that uh for that

824
00:31:51,840 --> 00:31:57,039
great talk and yeah go go get some sleep

825
00:31:55,760 --> 00:31:59,919
yeah yeah

826
00:31:57,039 --> 00:32:03,039
you can go on

827
00:31:59,919 --> 00:32:04,880
but yeah it's it's a tour

828
00:32:03,039 --> 00:32:07,760
okay um

829
00:32:04,880 --> 00:32:09,360
up next uh i understand this is a python

830
00:32:07,760 --> 00:32:12,559
conference yes

831
00:32:09,360 --> 00:32:15,760
uh but today uh quentin is going to be

832
00:32:12,559 --> 00:32:18,480
talking to us about adders

833
00:32:15,760 --> 00:32:20,399
yes um i'm assuming it can be heard well

834
00:32:18,480 --> 00:32:24,559
um thank you

835
00:32:20,399 --> 00:32:25,840
uh chris so uh my talk today is on

836
00:32:24,559 --> 00:32:28,000
uh

837
00:32:25,840 --> 00:32:29,679
obfuscational logic

838
00:32:28,000 --> 00:32:30,799
making binary outers where they

839
00:32:29,679 --> 00:32:32,880
shouldn't be

840
00:32:30,799 --> 00:32:34,880
a bit about me my name is quentin lovett

841
00:32:32,880 --> 00:32:36,559
i'm a fifth year mechatronic engineering

842
00:32:34,880 --> 00:32:37,679
student at the university of technology

843
00:32:36,559 --> 00:32:38,960
sydney

844
00:32:37,679 --> 00:32:41,840
i work for a

845
00:32:38,960 --> 00:32:43,519
cardio electrophysiology company

846
00:32:41,840 --> 00:32:45,039
and this is my eighth

847
00:32:43,519 --> 00:32:47,600
pipe

848
00:32:45,039 --> 00:32:50,640
so what are binary adders binary adders

849
00:32:47,600 --> 00:32:51,600
are a way of performing

850
00:32:50,640 --> 00:32:54,320
addition

851
00:32:51,600 --> 00:32:56,640
with binary numbers so this is the

852
00:32:54,320 --> 00:32:59,360
simplest binary adder this is a two bit

853
00:32:56,640 --> 00:33:01,600
so you can do uh one plus one or one

854
00:32:59,360 --> 00:33:03,919
plus zero um and so this has an

855
00:33:01,600 --> 00:33:05,360
exclusive or gate um which is one or the

856
00:33:03,919 --> 00:33:08,159
other but not both

857
00:33:05,360 --> 00:33:09,519
and an and gate so the output of this

858
00:33:08,159 --> 00:33:11,679
is a

859
00:33:09,519 --> 00:33:14,480
output bit and a carry bit so one plus

860
00:33:11,679 --> 00:33:17,120
one is zero plus the carry of one and

861
00:33:14,480 --> 00:33:18,960
binary is one zero

862
00:33:17,120 --> 00:33:21,200
so getting a little bit more complicated

863
00:33:18,960 --> 00:33:23,440
than that we here have a full adder and

864
00:33:21,200 --> 00:33:25,360
so this has the carry in bit

865
00:33:23,440 --> 00:33:27,360
and then has the

866
00:33:25,360 --> 00:33:29,519
display and the carry out

867
00:33:27,360 --> 00:33:32,960
and so in python um

868
00:33:29,519 --> 00:33:35,600
we can do this simply with oren and

869
00:33:32,960 --> 00:33:37,039
so if likes cats or likes dogs and likes

870
00:33:35,600 --> 00:33:40,080
cats or

871
00:33:37,039 --> 00:33:43,279
if snake or girly boy and programs

872
00:33:40,080 --> 00:33:46,159
and looking a bit deeper into this um

873
00:33:43,279 --> 00:33:47,519
we have the python boolean

874
00:33:46,159 --> 00:33:48,880
or binary

875
00:33:47,519 --> 00:33:52,080
operators

876
00:33:48,880 --> 00:33:53,840
or being the pipe ampersand the end and

877
00:33:52,080 --> 00:33:55,600
more important and importantly for this

878
00:33:53,840 --> 00:33:58,080
one we have xor

879
00:33:55,600 --> 00:33:59,919
which is the caret

880
00:33:58,080 --> 00:34:02,880
and so why do i care

881
00:33:59,919 --> 00:34:04,880
why i care is uh addition is too easy in

882
00:34:02,880 --> 00:34:08,799
python

883
00:34:04,880 --> 00:34:08,799
so i'm going to just pull up a demo here

884
00:34:09,919 --> 00:34:12,960
um

885
00:34:11,119 --> 00:34:15,520
so when we're doing this edition in

886
00:34:12,960 --> 00:34:19,520
python um open up the python interactive

887
00:34:15,520 --> 00:34:24,720
shell get an int of 9 and into 10

888
00:34:19,520 --> 00:34:24,720
a plus b and 21. perfect

889
00:34:25,520 --> 00:34:28,000
or not

890
00:34:29,280 --> 00:34:33,200
and so when looking at this

891
00:34:31,520 --> 00:34:35,200
i just got to say

892
00:34:33,200 --> 00:34:37,119
that i have been doing some funky stuff

893
00:34:35,200 --> 00:34:39,440
um your python

894
00:34:37,119 --> 00:34:41,760
might not be exactly the same as mine

895
00:34:39,440 --> 00:34:44,159
but um now for something a little bit

896
00:34:41,760 --> 00:34:44,159
different

897
00:34:44,800 --> 00:34:48,800
so we're gonna go

898
00:34:46,480 --> 00:34:50,159
and we're to get al

899
00:34:48,800 --> 00:34:52,879
so we're going to run

900
00:34:50,159 --> 00:34:55,520
a python interactive shell again

901
00:34:52,879 --> 00:34:59,280
and we're going to do a basic uh half

902
00:34:55,520 --> 00:35:01,280
adder so that's our one xor one

903
00:34:59,280 --> 00:35:04,560
or one and one and we're going to bit

904
00:35:01,280 --> 00:35:05,440
shift the carry bit um so that we now

905
00:35:04,560 --> 00:35:06,240
have

906
00:35:05,440 --> 00:35:07,440
uh

907
00:35:06,240 --> 00:35:10,000
that in the

908
00:35:07,440 --> 00:35:12,400
twos column

909
00:35:10,000 --> 00:35:15,200
so um getting a bit more complicated

910
00:35:12,400 --> 00:35:17,040
again um we can do a

911
00:35:15,200 --> 00:35:19,280
a little bit more complicated full adder

912
00:35:17,040 --> 00:35:20,960
we're going to get 9 plus 12 so that's

913
00:35:19,280 --> 00:35:22,960
the binary values

914
00:35:20,960 --> 00:35:25,200
we're going to reverse those so that the

915
00:35:22,960 --> 00:35:26,320
zero bit is in the zero column and we're

916
00:35:25,200 --> 00:35:28,960
just going to type out a little bit of

917
00:35:26,320 --> 00:35:30,560
code it's not too much but i think you

918
00:35:28,960 --> 00:35:32,800
could see that this might might be a

919
00:35:30,560 --> 00:35:34,400
little bit too much to do um

920
00:35:32,800 --> 00:35:36,560
do every time

921
00:35:34,400 --> 00:35:38,960
um

922
00:35:36,560 --> 00:35:41,680
if that's a bit too much for you um that

923
00:35:38,960 --> 00:35:43,280
you don't want to type out all that um

924
00:35:41,680 --> 00:35:45,520
every time you want to do

925
00:35:43,280 --> 00:35:48,480
any form of addition

926
00:35:45,520 --> 00:35:51,200
we have this my piece of resistance

927
00:35:48,480 --> 00:35:53,359
the fruits of my labor

928
00:35:51,200 --> 00:35:56,960
please enter a number one two three four

929
00:35:53,359 --> 00:36:01,200
plus four three two one uh 13 bits of

930
00:35:56,960 --> 00:36:03,599
binary gates calculated for you

931
00:36:01,200 --> 00:36:03,599
thank you

932
00:36:04,480 --> 00:36:08,720
thanks for that quentin

933
00:36:07,200 --> 00:36:10,240
um

934
00:36:08,720 --> 00:36:12,240
yeah and if you want to give a lightning

935
00:36:10,240 --> 00:36:14,640
talk next year about another kind of

936
00:36:12,240 --> 00:36:18,320
snake we haven't had any talks about

937
00:36:14,640 --> 00:36:20,960
asps at this conference so uh you know

938
00:36:18,320 --> 00:36:23,520
if you know about that please sign up to

939
00:36:20,960 --> 00:36:27,119
give a lightning talk next year

940
00:36:23,520 --> 00:36:29,440
right now we have waitama who is going

941
00:36:27,119 --> 00:36:31,200
to be talking about finding parasites

942
00:36:29,440 --> 00:36:32,560
with pythons

943
00:36:31,200 --> 00:36:34,880
uh hi everyone

944
00:36:32,560 --> 00:36:35,920
thanks for having me today um yeah so

945
00:36:34,880 --> 00:36:38,400
i'm going to give this talk about

946
00:36:35,920 --> 00:36:41,760
finding parasites with python

947
00:36:38,400 --> 00:36:43,440
specifically blood parasites in turtles

948
00:36:41,760 --> 00:36:45,040
um so

949
00:36:43,440 --> 00:36:47,280
uh this is a from a project that i

950
00:36:45,040 --> 00:36:49,599
worked on with ike signing in

951
00:36:47,280 --> 00:36:51,599
our phds uh we're both now postdocs at

952
00:36:49,599 --> 00:36:54,480
the doherty institute um just by

953
00:36:51,599 --> 00:36:55,680
coincidence which is cool so this is

954
00:36:54,480 --> 00:36:57,200
from a

955
00:36:55,680 --> 00:36:59,200
um

956
00:36:57,200 --> 00:37:01,920
a project called autoscope and basically

957
00:36:59,200 --> 00:37:04,800
we wanted to build a robotic microscope

958
00:37:01,920 --> 00:37:06,800
um so that we could modify existing

959
00:37:04,800 --> 00:37:08,320
microscopes to programmatically control

960
00:37:06,800 --> 00:37:11,200
the stage so how

961
00:37:08,320 --> 00:37:13,200
how it moves around and the focus knob

962
00:37:11,200 --> 00:37:16,640
and the camera and basically with the

963
00:37:13,200 --> 00:37:18,800
goal to produce digital slides um

964
00:37:16,640 --> 00:37:19,680
for production

965
00:37:18,800 --> 00:37:21,440
for

966
00:37:19,680 --> 00:37:23,119
downstream um machine learning

967
00:37:21,440 --> 00:37:25,520
applications

968
00:37:23,119 --> 00:37:26,800
so this is sort of born out of two memes

969
00:37:25,520 --> 00:37:29,520
my incessant

970
00:37:26,800 --> 00:37:31,680
uh goal to automate everything

971
00:37:29,520 --> 00:37:34,000
um that takes longer than five minutes

972
00:37:31,680 --> 00:37:36,240
and then the fact that if you stare into

973
00:37:34,000 --> 00:37:38,400
a microscope for several hours you're

974
00:37:36,240 --> 00:37:41,280
you basically go blind so we wanted a

975
00:37:38,400 --> 00:37:43,599
way to digitalize this process

976
00:37:41,280 --> 00:37:45,040
um so a little bit about these parasites

977
00:37:43,599 --> 00:37:46,079
so

978
00:37:45,040 --> 00:37:47,440
um

979
00:37:46,079 --> 00:37:48,560
this is a freshwater turtle and they

980
00:37:47,440 --> 00:37:50,160
have this blood parasite called

981
00:37:48,560 --> 00:37:52,160
hemograins

982
00:37:50,160 --> 00:37:53,839
this is a total blood cell this is the

983
00:37:52,160 --> 00:37:55,440
nucleus here and this is the parasite

984
00:37:53,839 --> 00:37:56,880
this hot dog looking thing and we think

985
00:37:55,440 --> 00:37:58,480
they're transmitted by freshwater

986
00:37:56,880 --> 00:38:00,560
leeches but we're not really sure and

987
00:37:58,480 --> 00:38:01,920
basically you take a blood smear and

988
00:38:00,560 --> 00:38:04,160
smear out the blood and then you look

989
00:38:01,920 --> 00:38:05,760
under the microscope in that blood smear

990
00:38:04,160 --> 00:38:07,839
for these parasites so give you a couple

991
00:38:05,760 --> 00:38:09,200
of seconds to find the parasite

992
00:38:07,839 --> 00:38:11,680
in this slide

993
00:38:09,200 --> 00:38:11,680
um

994
00:38:11,839 --> 00:38:14,960
okay i'll give you a hint

995
00:38:13,760 --> 00:38:17,040
there it is

996
00:38:14,960 --> 00:38:19,119
so it can be very difficult to do this

997
00:38:17,040 --> 00:38:20,560
and basically you have to to get a good

998
00:38:19,119 --> 00:38:22,320
level of

999
00:38:20,560 --> 00:38:23,520
paracetamia like the number of parasites

1000
00:38:22,320 --> 00:38:24,800
in the blood cell you have to count

1001
00:38:23,520 --> 00:38:28,160
around

1002
00:38:24,800 --> 00:38:29,920
1500 cells per slide

1003
00:38:28,160 --> 00:38:31,520
so obviously

1004
00:38:29,920 --> 00:38:34,400
this needs some automation

1005
00:38:31,520 --> 00:38:36,560
uh so how do we go about this um

1006
00:38:34,400 --> 00:38:40,000
i designed some 3d printed parts to

1007
00:38:36,560 --> 00:38:42,160
attach to the microscope um

1008
00:38:40,000 --> 00:38:44,400
and then printed them out

1009
00:38:42,160 --> 00:38:47,920
uh the stage is controlled by a stepper

1010
00:38:44,400 --> 00:38:50,160
motor so these 28 byj48 uh stepper

1011
00:38:47,920 --> 00:38:52,320
motors and basically uh these can be

1012
00:38:50,160 --> 00:38:53,680
controlled with a raspberry pi and run

1013
00:38:52,320 --> 00:38:55,520
on the five volt power supply of the

1014
00:38:53,680 --> 00:38:57,520
rossi pi so you don't need to supply a

1015
00:38:55,520 --> 00:39:00,400
separate separate um power supply for

1016
00:38:57,520 --> 00:39:02,320
them and you just send a signal um to

1017
00:39:00,400 --> 00:39:04,560
the channels of the

1018
00:39:02,320 --> 00:39:06,400
stepper motor and you can make it spin

1019
00:39:04,560 --> 00:39:10,800
uh in a set number of steps so you can

1020
00:39:06,400 --> 00:39:14,079
control um the how far the motor turns

1021
00:39:10,800 --> 00:39:15,440
uh for the camera i used opencv the

1022
00:39:14,079 --> 00:39:18,240
python

1023
00:39:15,440 --> 00:39:20,320
interface for that uh and a raspberry pi

1024
00:39:18,240 --> 00:39:21,839
8 megapixel camera that sits on the top

1025
00:39:20,320 --> 00:39:23,280
of um

1026
00:39:21,839 --> 00:39:25,599
the eyepiece

1027
00:39:23,280 --> 00:39:26,880
uh and so we used opencv to extract a

1028
00:39:25,599 --> 00:39:30,400
region of interest from there that we

1029
00:39:26,880 --> 00:39:32,240
could then do our processing on and uh

1030
00:39:30,400 --> 00:39:35,200
using the focus knob you can kind of see

1031
00:39:32,240 --> 00:39:36,720
it uh in this slide here uh

1032
00:39:35,200 --> 00:39:38,160
down the bottom we're able to do some

1033
00:39:36,720 --> 00:39:39,839
autofocusing just by looking at the

1034
00:39:38,160 --> 00:39:42,560
sharpness of the image so moving through

1035
00:39:39,839 --> 00:39:44,960
the x plane of the z plane and looking

1036
00:39:42,560 --> 00:39:47,359
at how sharp the image is

1037
00:39:44,960 --> 00:39:48,720
for the overall architecture for the

1038
00:39:47,359 --> 00:39:51,839
software basically there's this

1039
00:39:48,720 --> 00:39:53,680
sportoscope class that has

1040
00:39:51,839 --> 00:39:55,599
classes for the stage the focus and the

1041
00:39:53,680 --> 00:39:57,599
camera and this has two functions an

1042
00:39:55,599 --> 00:39:59,920
auto focus function

1043
00:39:57,599 --> 00:40:01,440
uh that enables you to focus uh

1044
00:39:59,920 --> 00:40:03,119
automatically focus through the plane of

1045
00:40:01,440 --> 00:40:05,520
the slide and then auto scan which

1046
00:40:03,119 --> 00:40:07,440
basically does this snake like pattern

1047
00:40:05,520 --> 00:40:10,319
um to produce

1048
00:40:07,440 --> 00:40:14,240
uh tiled uh output

1049
00:40:10,319 --> 00:40:14,240
so here's an example of it running

1050
00:40:16,640 --> 00:40:21,119
so yeah you can see the raspberry pi the

1051
00:40:18,240 --> 00:40:25,560
step motor in the background there

1052
00:40:21,119 --> 00:40:25,560
and then this is the slide on the stage

1053
00:40:31,599 --> 00:40:37,119
one minute

1054
00:40:34,240 --> 00:40:40,160
yep and then so it's moving along and so

1055
00:40:37,119 --> 00:40:44,000
to view the slide we use this gdel2 um

1056
00:40:40,160 --> 00:40:46,240
slides and i'll just open this quickly

1057
00:40:44,000 --> 00:40:48,079
so here's the slide

1058
00:40:46,240 --> 00:40:50,960
so you can zoom in here you can see it

1059
00:40:48,079 --> 00:40:52,720
loads there's a parasite and another

1060
00:40:50,960 --> 00:40:55,760
parasite

1061
00:40:52,720 --> 00:40:58,400
um just zoom out so yeah you can see the

1062
00:40:55,760 --> 00:40:58,400
slides there

1063
00:40:59,440 --> 00:41:04,560
so what's next to do take that data to

1064
00:41:01,680 --> 00:41:06,240
some ml so we can automatically use that

1065
00:41:04,560 --> 00:41:08,560
system to identify the parasites and

1066
00:41:06,240 --> 00:41:10,400
then uh train uh

1067
00:41:08,560 --> 00:41:12,720
a model to automatically identify these

1068
00:41:10,400 --> 00:41:14,800
parasites um and so yeah this is all

1069
00:41:12,720 --> 00:41:15,920
open source the code is on github and

1070
00:41:14,800 --> 00:41:18,400
the 3d

1071
00:41:15,920 --> 00:41:21,760
uh components are on thingiverse

1072
00:41:18,400 --> 00:41:24,800
cool thank you

1073
00:41:21,760 --> 00:41:26,480
thank you so much i for one am very glad

1074
00:41:24,800 --> 00:41:28,400
that there is another parasite talk

1075
00:41:26,480 --> 00:41:31,680
happening at lightning talks at like

1076
00:41:28,400 --> 00:41:33,680
online are you um up next we have zach

1077
00:41:31,680 --> 00:41:36,319
he has written code in order to write

1078
00:41:33,680 --> 00:41:36,319
your tests

1079
00:41:37,599 --> 00:41:41,280
thanks everyone i'm joining you from not

1080
00:41:39,440 --> 00:41:43,280
all country in canberra and i'm

1081
00:41:41,280 --> 00:41:44,960
speedrunning a 12-minute live demo in

1082
00:41:43,280 --> 00:41:46,560
five minutes so let's get started

1083
00:41:44,960 --> 00:41:48,480
hypothesis is a tool for people who

1084
00:41:46,560 --> 00:41:50,000
don't like writing tests because coming

1085
00:41:48,480 --> 00:41:52,160
up with all your test data is boring so

1086
00:41:50,000 --> 00:41:53,680
hypothesis does that for you i also got

1087
00:41:52,160 --> 00:41:55,200
bored writing hypothesis tests so i

1088
00:41:53,680 --> 00:41:56,480
wrote code to write hypothesis tests so

1089
00:41:55,200 --> 00:41:58,160
that you don't have to write tests or

1090
00:41:56,480 --> 00:42:00,480
write tests

1091
00:41:58,160 --> 00:42:02,000
to save for a writing test anyway you

1092
00:42:00,480 --> 00:42:03,599
can just tell hypothesis to write tests

1093
00:42:02,000 --> 00:42:05,359
for a sort function and hypothesis will

1094
00:42:03,599 --> 00:42:06,720
tell you there is no sort function if

1095
00:42:05,359 --> 00:42:08,560
you tell a hypothesis to write a test

1096
00:42:06,720 --> 00:42:09,599
for the sort end function it does that

1097
00:42:08,560 --> 00:42:11,599
it tells you that there's three

1098
00:42:09,599 --> 00:42:13,119
arguments here's some default values for

1099
00:42:11,599 --> 00:42:14,880
how to generate it and then it calls it

1100
00:42:13,119 --> 00:42:16,720
but just calling the sort function might

1101
00:42:14,880 --> 00:42:18,079
not crash because the python chord downs

1102
00:42:16,720 --> 00:42:19,599
aren't incompetent so you might want to

1103
00:42:18,079 --> 00:42:20,960
check if you sort the result of sorting

1104
00:42:19,599 --> 00:42:22,480
something it doesn't make any further

1105
00:42:20,960 --> 00:42:24,240
changes you do that just by telling

1106
00:42:22,480 --> 00:42:26,000
hypothesis that sorting is important

1107
00:42:24,240 --> 00:42:27,680
which it should be

1108
00:42:26,000 --> 00:42:28,880
if you want to check something else you

1109
00:42:27,680 --> 00:42:30,160
can check the two functions are

1110
00:42:28,880 --> 00:42:31,920
equivalent and if you want to check that

1111
00:42:30,160 --> 00:42:33,520
the ast.liter level function is

1112
00:42:31,920 --> 00:42:35,280
equivalent to the eval built-in at least

1113
00:42:33,520 --> 00:42:36,960
for all source code strings which

1114
00:42:35,280 --> 00:42:38,640
represent python literals you can tell

1115
00:42:36,960 --> 00:42:40,800
hypothesis to do that with just one

1116
00:42:38,640 --> 00:42:42,720
convenient line of terminal code which

1117
00:42:40,800 --> 00:42:44,319
is great and it tells you this and you

1118
00:42:42,720 --> 00:42:45,520
can see that hypothesis doesn't realize

1119
00:42:44,319 --> 00:42:47,359
that the

1120
00:42:45,520 --> 00:42:49,040
node or string argument is different

1121
00:42:47,359 --> 00:42:50,560
from the source code argument when they

1122
00:42:49,040 --> 00:42:52,000
should be the same so you'd have to edit

1123
00:42:50,560 --> 00:42:52,960
this but it's a great start to the test

1124
00:42:52,000 --> 00:42:55,119
right

1125
00:42:52,960 --> 00:42:56,880
or if you think about property-based

1126
00:42:55,119 --> 00:42:59,200
testing means algebraic properties of

1127
00:42:56,880 --> 00:43:00,640
binary operators like haskell hypothesis

1128
00:42:59,200 --> 00:43:03,440
could do that too i'm not sure why you'd

1129
00:43:00,640 --> 00:43:05,040
want to but it can so that's pretty cool

1130
00:43:03,440 --> 00:43:06,960
the most useful type though are what we

1131
00:43:05,040 --> 00:43:09,040
call ground trip tests where you can

1132
00:43:06,960 --> 00:43:10,640
tell hypothesis that you want to test

1133
00:43:09,040 --> 00:43:12,480
for gzip compression and it will

1134
00:43:10,640 --> 00:43:14,079
automatically work out that if a module

1135
00:43:12,480 --> 00:43:15,920
has a compressed function that also has

1136
00:43:14,079 --> 00:43:17,359
a decompress function if you compress

1137
00:43:15,920 --> 00:43:19,040
something then decompressing it should

1138
00:43:17,359 --> 00:43:20,400
probably give you the same thing back so

1139
00:43:19,040 --> 00:43:22,480
that's pretty cool all you need to do

1140
00:43:20,400 --> 00:43:24,400
here is tell it that data is a byte

1141
00:43:22,480 --> 00:43:26,720
string and that's a great test for g

1142
00:43:24,400 --> 00:43:28,319
compression

1143
00:43:26,720 --> 00:43:31,040
where it gets more complicated though is

1144
00:43:28,319 --> 00:43:32,720
json json it turns out is terrifyingly

1145
00:43:31,040 --> 00:43:35,200
complicated there's so many arguments

1146
00:43:32,720 --> 00:43:36,400
here that i have no idea about really

1147
00:43:35,200 --> 00:43:38,160
and

1148
00:43:36,400 --> 00:43:39,599
i that's a lie i do have some idea about

1149
00:43:38,160 --> 00:43:40,400
them but i don't have time to go into

1150
00:43:39,599 --> 00:43:42,160
them now

1151
00:43:40,400 --> 00:43:43,680
so i wrote a shorter version of this

1152
00:43:42,160 --> 00:43:46,240
test just by deleting a whole bunch of

1153
00:43:43,680 --> 00:43:48,640
arguments and describing json data to

1154
00:43:46,240 --> 00:43:50,640
hypothesis the description of json is

1155
00:43:48,640 --> 00:43:52,800
very simple recursive data the base case

1156
00:43:50,640 --> 00:43:54,720
is either none true or false a number or

1157
00:43:52,800 --> 00:43:56,560
a string or you can have a list of json

1158
00:43:54,720 --> 00:43:57,920
things or you can have a dictionary json

1159
00:43:56,560 --> 00:44:00,000
calls this an object but that's a

1160
00:43:57,920 --> 00:44:02,319
reserved term in python a dictionary

1161
00:44:00,000 --> 00:44:03,839
where you have string keys and json

1162
00:44:02,319 --> 00:44:05,280
values

1163
00:44:03,839 --> 00:44:06,720
and we're just going to tell hypothesis

1164
00:44:05,280 --> 00:44:08,400
that like whether or not we allow nand

1165
00:44:06,720 --> 00:44:10,079
is varied whether we check for circular

1166
00:44:08,400 --> 00:44:12,079
things whether we want ascii or unicode

1167
00:44:10,079 --> 00:44:13,760
output is all variable and then like do

1168
00:44:12,079 --> 00:44:15,119
we want to sort the keys to the objects

1169
00:44:13,760 --> 00:44:16,960
also varied

1170
00:44:15,119 --> 00:44:19,280
so let's see what happens when we run

1171
00:44:16,960 --> 00:44:22,599
this test

1172
00:44:19,280 --> 00:44:22,599
any guesses

1173
00:44:26,960 --> 00:44:30,800
they're on a 30-second delay zach

1174
00:44:46,960 --> 00:44:51,040
it doesn't usually take this

1175
00:44:49,760 --> 00:44:53,599
i guess this is what i get for doing a

1176
00:44:51,040 --> 00:44:53,599
live demo

1177
00:44:55,760 --> 00:44:59,119
yeah anyway let's just uh cancel that

1178
00:44:57,839 --> 00:45:01,200
and

1179
00:44:59,119 --> 00:45:03,040
it it turns out that this fails because

1180
00:45:01,200 --> 00:45:04,800
uh the allowed ad argument also

1181
00:45:03,040 --> 00:45:06,160
disallows infinity because non-finite

1182
00:45:04,800 --> 00:45:07,760
numbers aren't technically allowed by

1183
00:45:06,160 --> 00:45:09,680
the json spec and if you have not a

1184
00:45:07,760 --> 00:45:11,520
number not a number is unequal to itself

1185
00:45:09,680 --> 00:45:13,440
so the test fails so we'll just say that

1186
00:45:11,520 --> 00:45:15,119
we always allow non-finite numbers we're

1187
00:45:13,440 --> 00:45:16,640
default argument name but there we are

1188
00:45:15,119 --> 00:45:18,160
and we'll just assume that's like a

1189
00:45:16,640 --> 00:45:20,400
special assert that tells hypothesis

1190
00:45:18,160 --> 00:45:24,599
it's a bad example that it's equal if we

1191
00:45:20,400 --> 00:45:24,599
run this one hopefully it'll be faster

1192
00:45:44,480 --> 00:45:49,839
one minute

1193
00:45:47,200 --> 00:45:52,079
thanks chris all right so this one fails

1194
00:45:49,839 --> 00:45:53,760
with a really confusing error message it

1195
00:45:52,079 --> 00:45:55,760
turns out that a list containing not a

1196
00:45:53,760 --> 00:45:57,359
number compares equal to itself until

1197
00:45:55,760 --> 00:45:58,720
you round trip it through json and

1198
00:45:57,359 --> 00:46:01,200
that's because there's this list of

1199
00:45:58,720 --> 00:46:03,119
quality performance optimization in c

1200
00:46:01,200 --> 00:46:04,240
python where if a list is the other list

1201
00:46:03,119 --> 00:46:06,000
it doesn't bother checking whether the

1202
00:46:04,240 --> 00:46:08,560
elements are equal and it turns out that

1203
00:46:06,000 --> 00:46:10,160
it also starts by comparing elements by

1204
00:46:08,560 --> 00:46:11,680
identity to avoid deeply nested

1205
00:46:10,160 --> 00:46:13,200
comparisons great performance

1206
00:46:11,680 --> 00:46:15,599
optimization but it totally breaks to

1207
00:46:13,200 --> 00:46:17,040
the presence of dead so the real fix for

1208
00:46:15,599 --> 00:46:18,480
this test is just to tell hypothesis

1209
00:46:17,040 --> 00:46:20,079
that we don't actually want to deal with

1210
00:46:18,480 --> 00:46:21,520
not a number because you never want to

1211
00:46:20,079 --> 00:46:24,000
deal with not a number it's an awful

1212
00:46:21,520 --> 00:46:25,520
value and if we run the final real test

1213
00:46:24,000 --> 00:46:29,480
then this one should pass hopefully

1214
00:46:25,520 --> 00:46:29,480
within the next 19 seconds

1215
00:46:32,480 --> 00:46:35,839
and there we have it that's hypothesis

1216
00:46:34,240 --> 00:46:38,640
you could use it to avoid writing tests

1217
00:46:35,839 --> 00:46:40,240
and still fight bugs thank you very much

1218
00:46:38,640 --> 00:46:42,640
i was waiting for that last one with

1219
00:46:40,240 --> 00:46:44,240
baited breath i'm i'm glad you finished

1220
00:46:42,640 --> 00:46:45,920
within the five minutes today zach good

1221
00:46:44,240 --> 00:46:48,880
job okay

1222
00:46:45,920 --> 00:46:51,520
up next uh we have a

1223
00:46:48,880 --> 00:46:56,079
a slight escalation because riley is

1224
00:46:51,520 --> 00:46:58,400
going to tell us where the stairs are

1225
00:46:56,079 --> 00:47:00,240
alrighty hello pycon i gave a talk last

1226
00:46:58,400 --> 00:47:02,640
year which coincided with the public

1227
00:47:00,240 --> 00:47:05,119
debut of a package called staircase and

1228
00:47:02,640 --> 00:47:07,119
this is a five-minute update a

1229
00:47:05,119 --> 00:47:09,200
where are they now and to be fair who

1230
00:47:07,119 --> 00:47:10,800
amongst us couldn't sum up the last year

1231
00:47:09,200 --> 00:47:12,880
in five minutes

1232
00:47:10,800 --> 00:47:14,400
firstly though let's quickly revisit

1233
00:47:12,880 --> 00:47:16,480
what the package is and for that i'm

1234
00:47:14,400 --> 00:47:19,040
going to focus on one use case which is

1235
00:47:16,480 --> 00:47:20,720
where we have timestamp data now you're

1236
00:47:19,040 --> 00:47:23,839
a python audience so when i say

1237
00:47:20,720 --> 00:47:25,520
timestamp data you probably think pandas

1238
00:47:23,839 --> 00:47:28,160
so to set the scene i'm going to perform

1239
00:47:25,520 --> 00:47:30,800
a monologue from the pandas website

1240
00:47:28,160 --> 00:47:32,800
timestamp data is the most basic time of

1241
00:47:30,800 --> 00:47:35,280
time series data that associates values

1242
00:47:32,800 --> 00:47:38,480
with points in time for pandas objects

1243
00:47:35,280 --> 00:47:40,559
it means using the points in time

1244
00:47:38,480 --> 00:47:42,880
now those points in time are implemented

1245
00:47:40,559 --> 00:47:44,400
in pandas by the time stamp class and

1246
00:47:42,880 --> 00:47:46,160
you can model a time series with a date

1247
00:47:44,400 --> 00:47:48,480
time index which is an index of time

1248
00:47:46,160 --> 00:47:50,960
stamps and a vector of values and we

1249
00:47:48,480 --> 00:47:53,280
commonly assign this to the x and y axis

1250
00:47:50,960 --> 00:47:54,720
respectively but they also have this

1251
00:47:53,280 --> 00:47:56,480
period class which represents a

1252
00:47:54,720 --> 00:47:58,319
particular span of time

1253
00:47:56,480 --> 00:48:00,640
like a particular error or a particular

1254
00:47:58,319 --> 00:48:02,319
day or a particular week and there's

1255
00:48:00,640 --> 00:48:04,400
also this period index and i'm sure you

1256
00:48:02,319 --> 00:48:06,480
can guess what that means now if you're

1257
00:48:04,400 --> 00:48:09,040
pairing a period index with some values

1258
00:48:06,480 --> 00:48:12,160
and the periods don't overlap each other

1259
00:48:09,040 --> 00:48:14,559
then what you have is a step function

1260
00:48:12,160 --> 00:48:16,400
so it's a function over a time domain in

1261
00:48:14,559 --> 00:48:19,359
this case which is piecewise constant

1262
00:48:16,400 --> 00:48:21,760
and it's made up of horizontal lines

1263
00:48:19,359 --> 00:48:24,160
but the periods in a period index are

1264
00:48:21,760 --> 00:48:26,960
all the same size and they're limited to

1265
00:48:24,160 --> 00:48:29,119
specific frequencies like hour or day

1266
00:48:26,960 --> 00:48:32,319
and so you are constrained in what you

1267
00:48:29,119 --> 00:48:34,960
can do but this is where staircase steps

1268
00:48:32,319 --> 00:48:37,599
in pun intended so it doesn't have those

1269
00:48:34,960 --> 00:48:40,480
limits staircase is for step functions

1270
00:48:37,599 --> 00:48:42,559
set functions over a time domain step

1271
00:48:40,480 --> 00:48:45,520
functions over the real numbers

1272
00:48:42,559 --> 00:48:47,680
and assigning values to intervals over

1273
00:48:45,520 --> 00:48:50,559
domain rather than single points opens

1274
00:48:47,680 --> 00:48:52,880
you up to thinking about problems and

1275
00:48:50,559 --> 00:48:55,280
their solutions and analysis in a

1276
00:48:52,880 --> 00:48:56,880
different kind of way

1277
00:48:55,280 --> 00:48:58,800
now there's lots of fun things that you

1278
00:48:56,880 --> 00:49:00,240
can do with step functions much of the

1279
00:48:58,800 --> 00:49:02,400
same things that you can do with

1280
00:49:00,240 --> 00:49:04,640
ordinary time series or functions in

1281
00:49:02,400 --> 00:49:06,079
general you can add them you can divide

1282
00:49:04,640 --> 00:49:07,599
them you can compare them to see which

1283
00:49:06,079 --> 00:49:10,000
one is bigger than another you can

1284
00:49:07,599 --> 00:49:12,000
calculate their mean or median or

1285
00:49:10,000 --> 00:49:12,960
analyze the distribution of values and

1286
00:49:12,000 --> 00:49:15,040
so on

1287
00:49:12,960 --> 00:49:17,280
and you can convert them to an ordinary

1288
00:49:15,040 --> 00:49:18,960
time series if you wish by sampling the

1289
00:49:17,280 --> 00:49:21,119
step function at regular points or

1290
00:49:18,960 --> 00:49:22,480
aggregating it over a sequence of

1291
00:49:21,119 --> 00:49:23,599
periods

1292
00:49:22,480 --> 00:49:24,800
so

1293
00:49:23,599 --> 00:49:27,200
what's new

1294
00:49:24,800 --> 00:49:29,040
there was a major release a month ago

1295
00:49:27,200 --> 00:49:31,440
which saw every part of the library

1296
00:49:29,040 --> 00:49:34,079
rewritten to use pandas and numpy under

1297
00:49:31,440 --> 00:49:36,960
the hood so it's a lot faster it has

1298
00:49:34,079 --> 00:49:38,640
become very closely aligned with pandas

1299
00:49:36,960 --> 00:49:42,000
stylistically and there's a lot of

1300
00:49:38,640 --> 00:49:43,359
features which have analogies in pandas

1301
00:49:42,000 --> 00:49:46,000
for example two of the biggest

1302
00:49:43,359 --> 00:49:48,480
enhancements in version two are slicing

1303
00:49:46,000 --> 00:49:51,359
and masking so with slicing you start

1304
00:49:48,480 --> 00:49:52,640
with a step function you cut it up into

1305
00:49:51,359 --> 00:49:54,319
slices

1306
00:49:52,640 --> 00:49:56,559
and then you perform some calculation

1307
00:49:54,319 --> 00:49:58,800
over the slices and combine the results

1308
00:49:56,559 --> 00:50:01,520
and this is analogous to

1309
00:49:58,800 --> 00:50:04,720
group bias in pandas

1310
00:50:01,520 --> 00:50:07,520
the other is masking which is analogous

1311
00:50:04,720 --> 00:50:09,760
in pandas to masking and

1312
00:50:07,520 --> 00:50:12,240
we all understand what masking is so for

1313
00:50:09,760 --> 00:50:13,359
step functions this means allowing

1314
00:50:12,240 --> 00:50:15,520
intervals

1315
00:50:13,359 --> 00:50:16,880
where the step function is undefined or

1316
00:50:15,520 --> 00:50:18,640
doesn't exist

1317
00:50:16,880 --> 00:50:19,680
and this is a great feature to have for

1318
00:50:18,640 --> 00:50:21,680
example

1319
00:50:19,680 --> 00:50:23,760
i may have a step function which

1320
00:50:21,680 --> 00:50:24,880
represents the heart rate of a patient

1321
00:50:23,760 --> 00:50:26,880
over time

1322
00:50:24,880 --> 00:50:29,280
now if their heart rate never changes

1323
00:50:26,880 --> 00:50:31,200
then this step function is a flat line

1324
00:50:29,280 --> 00:50:33,119
and let's face it it probably has a

1325
00:50:31,200 --> 00:50:34,640
value of zero

1326
00:50:33,119 --> 00:50:36,400
but if they're not flat lining then

1327
00:50:34,640 --> 00:50:38,480
analyzing the distribution of the values

1328
00:50:36,400 --> 00:50:41,359
might be an interesting thing to do and

1329
00:50:38,480 --> 00:50:43,839
this is very easy with staircase

1330
00:50:41,359 --> 00:50:46,240
but maybe i want to exclude the times

1331
00:50:43,839 --> 00:50:48,000
that the patient is asleep

1332
00:50:46,240 --> 00:50:50,559
and that could happen at

1333
00:50:48,000 --> 00:50:52,480
random times for arbitrary intervals

1334
00:50:50,559 --> 00:50:54,880
especially if they're sharing a room in

1335
00:50:52,480 --> 00:50:57,119
a public hospital and if i happen to

1336
00:50:54,880 --> 00:50:58,079
know when the patient was

1337
00:50:57,119 --> 00:50:59,599
sleeping

1338
00:50:58,079 --> 00:51:02,319
and i knew when they were awake like

1339
00:50:59,599 --> 00:51:05,200
some sort of creepy analytical santa

1340
00:51:02,319 --> 00:51:07,599
claus then it's very easy to make to

1341
00:51:05,200 --> 00:51:10,800
mask those intervals of time and ignore

1342
00:51:07,599 --> 00:51:13,920
them when performing the analysis

1343
00:51:10,800 --> 00:51:16,240
so it's a pretty full featured package

1344
00:51:13,920 --> 00:51:18,880
there is always more to be done

1345
00:51:16,240 --> 00:51:20,559
one of today's social events is the

1346
00:51:18,880 --> 00:51:23,119
staircase sprint

1347
00:51:20,559 --> 00:51:25,920
it's not really a sprint because it will

1348
00:51:23,119 --> 00:51:28,240
just go on forever you can join whenever

1349
00:51:25,920 --> 00:51:29,040
you like even tomorrow or the weeks to

1350
00:51:28,240 --> 00:51:31,520
come

1351
00:51:29,040 --> 00:51:34,319
i've put a backlog of tasks many of

1352
00:51:31,520 --> 00:51:36,240
which are beginner friendly um

1353
00:51:34,319 --> 00:51:38,640
in a repository so if you're wanting to

1354
00:51:36,240 --> 00:51:40,800
dip your toe in open source contribution

1355
00:51:38,640 --> 00:51:45,839
and you don't know where to start then

1356
00:51:40,800 --> 00:51:45,839
uh come along and um get wet thank you

1357
00:51:46,000 --> 00:51:52,240
thanks for that riley

1358
00:51:47,920 --> 00:51:55,520
okay up next uh we have evan who is

1359
00:51:52,240 --> 00:51:58,160
going to be telling us about uh email

1360
00:51:55,520 --> 00:51:58,160
apparently

1361
00:51:58,240 --> 00:52:02,079
hello

1362
00:51:59,680 --> 00:52:02,839
uh this is the story of

1363
00:52:02,079 --> 00:52:07,200
my

1364
00:52:02,839 --> 00:52:10,720
email this is my email

1365
00:52:07,200 --> 00:52:12,960
this is also my email

1366
00:52:10,720 --> 00:52:16,079
this one too

1367
00:52:12,960 --> 00:52:19,200
oh and this one as well

1368
00:52:16,079 --> 00:52:20,640
these are all different emails

1369
00:52:19,200 --> 00:52:23,440
and

1370
00:52:20,640 --> 00:52:26,720
they're all my emails

1371
00:52:23,440 --> 00:52:30,160
but i only have one email

1372
00:52:26,720 --> 00:52:32,960
and that is this email

1373
00:52:30,160 --> 00:52:36,559
how are they all my emails

1374
00:52:32,960 --> 00:52:39,200
through the power of plus addressing

1375
00:52:36,559 --> 00:52:40,880
you simply take your email address

1376
00:52:39,200 --> 00:52:42,559
you add a plus

1377
00:52:40,880 --> 00:52:45,599
and your tag

1378
00:52:42,559 --> 00:52:47,839
and it'll appear in your inbox

1379
00:52:45,599 --> 00:52:48,960
and you can even add dots anywhere you

1380
00:52:47,839 --> 00:52:51,520
want

1381
00:52:48,960 --> 00:52:54,000
and they'll all be ignored

1382
00:52:51,520 --> 00:52:56,720
and this isn't just for gmail

1383
00:52:54,000 --> 00:52:58,000
it also works for outlook

1384
00:52:56,720 --> 00:52:59,599
yahoo

1385
00:52:58,000 --> 00:53:01,200
protonmail

1386
00:52:59,599 --> 00:53:03,359
and fastmail

1387
00:53:01,200 --> 00:53:05,599
and many more

1388
00:53:03,359 --> 00:53:07,760
fastmail even supports subdomain

1389
00:53:05,599 --> 00:53:09,760
addressing and they're great because

1390
00:53:07,760 --> 00:53:11,920
they look just like regular email

1391
00:53:09,760 --> 00:53:12,839
addresses

1392
00:53:11,920 --> 00:53:14,480
but

1393
00:53:12,839 --> 00:53:17,440
why

1394
00:53:14,480 --> 00:53:20,400
well these tags let you be in control of

1395
00:53:17,440 --> 00:53:22,160
how you feel to your email regardless of

1396
00:53:20,400 --> 00:53:24,640
who it's from

1397
00:53:22,160 --> 00:53:26,880
and they also let you easily register

1398
00:53:24,640 --> 00:53:29,200
accounts for the same service without

1399
00:53:26,880 --> 00:53:31,280
having to make a whole new email

1400
00:53:29,200 --> 00:53:34,240
meaning that you can sign up for as many

1401
00:53:31,280 --> 00:53:36,319
freebies from that website as you want

1402
00:53:34,240 --> 00:53:38,480
but more importantly it means that you

1403
00:53:36,319 --> 00:53:40,800
can better protect your email from the

1404
00:53:38,480 --> 00:53:43,200
ever increasing companies that leak it

1405
00:53:40,800 --> 00:53:45,440
letting you know who leaked your email

1406
00:53:43,200 --> 00:53:47,680
and giving you the option to filter it

1407
00:53:45,440 --> 00:53:49,599
all to spam

1408
00:53:47,680 --> 00:53:52,880
you can even test if your email has been

1409
00:53:49,599 --> 00:53:53,920
leaked at having beenpwned.com

1410
00:53:52,880 --> 00:53:57,119
you can see

1411
00:53:53,920 --> 00:54:00,079
my email is fine but one day i got a

1412
00:53:57,119 --> 00:54:02,640
spam email and sure enough i confirmed

1413
00:54:00,079 --> 00:54:05,119
that a dodgy website leaked it

1414
00:54:02,640 --> 00:54:07,200
but i don't care because it's not my

1415
00:54:05,119 --> 00:54:09,839
main email

1416
00:54:07,200 --> 00:54:11,760
lastly it's also just really funny to

1417
00:54:09,839 --> 00:54:13,920
just give out your email now

1418
00:54:11,760 --> 00:54:17,200
i've heard on multiple occasions

1419
00:54:13,920 --> 00:54:20,319
wow you made an email just for me

1420
00:54:17,200 --> 00:54:21,280
uh are you sure that's correct

1421
00:54:20,319 --> 00:54:25,280
and

1422
00:54:21,280 --> 00:54:28,079
oh i didn't know you work for us

1423
00:54:25,280 --> 00:54:29,760
there are three things to note though

1424
00:54:28,079 --> 00:54:31,920
you will have to come up with some kind

1425
00:54:29,760 --> 00:54:33,839
of system for generating and remembering

1426
00:54:31,920 --> 00:54:36,160
these tags

1427
00:54:33,839 --> 00:54:38,319
and unfortunately this has been used in

1428
00:54:36,160 --> 00:54:40,640
phishing like scams like this netflix

1429
00:54:38,319 --> 00:54:43,280
one where attackers could sign up with

1430
00:54:40,640 --> 00:54:45,359
someone else's similar email address and

1431
00:54:43,280 --> 00:54:47,520
have the victim think that their own

1432
00:54:45,359 --> 00:54:49,839
account needed updating when in fact it

1433
00:54:47,520 --> 00:54:52,720
was the attackers

1434
00:54:49,839 --> 00:54:54,640
and finally just like people not all

1435
00:54:52,720 --> 00:54:57,839
websites know about this cool trick

1436
00:54:54,640 --> 00:55:00,720
which can be a bit annoying at times

1437
00:54:57,839 --> 00:55:02,960
so spread the word if you're a web dev

1438
00:55:00,720 --> 00:55:05,280
stop rolling your own email validation

1439
00:55:02,960 --> 00:55:07,359
just use a library instead

1440
00:55:05,280 --> 00:55:10,480
and if you're an email dev

1441
00:55:07,359 --> 00:55:11,359
implement subdomain addressing

1442
00:55:10,480 --> 00:55:13,280
oh

1443
00:55:11,359 --> 00:55:14,799
and one more thing

1444
00:55:13,280 --> 00:55:17,280
this

1445
00:55:14,799 --> 00:55:19,680
is not my email

1446
00:55:17,280 --> 00:55:22,079
this is my email

1447
00:55:19,680 --> 00:55:23,760
i don't give out my email publicly and

1448
00:55:22,079 --> 00:55:25,839
neither should you

1449
00:55:23,760 --> 00:55:27,359
that is the best way to protect your

1450
00:55:25,839 --> 00:55:29,760
privacy

1451
00:55:27,359 --> 00:55:29,760
thank you

1452
00:55:30,640 --> 00:55:36,160
thanks for that evan i i was just about

1453
00:55:33,280 --> 00:55:38,400
to ask if evan gmail was your your email

1454
00:55:36,160 --> 00:55:42,319
address and that would explain why

1455
00:55:38,400 --> 00:55:43,920
uh why i had a bounce earlier on but uh

1456
00:55:42,319 --> 00:55:48,920
okay uh

1457
00:55:43,920 --> 00:55:48,920
up next we have apollo

1458
00:55:49,599 --> 00:55:55,119
so let's go

1459
00:55:52,319 --> 00:55:58,000
hello everyone and welcome to my gender

1460
00:55:55,119 --> 00:55:59,359
reveal party my quantum gender reveal

1461
00:55:58,000 --> 00:56:01,680
party

1462
00:55:59,359 --> 00:56:03,920
today i'm going to explain why my gender

1463
00:56:01,680 --> 00:56:05,920
is a quantum mechanical object and then

1464
00:56:03,920 --> 00:56:09,280
we're going to measure it together using

1465
00:56:05,920 --> 00:56:11,599
a real quantum computer

1466
00:56:09,280 --> 00:56:13,599
so some people have a very strong

1467
00:56:11,599 --> 00:56:16,000
internal sense of their gender

1468
00:56:13,599 --> 00:56:17,440
knowing that they're a woman or a man or

1469
00:56:16,000 --> 00:56:19,040
something else

1470
00:56:17,440 --> 00:56:21,440
these genders behave in ways that we're

1471
00:56:19,040 --> 00:56:24,160
used to so i like to think of them as

1472
00:56:21,440 --> 00:56:26,480
the classical genders

1473
00:56:24,160 --> 00:56:28,000
unfortunately my gender has never been

1474
00:56:26,480 --> 00:56:29,520
that simple

1475
00:56:28,000 --> 00:56:31,760
i've been trying to find a nice label to

1476
00:56:29,520 --> 00:56:34,480
slap on it for a few years now and the

1477
00:56:31,760 --> 00:56:36,480
best i've found is gender void

1478
00:56:34,480 --> 00:56:38,640
but even that's not right sometimes my

1479
00:56:36,480 --> 00:56:41,280
gender doesn't feel like a void at all

1480
00:56:38,640 --> 00:56:44,000
sometimes i'm a soft guy

1481
00:56:41,280 --> 00:56:46,319
sometimes a screaming magpie coming for

1482
00:56:44,000 --> 00:56:48,160
your helmet

1483
00:56:46,319 --> 00:56:50,400
before i worked as a tech consultant i

1484
00:56:48,160 --> 00:56:52,400
studied theoretical physics and there i

1485
00:56:50,400 --> 00:56:53,520
discovered the wonderful world of

1486
00:56:52,400 --> 00:56:55,920
quantum

1487
00:56:53,520 --> 00:56:58,799
and my gender

1488
00:56:55,920 --> 00:57:00,720
so what does it mean to be quantum

1489
00:56:58,799 --> 00:57:02,799
quantum mechanics is just a model we use

1490
00:57:00,720 --> 00:57:04,720
in physics to describe really small

1491
00:57:02,799 --> 00:57:06,400
things like atoms

1492
00:57:04,720 --> 00:57:08,720
but there's no reason it couldn't apply

1493
00:57:06,400 --> 00:57:10,480
to bigger things too

1494
00:57:08,720 --> 00:57:12,400
fundamentally something is quantum

1495
00:57:10,480 --> 00:57:14,720
mechanical when it displays strange

1496
00:57:12,400 --> 00:57:16,319
behaviors like superposition and

1497
00:57:14,720 --> 00:57:18,079
entanglement

1498
00:57:16,319 --> 00:57:20,240
unfortunately i only have time to talk

1499
00:57:18,079 --> 00:57:23,599
about superposition for now but feel

1500
00:57:20,240 --> 00:57:26,480
free to ask about entanglement later

1501
00:57:23,599 --> 00:57:28,720
so superposition means you can be in

1502
00:57:26,480 --> 00:57:30,640
multiple states at once

1503
00:57:28,720 --> 00:57:31,520
kind of

1504
00:57:30,640 --> 00:57:33,440
it's

1505
00:57:31,520 --> 00:57:35,839
classically explained using

1506
00:57:33,440 --> 00:57:38,319
schrodinger's cat where his abused cat

1507
00:57:35,839 --> 00:57:41,280
is in a superposition of being alive and

1508
00:57:38,319 --> 00:57:43,359
dead until you open the box and measure

1509
00:57:41,280 --> 00:57:44,480
the cap

1510
00:57:43,359 --> 00:57:46,240
but really

1511
00:57:44,480 --> 00:57:48,960
a better way to think of it is that your

1512
00:57:46,240 --> 00:57:50,400
state is blurred across multiple values

1513
00:57:48,960 --> 00:57:52,880
with different likelihoods of a

1514
00:57:50,400 --> 00:57:55,599
particular state being measured

1515
00:57:52,880 --> 00:57:57,599
this blurry probability blob is called a

1516
00:57:55,599 --> 00:58:01,359
wave function

1517
00:57:57,599 --> 00:58:03,680
now that sounds just like my gender

1518
00:58:01,359 --> 00:58:07,680
i'm blurred across these states a

1519
00:58:03,680 --> 00:58:09,119
probability distribution across genders

1520
00:58:07,680 --> 00:58:11,119
it's important to note that when a

1521
00:58:09,119 --> 00:58:13,520
classical object interacts with a

1522
00:58:11,119 --> 00:58:16,160
quantum object the wave function

1523
00:58:13,520 --> 00:58:19,520
collapses to a classical value

1524
00:58:16,160 --> 00:58:21,440
like when someone calls me seroman

1525
00:58:19,520 --> 00:58:24,000
so now we're going to do something

1526
00:58:21,440 --> 00:58:27,040
similar i'm going to construct an actual

1527
00:58:24,000 --> 00:58:28,160
quantum state on a real quantum computer

1528
00:58:27,040 --> 00:58:30,400
and then we're going to measure that

1529
00:58:28,160 --> 00:58:32,880
quantum state which will collapse it

1530
00:58:30,400 --> 00:58:34,880
into a classical value

1531
00:58:32,880 --> 00:58:38,160
and to do so i'm going to be using

1532
00:58:34,880 --> 00:58:41,760
kisket which is an open source library

1533
00:58:38,160 --> 00:58:44,880
and ibm's free web ui to visually look

1534
00:58:41,760 --> 00:58:47,280
at what's going on in the code

1535
00:58:44,880 --> 00:58:49,520
so what's happening here is on the right

1536
00:58:47,280 --> 00:58:51,760
we have the kisket code and on the left

1537
00:58:49,520 --> 00:58:52,720
we have a visualization of quantum

1538
00:58:51,760 --> 00:58:54,480
circuit

1539
00:58:52,720 --> 00:58:57,359
and then a visualization of the quantum

1540
00:58:54,480 --> 00:58:59,359
state down the bottom

1541
00:58:57,359 --> 00:59:01,839
this quantum state is

1542
00:58:59,359 --> 00:59:04,000
one cubit which is just an atom that's

1543
00:59:01,839 --> 00:59:05,520
been prepared in a fancy way so that it

1544
00:59:04,000 --> 00:59:09,200
can only be

1545
00:59:05,520 --> 00:59:11,119
in two states on off

1546
00:59:09,200 --> 00:59:13,119
so i said earlier that my gender is like

1547
00:59:11,119 --> 00:59:15,599
a superposition so what i'm going to do

1548
00:59:13,119 --> 00:59:18,400
today is i'm going to put that qubit in

1549
00:59:15,599 --> 00:59:20,480
a 50 50 superposition

1550
00:59:18,400 --> 00:59:22,079
of on and off

1551
00:59:20,480 --> 00:59:24,559
although on and off aren't very good

1552
00:59:22,079 --> 00:59:26,559
genders are they hmm i guess we should

1553
00:59:24,559 --> 00:59:27,839
label them something else

1554
00:59:26,559 --> 00:59:30,319
i guess the traditional choice here

1555
00:59:27,839 --> 00:59:32,400
would be male and female but i'd like to

1556
00:59:30,319 --> 00:59:35,680
go with something a bit spicier

1557
00:59:32,400 --> 00:59:38,160
the true original binary the two genders

1558
00:59:35,680 --> 00:59:40,160
fundamentally opposed from each other

1559
00:59:38,160 --> 00:59:42,160
i'm talking of course about pineapple

1560
00:59:40,160 --> 00:59:43,920
and pizza

1561
00:59:42,160 --> 00:59:46,640
so what we're going to do next is we're

1562
00:59:43,920 --> 00:59:48,640
going to measure our qubit which is in a

1563
00:59:46,640 --> 00:59:49,599
50 50 superposition of pineapple and

1564
00:59:48,640 --> 00:59:51,200
pizza

1565
00:59:49,599 --> 00:59:52,480
and this is just a simulated outcome

1566
00:59:51,200 --> 00:59:54,240
down here

1567
00:59:52,480 --> 00:59:55,520
and then we can run it on a quantum

1568
00:59:54,240 --> 00:59:56,799
computer

1569
00:59:55,520 --> 00:59:59,119
over here

1570
00:59:56,799 --> 01:00:01,119
now unfortunately because this is a free

1571
00:59:59,119 --> 01:00:02,480
and open service there's a lot of other

1572
01:00:01,119 --> 01:00:03,839
people trying to use these quantum

1573
01:00:02,480 --> 01:00:05,440
computers which means you can wait

1574
01:00:03,839 --> 01:00:07,520
several hours for your measurement to be

1575
01:00:05,440 --> 01:00:09,839
run but i thought of this in advance and

1576
01:00:07,520 --> 01:00:11,599
i ran the job earlier this morning

1577
01:00:09,839 --> 01:00:13,599
so we have a result here that i haven't

1578
01:00:11,599 --> 01:00:15,119
looked at yet

1579
01:00:13,599 --> 01:00:16,160
so just to quickly rehash over what

1580
01:00:15,119 --> 01:00:19,040
we've done

1581
01:00:16,160 --> 01:00:21,280
uh i set up a qubit in a 50 50 super

1582
01:00:19,040 --> 01:00:23,440
position of pineapple and pizza

1583
01:00:21,280 --> 01:00:26,240
and then i just measured that qubit to

1584
01:00:23,440 --> 01:00:29,839
see which gender it is

1585
01:00:26,240 --> 01:00:29,839
and now we're going to find out

1586
01:00:32,000 --> 01:00:36,640
this is a very tense moment

1587
01:00:34,400 --> 01:00:40,079
here's our histogram down here

1588
01:00:36,640 --> 01:00:42,400
my gender is pineapple yay

1589
01:00:40,079 --> 01:00:44,319
thanks for coming everyone my name is

1590
01:00:42,400 --> 01:00:47,440
apollo my pronouns are they them and my

1591
01:00:44,319 --> 01:00:47,440
gender is pineapple

1592
01:00:48,319 --> 01:00:50,559
thanks apollo

1593
01:00:49,680 --> 01:00:54,000
uh

1594
01:00:50,559 --> 01:00:57,520
right just waiting for patrick to uh to

1595
01:00:54,000 --> 01:00:59,520
get themselves together ah there we are

1596
01:00:57,520 --> 01:01:01,520
okay yep you had an uncertain amount of

1597
01:00:59,520 --> 01:01:04,079
time for for your talk just then apollo

1598
01:01:01,520 --> 01:01:06,480
while uh while patrick had it had a

1599
01:01:04,079 --> 01:01:10,400
slight uh had a slight moment

1600
01:01:06,480 --> 01:01:14,720
hey up next we have david who is going

1601
01:01:10,400 --> 01:01:14,720
to tell us about gets a sorry git

1602
01:01:15,359 --> 01:01:18,559
hi

1603
01:01:16,319 --> 01:01:20,400
i'm davo here are n git commands you

1604
01:01:18,559 --> 01:01:23,040
didn't know about

1605
01:01:20,400 --> 01:01:24,720
diff 3 merge conflict style this adds a

1606
01:01:23,040 --> 01:01:27,520
section to merge conflicts with a

1607
01:01:24,720 --> 01:01:30,000
content with a common ancestor so you

1608
01:01:27,520 --> 01:01:32,240
can actually see what's changed since

1609
01:01:30,000 --> 01:01:35,440
you last branched here is an example

1610
01:01:32,240 --> 01:01:37,839
from my talon voice coding config

1611
01:01:35,440 --> 01:01:39,680
the default diff algorithm can produce

1612
01:01:37,839 --> 01:01:41,680
unreadable gif sometimes

1613
01:01:39,680 --> 01:01:43,920
but we have others

1614
01:01:41,680 --> 01:01:46,400
the patience algorithm

1615
01:01:43,920 --> 01:01:48,319
tries to keep sequences together much

1616
01:01:46,400 --> 01:01:50,319
like the card game which you might

1617
01:01:48,319 --> 01:01:51,920
otherwise know as solitaire

1618
01:01:50,319 --> 01:01:54,160
but my explanation doesn't really do it

1619
01:01:51,920 --> 01:01:55,760
justice so see james coughlin's blog

1620
01:01:54,160 --> 01:01:59,039
post which i've linked there

1621
01:01:55,760 --> 01:02:00,640
uh it is also sometimes slow

1622
01:01:59,039 --> 01:02:01,760
but there is also the histogram

1623
01:02:00,640 --> 01:02:03,280
algorithm which is based on the

1624
01:02:01,760 --> 01:02:05,119
patient's algorithm and it's been

1625
01:02:03,280 --> 01:02:08,480
benchmarked to be faster than the other

1626
01:02:05,119 --> 01:02:10,160
three diff algorithms in git

1627
01:02:08,480 --> 01:02:12,400
you can color move

1628
01:02:10,160 --> 01:02:14,559
lines that have moved in git diff

1629
01:02:12,400 --> 01:02:16,240
here you can see chunks of lines that

1630
01:02:14,559 --> 01:02:18,960
have moved in different colors in my

1631
01:02:16,240 --> 01:02:18,960
vincon peak

1632
01:02:19,359 --> 01:02:23,839
ever forget what a and b refer to when

1633
01:02:21,680 --> 01:02:26,640
you run git diff well you can change

1634
01:02:23,839 --> 01:02:29,200
those so that they give you context uh

1635
01:02:26,640 --> 01:02:30,720
here we have uh i for index w for

1636
01:02:29,200 --> 01:02:33,119
working tree but you can also have c for

1637
01:02:30,720 --> 01:02:33,119
commit

1638
01:02:33,280 --> 01:02:38,160
has your editor get integration failed

1639
01:02:35,680 --> 01:02:40,160
in a merge conflict try get merge tool

1640
01:02:38,160 --> 01:02:42,079
this opens interactive three-way diff

1641
01:02:40,160 --> 01:02:44,160
tools that you have installed

1642
01:02:42,079 --> 01:02:46,720
uh here are a few examples that i've

1643
01:02:44,160 --> 01:02:48,640
used and you can also

1644
01:02:46,720 --> 01:02:50,559
use a diff tool instead of staring at

1645
01:02:48,640 --> 01:02:54,319
git diff

1646
01:02:50,559 --> 01:02:54,319
to open these same tools

1647
01:02:55,680 --> 01:02:59,760
interactive rebase is great you can chop

1648
01:02:57,680 --> 01:03:02,000
and change your history however you want

1649
01:02:59,760 --> 01:03:04,079
here's an example listing you you get

1650
01:03:02,000 --> 01:03:06,400
when you run rebase i

1651
01:03:04,079 --> 01:03:08,720
however you might not need interactive

1652
01:03:06,400 --> 01:03:11,280
rebase

1653
01:03:08,720 --> 01:03:12,480
git rebase auto squash will find

1654
01:03:11,280 --> 01:03:14,400
formatted

1655
01:03:12,480 --> 01:03:17,039
commit messages generated by the fix up

1656
01:03:14,400 --> 01:03:19,200
and squash options to git commit

1657
01:03:17,039 --> 01:03:22,400
these do the same thing as the actions

1658
01:03:19,200 --> 01:03:26,880
of the same name in rebate i

1659
01:03:22,400 --> 01:03:29,680
and you also have the amend

1660
01:03:26,880 --> 01:03:31,920
option to to fix up as well which is

1661
01:03:29,680 --> 01:03:36,319
basically fix up and reword

1662
01:03:31,920 --> 01:03:36,319
much like git commit dash dash admit

1663
01:03:36,480 --> 01:03:40,960
have you ever bought a rebase don't

1664
01:03:38,720 --> 01:03:42,079
worry git makes it really hard to lose

1665
01:03:40,960 --> 01:03:43,920
commitments

1666
01:03:42,079 --> 01:03:46,319
git appends to a log whenever something

1667
01:03:43,920 --> 01:03:49,280
happens you can view this with a git ref

1668
01:03:46,319 --> 01:03:52,640
log or you can refer to entries in this

1669
01:03:49,280 --> 01:03:54,799
log using syntax uh wherever you can use

1670
01:03:52,640 --> 01:03:56,640
a commit hash

1671
01:03:54,799 --> 01:03:57,520
you can count from the most recent or

1672
01:03:56,640 --> 01:03:59,440
you can

1673
01:03:57,520 --> 01:04:02,000
specify dates as well

1674
01:03:59,440 --> 01:04:03,440
this defaults to head which is uh what

1675
01:04:02,000 --> 01:04:06,000
you've checked out as well if you don't

1676
01:04:03,440 --> 01:04:10,559
specify a a branch name

1677
01:04:06,000 --> 01:04:12,160
uh see dang it git for more oh no tips

1678
01:04:10,559 --> 01:04:15,200
thank you here's a link to my slides

1679
01:04:12,160 --> 01:04:18,000
also canva is hiring

1680
01:04:15,200 --> 01:04:22,160
thanks for that dave oh

1681
01:04:18,000 --> 01:04:23,920
uh up next we have paul who uh only

1682
01:04:22,160 --> 01:04:25,520
received his uh

1683
01:04:23,920 --> 01:04:27,520
his acceptance at the last minute i

1684
01:04:25,520 --> 01:04:29,039
apologize for my email configuration

1685
01:04:27,520 --> 01:04:31,359
being terrible

1686
01:04:29,039 --> 01:04:31,359
uh

1687
01:04:31,520 --> 01:04:34,400
let's uh let's go

1688
01:04:34,480 --> 01:04:39,280
uh hello yeah so hopefully you can see

1689
01:04:36,480 --> 01:04:41,039
that slide set um

1690
01:04:39,280 --> 01:04:42,880
yeah this is a link to an electric

1691
01:04:41,039 --> 01:04:44,559
article it says tesla is collecting

1692
01:04:42,880 --> 01:04:47,039
insane amounts of data from its full

1693
01:04:44,559 --> 01:04:50,000
self-driving test fleet for in fact it's

1694
01:04:47,039 --> 01:04:52,240
also collecting uh video data from its

1695
01:04:50,000 --> 01:04:54,640
normal fleet which numbers about a

1696
01:04:52,240 --> 01:04:56,559
million cars um

1697
01:04:54,640 --> 01:05:00,240
what happens is um

1698
01:04:56,559 --> 01:05:02,319
to um to explain things i um have skin

1699
01:05:00,240 --> 01:05:03,440
in the game i actually own a tesla model

1700
01:05:02,319 --> 01:05:06,480
3

1701
01:05:03,440 --> 01:05:09,359
so i'm following this this basically all

1702
01:05:06,480 --> 01:05:12,400
the time to see exactly what's going on

1703
01:05:09,359 --> 01:05:15,200
but what happens is that uh

1704
01:05:12,400 --> 01:05:17,119
the tesla cars have a certain level of

1705
01:05:15,200 --> 01:05:19,359
automation either called autopilot or

1706
01:05:17,119 --> 01:05:22,000
full-self driving

1707
01:05:19,359 --> 01:05:23,680
and within that

1708
01:05:22,000 --> 01:05:25,280
they have a

1709
01:05:23,680 --> 01:05:26,960
training set up

1710
01:05:25,280 --> 01:05:28,720
with neural networks that they're

1711
01:05:26,960 --> 01:05:29,920
continually training on the next version

1712
01:05:28,720 --> 01:05:32,480
of software

1713
01:05:29,920 --> 01:05:34,640
and whenever they get to a point where

1714
01:05:32,480 --> 01:05:36,880
they are

1715
01:05:34,640 --> 01:05:38,480
have some misbehavior that they're

1716
01:05:36,880 --> 01:05:39,920
trying to cure

1717
01:05:38,480 --> 01:05:41,680
they do something called querying the

1718
01:05:39,920 --> 01:05:42,799
fleet they actually ask the fleet to

1719
01:05:41,680 --> 01:05:45,599
send

1720
01:05:42,799 --> 01:05:48,319
video data back to

1721
01:05:45,599 --> 01:05:51,280
the mothership uh in order for that to

1722
01:05:48,319 --> 01:05:53,760
be incorporated into the next neural net

1723
01:05:51,280 --> 01:05:54,960
yay for neural networks and machine

1724
01:05:53,760 --> 01:05:58,720
learning

1725
01:05:54,960 --> 01:06:01,039
the problem with that is that that data

1726
01:05:58,720 --> 01:06:03,520
could potentially be used

1727
01:06:01,039 --> 01:06:06,799
either to identify people or at least

1728
01:06:03,520 --> 01:06:10,000
identify cars or possibly just probe

1729
01:06:06,799 --> 01:06:13,119
into areas which are restricted

1730
01:06:10,000 --> 01:06:15,359
in other words there's definitely issues

1731
01:06:13,119 --> 01:06:17,760
surrounding privacy

1732
01:06:15,359 --> 01:06:18,559
and every time i've queried this privacy

1733
01:06:17,760 --> 01:06:20,799
issue

1734
01:06:18,559 --> 01:06:23,599
basically on youtube

1735
01:06:20,799 --> 01:06:25,440
i'm i'm staring into the void as to what

1736
01:06:23,599 --> 01:06:26,799
it is that's actually being done about

1737
01:06:25,440 --> 01:06:28,960
it

1738
01:06:26,799 --> 01:06:30,960
so a number of things could happen to

1739
01:06:28,960 --> 01:06:32,799
alleviate it to some extent like for

1740
01:06:30,960 --> 01:06:35,520
example having data centers in different

1741
01:06:32,799 --> 01:06:37,839
countries so you don't ship video data

1742
01:06:35,520 --> 01:06:39,760
off to somewhere else

1743
01:06:37,839 --> 01:06:41,039
but i think the ultimate cure would be

1744
01:06:39,760 --> 01:06:43,760
to have

1745
01:06:41,039 --> 01:06:46,079
the neural network in each car

1746
01:06:43,760 --> 01:06:50,640
be able to respond to queries for video

1747
01:06:46,079 --> 01:06:52,720
data in such a way as it obfuscates

1748
01:06:50,640 --> 01:06:55,359
anything which could be used to identify

1749
01:06:52,720 --> 01:06:58,240
people such as people's faces

1750
01:06:55,359 --> 01:07:00,079
or license plate numbers or any of that

1751
01:06:58,240 --> 01:07:02,480
sort of thing

1752
01:07:00,079 --> 01:07:05,039
but so far i haven't seen

1753
01:07:02,480 --> 01:07:08,079
anybody refer to this as

1754
01:07:05,039 --> 01:07:10,960
as a problem specifically with tesla i

1755
01:07:08,079 --> 01:07:14,079
have seen some research

1756
01:07:10,960 --> 01:07:17,359
in terms of the intersection between

1757
01:07:14,079 --> 01:07:21,440
data privacy and video data for training

1758
01:07:17,359 --> 01:07:22,559
data in a neural network so if anybody

1759
01:07:21,440 --> 01:07:24,880
has

1760
01:07:22,559 --> 01:07:28,079
any leads on who is doing that sort of

1761
01:07:24,880 --> 01:07:31,440
research in australia or or elsewhere

1762
01:07:28,079 --> 01:07:34,000
even please uh please hit me up at paul

1763
01:07:31,440 --> 01:07:37,520
leopardy gmail.com

1764
01:07:34,000 --> 01:07:39,440
and uh that's it thanks very much

1765
01:07:37,520 --> 01:07:40,400
thanks for that paul

1766
01:07:39,440 --> 01:07:42,480
i

1767
01:07:40,400 --> 01:07:44,160
always appreciate hearing about teslas

1768
01:07:42,480 --> 01:07:47,680
from tesla owners

1769
01:07:44,160 --> 01:07:53,039
up next we have peter who is going to be

1770
01:07:47,680 --> 01:07:55,200
uh asking us a very very vital question

1771
01:07:53,039 --> 01:07:57,119
thank you very much chris uh hi everyone

1772
01:07:55,200 --> 01:07:58,640
i'm peter and this

1773
01:07:57,119 --> 01:08:00,960
is my alarm clock

1774
01:07:58,640 --> 01:08:02,799
now unfortunately in order to give this

1775
01:08:00,960 --> 01:08:04,480
uh lightning talk i had to unplug my

1776
01:08:02,799 --> 01:08:06,400
alarm clock and plug it in near my

1777
01:08:04,480 --> 01:08:08,559
computer which means it's lost the time

1778
01:08:06,400 --> 01:08:10,160
so my important question is

1779
01:08:08,559 --> 01:08:12,319
can you please help me to set the time

1780
01:08:10,160 --> 01:08:13,200
on my alarm clock i'm sure that between

1781
01:08:12,319 --> 01:08:15,760
me

1782
01:08:13,200 --> 01:08:16,560
and several hundred of the most uh tech

1783
01:08:15,760 --> 01:08:18,560
uh

1784
01:08:16,560 --> 01:08:20,640
knowledgeable people in the world will

1785
01:08:18,560 --> 01:08:22,239
be able to work this out so i'll give

1786
01:08:20,640 --> 01:08:23,679
you a quick rundown there's seven

1787
01:08:22,239 --> 01:08:25,120
buttons so that should be pretty easy to

1788
01:08:23,679 --> 01:08:28,480
work out what everyone does they are

1789
01:08:25,120 --> 01:08:31,839
labeled off slash fm slash am min slash

1790
01:08:28,480 --> 01:08:34,880
down a l1 slash roll minus snooze slash

1791
01:08:31,839 --> 01:08:36,480
sleep al2 slash vol plus

1792
01:08:34,880 --> 01:08:39,120
hour slash up

1793
01:08:36,480 --> 01:08:41,040
and time slash m plus

1794
01:08:39,120 --> 01:08:42,560
now i'm sure that's going to be really

1795
01:08:41,040 --> 01:08:45,520
easy to uh

1796
01:08:42,560 --> 01:08:47,440
to uh configure so uh whenever i try and

1797
01:08:45,520 --> 01:08:48,880
do something please just like shout out

1798
01:08:47,440 --> 01:08:50,640
what you think the button i should press

1799
01:08:48,880 --> 01:08:51,759
is and if you get it right you can feel

1800
01:08:50,640 --> 01:08:52,480
smart

1801
01:08:51,759 --> 01:08:54,560
so

1802
01:08:52,480 --> 01:08:56,000
let's start by setting the time so uh

1803
01:08:54,560 --> 01:08:57,359
i've done this before on other clocks so

1804
01:08:56,000 --> 01:08:59,199
i know that to set the time you press

1805
01:08:57,359 --> 01:09:01,759
the time button first so let's press the

1806
01:08:59,199 --> 01:09:03,040
time button hey it stopped uh

1807
01:09:01,759 --> 01:09:05,600
it stopped flashing so it's done

1808
01:09:03,040 --> 01:09:07,520
something let's try the hour button or

1809
01:09:05,600 --> 01:09:09,199
the minute button that's going out uh no

1810
01:09:07,520 --> 01:09:11,679
hour minute no

1811
01:09:09,199 --> 01:09:14,000
hour and minute didn't do anything

1812
01:09:11,679 --> 01:09:16,319
i'm going to try the time button again

1813
01:09:14,000 --> 01:09:17,279
let's see what happens time

1814
01:09:16,319 --> 01:09:19,359
nothing

1815
01:09:17,279 --> 01:09:20,640
hour nothing minute

1816
01:09:19,359 --> 01:09:23,199
nothing

1817
01:09:20,640 --> 01:09:25,199
all right if you guessed do a long press

1818
01:09:23,199 --> 01:09:26,960
of the time button

1819
01:09:25,199 --> 01:09:28,560
hey it started flashing now i'm going to

1820
01:09:26,960 --> 01:09:29,839
be able to do something

1821
01:09:28,560 --> 01:09:31,520
all right long pressure that the time

1822
01:09:29,839 --> 01:09:34,880
button was oh it stopped flashing all

1823
01:09:31,520 --> 01:09:38,400
right minute no hour no all right

1824
01:09:34,880 --> 01:09:38,400
long press of the time button

1825
01:09:38,880 --> 01:09:43,040
is it working oh long press hey there we

1826
01:09:41,359 --> 01:09:45,040
go started flashing press minutes

1827
01:09:43,040 --> 01:09:47,040
straight away how australia hey it works

1828
01:09:45,040 --> 01:09:48,239
all right the time is now 114 i don't

1829
01:09:47,040 --> 01:09:50,159
make the rules if you don't believe me

1830
01:09:48,239 --> 01:09:51,920
check my clock

1831
01:09:50,159 --> 01:09:54,960
so now that we have the time set we are

1832
01:09:51,920 --> 01:09:58,000
going to try and tune the radio so i

1833
01:09:54,960 --> 01:09:59,840
have off slash fm slash am so i'm going

1834
01:09:58,000 --> 01:10:01,440
to try and put us into fm let's see what

1835
01:09:59,840 --> 01:10:02,400
happens

1836
01:10:01,440 --> 01:10:04,960
fm

1837
01:10:02,400 --> 01:10:08,960
we have a frequency oh we have static

1838
01:10:04,960 --> 01:10:10,080
let's try up to tune it

1839
01:10:08,960 --> 01:10:11,840
oh

1840
01:10:10,080 --> 01:10:13,520
all right i'm doing a long press to see

1841
01:10:11,840 --> 01:10:15,440
if it'll automatically tune and i don't

1842
01:10:13,520 --> 01:10:16,719
have to try and guess what frequency i

1843
01:10:15,440 --> 01:10:18,400
want

1844
01:10:16,719 --> 01:10:21,440
let's see

1845
01:10:18,400 --> 01:10:23,280
do we have a station

1846
01:10:21,440 --> 01:10:25,280
hey we do all right let's turn it off

1847
01:10:23,280 --> 01:10:27,040
now uh there's an off button i'm

1848
01:10:25,280 --> 01:10:28,960
pressing it

1849
01:10:27,040 --> 01:10:31,520
and it sent me to am

1850
01:10:28,960 --> 01:10:34,080
well it is off slash fm am so let's try

1851
01:10:31,520 --> 01:10:36,960
it again to get to off

1852
01:10:34,080 --> 01:10:39,199
uh took me back to fm

1853
01:10:36,960 --> 01:10:40,560
all right if your guess was do a long

1854
01:10:39,199 --> 01:10:42,480
press of the off button you got it

1855
01:10:40,560 --> 01:10:44,400
correct congratulations

1856
01:10:42,480 --> 01:10:46,880
all right we have the time set we have

1857
01:10:44,400 --> 01:10:48,719
the radio set now we are going to set

1858
01:10:46,880 --> 01:10:51,040
the alarm so i'm going to press the

1859
01:10:48,719 --> 01:10:52,480
alarm one a little light came on i'm

1860
01:10:51,040 --> 01:10:55,760
going to press it again because i want a

1861
01:10:52,480 --> 01:10:57,360
radio alarm and i'm going to set it for

1862
01:10:55,760 --> 01:10:59,280
a one

1863
01:10:57,360 --> 01:11:01,360
minute and

1864
01:10:59,280 --> 01:11:02,640
i think we'll go for i can't remember

1865
01:11:01,360 --> 01:11:04,320
what time it said on my clock i should

1866
01:11:02,640 --> 01:11:07,360
have checked let's go for 11 past and

1867
01:11:04,320 --> 01:11:07,360
see if that looks right

1868
01:11:07,520 --> 01:11:11,280
oh the current time is 15 past all right

1869
01:11:09,360 --> 01:11:12,719
let's change that to 16 pass so that it

1870
01:11:11,280 --> 01:11:15,520
actually goes off while the lightning

1871
01:11:12,719 --> 01:11:15,520
talk is happening

1872
01:11:15,920 --> 01:11:20,080
there we go we have an alarm set for 16

1873
01:11:18,640 --> 01:11:21,520
past

1874
01:11:20,080 --> 01:11:23,440
so hey

1875
01:11:21,520 --> 01:11:25,120
good i've got a minute 30 so that will

1876
01:11:23,440 --> 01:11:26,480
probably go off with enough time to

1877
01:11:25,120 --> 01:11:28,239
spare

1878
01:11:26,480 --> 01:11:29,840
so i got this alarm clock a few years

1879
01:11:28,239 --> 01:11:31,760
back because i really wanted one that

1880
01:11:29,840 --> 01:11:33,040
had red colored numbers on it because

1881
01:11:31,760 --> 01:11:34,880
i've heard that having red lights in

1882
01:11:33,040 --> 01:11:37,040
your bedroom is less disruptive for your

1883
01:11:34,880 --> 01:11:39,040
sleep than having any other color

1884
01:11:37,040 --> 01:11:40,719
and i gotta say it certainly has red

1885
01:11:39,040 --> 01:11:42,000
lights and i really wish i'd just

1886
01:11:40,719 --> 01:11:43,600
actually tried out all the buttons to

1887
01:11:42,000 --> 01:11:45,120
see how to program it first because this

1888
01:11:43,600 --> 01:11:47,840
thing has been confusing

1889
01:11:45,120 --> 01:11:49,360
hey the alarm came off this is great it

1890
01:11:47,840 --> 01:11:51,280
showed me the frequency let's turn it

1891
01:11:49,360 --> 01:11:53,760
off now i remember how to do that it's a

1892
01:11:51,280 --> 01:11:55,920
long press of one

1893
01:11:53,760 --> 01:11:58,880
gonna hold that down and we are going to

1894
01:11:55,920 --> 01:12:00,480
see that the alarm has turned off

1895
01:11:58,880 --> 01:12:02,239
and oh it stayed off all right that's a

1896
01:12:00,480 --> 01:12:03,600
surprise i spent like an hour trying to

1897
01:12:02,239 --> 01:12:05,120
work out how to make this thing not

1898
01:12:03,600 --> 01:12:06,400
always work because it's extremely

1899
01:12:05,120 --> 01:12:08,159
unreliable

1900
01:12:06,400 --> 01:12:09,760
uh sometimes when you do a long press of

1901
01:12:08,159 --> 01:12:12,880
the off button it turns off and then

1902
01:12:09,760 --> 01:12:15,840
turns on again and i get really confused

1903
01:12:12,880 --> 01:12:17,360
but we got there in the end so

1904
01:12:15,840 --> 01:12:19,120
back to me

1905
01:12:17,360 --> 01:12:21,840
thank you very much for helping me set

1906
01:12:19,120 --> 01:12:23,440
my clock i hope you uh had an easy time

1907
01:12:21,840 --> 01:12:26,480
of uh setting the time tuning the radio

1908
01:12:23,440 --> 01:12:29,199
and setting the alarm the lesson here is

1909
01:12:26,480 --> 01:12:31,600
if you are ever uh trying to design a

1910
01:12:29,199 --> 01:12:33,360
user interface for something please do a

1911
01:12:31,600 --> 01:12:35,280
test of your user interface get your

1912
01:12:33,360 --> 01:12:36,800
actual users to try it out see what they

1913
01:12:35,280 --> 01:12:38,560
think and ask them when you have just

1914
01:12:36,800 --> 01:12:40,080
woken up do you want to see the

1915
01:12:38,560 --> 01:12:41,840
frequency of the station you've tuned to

1916
01:12:40,080 --> 01:12:43,760
or do you care about the time i'm not

1917
01:12:41,840 --> 01:12:46,400
sure ask your users

1918
01:12:43,760 --> 01:12:48,560
thank you very much chris and lily

1919
01:12:46,400 --> 01:12:50,800
thanks for that peter i

1920
01:12:48,560 --> 01:12:52,880
i did end up yanking my earpiece out of

1921
01:12:50,800 --> 01:12:53,679
my ear when that am static came on it

1922
01:12:52,880 --> 01:12:55,600
was

1923
01:12:53,679 --> 01:12:58,320
the most painful lightning talk i've

1924
01:12:55,600 --> 01:13:00,560
experienced at this point but uh nothing

1925
01:12:58,320 --> 01:13:02,080
like the pain of actually trying to set

1926
01:13:00,560 --> 01:13:04,880
that thing uh

1927
01:13:02,080 --> 01:13:06,719
yeah i'm glad that there was a very good

1928
01:13:04,880 --> 01:13:08,480
lesson at the end of this because

1929
01:13:06,719 --> 01:13:10,080
otherwise it was just going to be five

1930
01:13:08,480 --> 01:13:12,960
minutes of let me show you this

1931
01:13:10,080 --> 01:13:15,520
incredibly cursed clock which i mean is

1932
01:13:12,960 --> 01:13:16,800
also entertaining in and of its own

1933
01:13:15,520 --> 01:13:20,640
right so

1934
01:13:16,800 --> 01:13:23,679
anyway next up we have caitlin mcleod

1935
01:13:20,640 --> 01:13:25,520
who is going to teach us some lessons

1936
01:13:23,679 --> 01:13:27,199
about speed running

1937
01:13:25,520 --> 01:13:29,040
when they were working on this talk we

1938
01:13:27,199 --> 01:13:31,760
were chatting in the discord and they

1939
01:13:29,040 --> 01:13:33,920
mentioned to me wait libby

1940
01:13:31,760 --> 01:13:36,159
i can't swear in these lightning talks

1941
01:13:33,920 --> 01:13:37,440
to which i responded uh

1942
01:13:36,159 --> 01:13:40,000
and they said

1943
01:13:37,440 --> 01:13:42,080
libby this is suddenly very difficult i

1944
01:13:40,000 --> 01:13:43,360
did say the f word when i said that time

1945
01:13:42,080 --> 01:13:44,880
starts

1946
01:13:43,360 --> 01:13:47,040
all right time starts hello i'm kate

1947
01:13:44,880 --> 01:13:48,320
kate latte oh caitlin uh i use shio they

1948
01:13:47,040 --> 01:13:50,800
pronouns both are good today we're

1949
01:13:48,320 --> 01:13:52,480
having speedrun lessons hello uh if you

1950
01:13:50,800 --> 01:13:54,400
haven't played celeste before this is a

1951
01:13:52,480 --> 01:13:56,239
game about an anxious trans girl called

1952
01:13:54,400 --> 01:13:58,880
madeline who finds an ancient abandoned

1953
01:13:56,239 --> 01:14:00,159
cursed mountain and uh wins

1954
01:13:58,880 --> 01:14:01,280
uh so some of the lessons that i've

1955
01:14:00,159 --> 01:14:03,520
learned while speed running this game

1956
01:14:01,280 --> 01:14:05,040
for the past like five months uh are as

1957
01:14:03,520 --> 01:14:06,719
follows

1958
01:14:05,040 --> 01:14:09,280
heck okay

1959
01:14:06,719 --> 01:14:09,280
great start

1960
01:14:09,440 --> 01:14:12,800
oh no

1961
01:14:11,280 --> 01:14:14,400
oh no

1962
01:14:12,800 --> 01:14:16,080
uh the first one that i want to really

1963
01:14:14,400 --> 01:14:18,239
talk about is to be honest about what

1964
01:14:16,080 --> 01:14:20,640
drives you so um myself i'm a very

1965
01:14:18,239 --> 01:14:21,840
people and community focused person

1966
01:14:20,640 --> 01:14:23,280
uh but there are a lot of people on

1967
01:14:21,840 --> 01:14:24,800
speedrunning who are very competitive

1968
01:14:23,280 --> 01:14:26,400
and they just wanna like get better at

1969
01:14:24,800 --> 01:14:27,360
the game they wanna show off how cool

1970
01:14:26,400 --> 01:14:29,440
they are

1971
01:14:27,360 --> 01:14:31,199
and they want to beat their friends uh

1972
01:14:29,440 --> 01:14:33,440
that's not always sustainable

1973
01:14:31,199 --> 01:14:34,400
um there's a lot of people who burn out

1974
01:14:33,440 --> 01:14:36,239
because

1975
01:14:34,400 --> 01:14:38,159
they just can't beat the people who are

1976
01:14:36,239 --> 01:14:39,840
very good at the game

1977
01:14:38,159 --> 01:14:41,120
um but there's also a lot of people

1978
01:14:39,840 --> 01:14:42,960
who've stayed in there for a long time

1979
01:14:41,120 --> 01:14:45,920
so that's pretty nice

1980
01:14:42,960 --> 01:14:45,920
um the oh

1981
01:14:46,719 --> 01:14:50,239
i'm just gonna get past the screen there

1982
01:14:48,400 --> 01:14:52,159
we go

1983
01:14:50,239 --> 01:14:54,640
the other big one uh that i really love

1984
01:14:52,159 --> 01:14:56,960
about celeste is and about python as

1985
01:14:54,640 --> 01:14:58,560
well is how welcoming communities really

1986
01:14:56,960 --> 01:15:00,560
bring in fresh blood

1987
01:14:58,560 --> 01:15:02,800
um

1988
01:15:00,560 --> 01:15:04,320
you don't really think about it uh

1989
01:15:02,800 --> 01:15:06,560
when you you're involved in these

1990
01:15:04,320 --> 01:15:08,320
communities for so long but compared to

1991
01:15:06,560 --> 01:15:09,520
other speed running communities and

1992
01:15:08,320 --> 01:15:11,440
especially

1993
01:15:09,520 --> 01:15:13,840
um celeste is really friendly when i

1994
01:15:11,440 --> 01:15:14,960
started playing on twitch um

1995
01:15:13,840 --> 01:15:18,560
and when other people started

1996
01:15:14,960 --> 01:15:20,560
speedrunning on twitch uh before i did

1997
01:15:18,560 --> 01:15:22,320
uh speedrunners just hang out in new

1998
01:15:20,560 --> 01:15:24,560
speedrunners chats and tell them like

1999
01:15:22,320 --> 01:15:27,040
cool strats and it's really cute

2000
01:15:24,560 --> 01:15:28,719
um and python we've got a lot of good oh

2001
01:15:27,040 --> 01:15:30,560
i'm i'm quite nervous i'm messing up a

2002
01:15:28,719 --> 01:15:32,480
lot

2003
01:15:30,560 --> 01:15:34,320
we've got a lot of good uh ways to get

2004
01:15:32,480 --> 01:15:35,920
new people on to the community and the

2005
01:15:34,320 --> 01:15:37,280
best thing about that is that even

2006
01:15:35,920 --> 01:15:39,280
though people are burning out and like

2007
01:15:37,280 --> 01:15:41,360
moving out into other areas they've got

2008
01:15:39,280 --> 01:15:43,280
fresh people coming in and playing with

2009
01:15:41,360 --> 01:15:45,040
new strats and coming up with new ideas

2010
01:15:43,280 --> 01:15:46,320
about how to save time in the game and

2011
01:15:45,040 --> 01:15:47,840
that's really good

2012
01:15:46,320 --> 01:15:49,280
it's really healthy to have that kind of

2013
01:15:47,840 --> 01:15:50,719
community and i'm really glad that we've

2014
01:15:49,280 --> 01:15:53,600
got that

2015
01:15:50,719 --> 01:15:55,360
the next lesson that we have is

2016
01:15:53,600 --> 01:15:57,120
uh practice to practice and practice

2017
01:15:55,360 --> 01:15:58,560
some more oh my cat really wants to

2018
01:15:57,120 --> 01:16:00,960
leave the room okay

2019
01:15:58,560 --> 01:16:00,960
you can't

2020
01:16:01,440 --> 01:16:04,719
uh

2021
01:16:03,040 --> 01:16:06,320
and celeste like

2022
01:16:04,719 --> 01:16:07,520
a lot of this is muscle memory right you

2023
01:16:06,320 --> 01:16:08,880
really need to get it into your muscle

2024
01:16:07,520 --> 01:16:10,159
memory you really need to learn the

2025
01:16:08,880 --> 01:16:12,400
strats

2026
01:16:10,159 --> 01:16:13,120
and learn to get better at the game

2027
01:16:12,400 --> 01:16:14,880
and

2028
01:16:13,120 --> 01:16:16,400
if you're

2029
01:16:14,880 --> 01:16:18,159
i'm sorry if my camera goes out he's

2030
01:16:16,400 --> 01:16:21,840
just nudged the cable

2031
01:16:18,159 --> 01:16:21,840
oh really mickey

2032
01:16:21,920 --> 01:16:26,320
if you are uh

2033
01:16:24,320 --> 01:16:28,080
burnt out from practicing the same level

2034
01:16:26,320 --> 01:16:31,199
again and again and again and or you're

2035
01:16:28,080 --> 01:16:32,480
burnt out from only having one focus uh

2036
01:16:31,199 --> 01:16:34,800
at work

2037
01:16:32,480 --> 01:16:37,199
uh and in your uh programming things

2038
01:16:34,800 --> 01:16:39,520
it's good to try and practice something

2039
01:16:37,199 --> 01:16:40,719
else uh there are other categories

2040
01:16:39,520 --> 01:16:42,400
there are other tools you could give

2041
01:16:40,719 --> 01:16:43,920
machine learning a try for example in

2042
01:16:42,400 --> 01:16:45,199
python they're all options that you've

2043
01:16:43,920 --> 01:16:47,600
got

2044
01:16:45,199 --> 01:16:49,679
um so just yeah have it have an eye on

2045
01:16:47,600 --> 01:16:51,600
it practice is really good and it's

2046
01:16:49,679 --> 01:16:53,280
really good to be able to spend time on

2047
01:16:51,600 --> 01:16:55,920
the things that you love

2048
01:16:53,280 --> 01:16:55,920
but also

2049
01:16:56,800 --> 01:17:00,719
there we go

2050
01:16:58,800 --> 01:17:03,120
uh you really want to keep an eye on

2051
01:17:00,719 --> 01:17:04,719
your uh like

2052
01:17:03,120 --> 01:17:06,719
how you're feeling about the thing if

2053
01:17:04,719 --> 01:17:08,000
it's feeling like a drag

2054
01:17:06,719 --> 01:17:09,280
then you're not going to be as effective

2055
01:17:08,000 --> 01:17:10,800
at it

2056
01:17:09,280 --> 01:17:12,719
and obviously celeste is a video game

2057
01:17:10,800 --> 01:17:14,800
right like it's designed to be fun

2058
01:17:12,719 --> 01:17:15,760
uh instead of a troll or something that

2059
01:17:14,800 --> 01:17:18,320
you missed

2060
01:17:15,760 --> 01:17:18,320
heck me

2061
01:17:19,440 --> 01:17:22,880
thank you okay

2062
01:17:21,520 --> 01:17:25,120
it's something that you necessarily do

2063
01:17:22,880 --> 01:17:26,960
for money um

2064
01:17:25,120 --> 01:17:29,120
but that can be it another thing that

2065
01:17:26,960 --> 01:17:30,960
i've really loved about celeste is how

2066
01:17:29,120 --> 01:17:32,960
generous the community is

2067
01:17:30,960 --> 01:17:34,800
not just with their time but also with

2068
01:17:32,960 --> 01:17:37,600
the money there's a lot of speed running

2069
01:17:34,800 --> 01:17:39,760
marathons that raise a lot of cash for

2070
01:17:37,600 --> 01:17:42,239
important charities like trans lifeline

2071
01:17:39,760 --> 01:17:44,719
transliberation one minute

2072
01:17:42,239 --> 01:17:46,640
one minute oh that's fine we might

2073
01:17:44,719 --> 01:17:48,560
finish the level we might not i did mess

2074
01:17:46,640 --> 01:17:50,800
up a bit

2075
01:17:48,560 --> 01:17:50,800
oh

2076
01:17:51,520 --> 01:17:56,280
i'm just gonna focus for this screen oh

2077
01:17:53,280 --> 01:17:56,280
no

2078
01:17:59,199 --> 01:18:03,360
um but also collaboration has been

2079
01:18:00,960 --> 01:18:05,120
really important in this community

2080
01:18:03,360 --> 01:18:06,960
i was part of a celeste speedrunning

2081
01:18:05,120 --> 01:18:09,600
tournament um which had really

2082
01:18:06,960 --> 01:18:12,000
impressive technical capabilities for

2083
01:18:09,600 --> 01:18:14,159
restreaming so many different people and

2084
01:18:12,000 --> 01:18:15,840
recording their runs and their races

2085
01:18:14,159 --> 01:18:16,719
we're gonna finish

2086
01:18:15,840 --> 01:18:18,880
um

2087
01:18:16,719 --> 01:18:20,800
the most important uh speedrunning tip

2088
01:18:18,880 --> 01:18:22,560
that i've got for you all however is to

2089
01:18:20,800 --> 01:18:23,840
just fast bubble every single bubble so

2090
01:18:22,560 --> 01:18:26,400
you don't lose your speed thank you very

2091
01:18:23,840 --> 01:18:27,920
much that was a 407 within 6 seconds of

2092
01:18:26,400 --> 01:18:29,040
my pv that's pretty good

2093
01:18:27,920 --> 01:18:30,880
and i just want to get to this part of

2094
01:18:29,040 --> 01:18:32,239
the cutscene

2095
01:18:30,880 --> 01:18:35,040
and this is theater falling down from a

2096
01:18:32,239 --> 01:18:35,040
cliff nine

2097
01:18:36,239 --> 01:18:41,280
we nailed it thank you very much

2098
01:18:38,960 --> 01:18:43,760
okay thank you for that caitlyn

2099
01:18:41,280 --> 01:18:46,960
hey it's time for our last lightning

2100
01:18:43,760 --> 01:18:50,000
talk uh after caitlyn interrupted our

2101
01:18:46,960 --> 01:18:52,320
stream of peter and paul we've now got

2102
01:18:50,000 --> 01:18:53,360
marius who's going to be telling us

2103
01:18:52,320 --> 01:18:55,760
about

2104
01:18:53,360 --> 01:18:57,440
something extremely important with

2105
01:18:55,760 --> 01:19:00,080
tensorflow

2106
01:18:57,440 --> 01:19:02,800
yeah hi um i'm coming to you from diary

2107
01:19:00,080 --> 01:19:04,640
country in southwest sydney today

2108
01:19:02,800 --> 01:19:06,640
um yeah and i was just going to talk

2109
01:19:04,640 --> 01:19:08,640
people through my latest you know

2110
01:19:06,640 --> 01:19:10,239
weekend projects my weekend projects

2111
01:19:08,640 --> 01:19:11,679
never take a weekend they're always a

2112
01:19:10,239 --> 01:19:14,000
bit longer

2113
01:19:11,679 --> 01:19:16,000
but basically this was uh the idea of

2114
01:19:14,000 --> 01:19:17,679
generating metal band names using

2115
01:19:16,000 --> 01:19:20,080
machine learning

2116
01:19:17,679 --> 01:19:22,719
so using python and tensorflow

2117
01:19:20,080 --> 01:19:24,960
um yeah so i'm a data scientist at work

2118
01:19:22,719 --> 01:19:26,800
and a metal fan at home

2119
01:19:24,960 --> 01:19:28,800
um and i thought metal band names were

2120
01:19:26,800 --> 01:19:31,040
kind of a fun thing to play around with

2121
01:19:28,800 --> 01:19:33,199
with some kind of data science tools

2122
01:19:31,040 --> 01:19:35,120
because if you're a fan of the genre you

2123
01:19:33,199 --> 01:19:37,679
sort of know that there's kind of some

2124
01:19:35,120 --> 01:19:39,679
correlation between band names and all

2125
01:19:37,679 --> 01:19:42,640
these hyper-specific sub-genres that

2126
01:19:39,679 --> 01:19:45,280
people use like atmospheric black metal

2127
01:19:42,640 --> 01:19:46,880
um but it's not really easy to pin down

2128
01:19:45,280 --> 01:19:49,360
so it's sort of fun to play around in

2129
01:19:46,880 --> 01:19:51,679
that space and see if you can kind of

2130
01:19:49,360 --> 01:19:52,960
approximate that kind of fuzzy feeling

2131
01:19:51,679 --> 01:19:55,600
of how some

2132
01:19:52,960 --> 01:19:56,400
band names are sort of suited to some

2133
01:19:55,600 --> 01:19:58,880
um

2134
01:19:56,400 --> 01:20:01,520
certain genres so i think you know blood

2135
01:19:58,880 --> 01:20:04,719
incantation great death metal band name

2136
01:20:01,520 --> 01:20:06,320
um they love these four syllable words

2137
01:20:04,719 --> 01:20:08,080
but yeah some of the other things are a

2138
01:20:06,320 --> 01:20:10,639
little bit harder to pin down you can't

2139
01:20:08,080 --> 01:20:12,719
necessarily spell out rules for why

2140
01:20:10,639 --> 01:20:14,880
bands are called certain things you just

2141
01:20:12,719 --> 01:20:17,280
sort of know it when you see them

2142
01:20:14,880 --> 01:20:18,800
um yeah so it's kind of a complex

2143
01:20:17,280 --> 01:20:21,199
influence space with some kind of

2144
01:20:18,800 --> 01:20:23,120
structure of correlation to it

2145
01:20:21,199 --> 01:20:24,960
there's lots of labeled examples to it

2146
01:20:23,120 --> 01:20:26,000
because there's obsessive people on the

2147
01:20:24,960 --> 01:20:27,840
internet

2148
01:20:26,000 --> 01:20:29,840
cataloging all these bands and listing

2149
01:20:27,840 --> 01:20:31,920
all their genres

2150
01:20:29,840 --> 01:20:34,159
you can check that the generative model

2151
01:20:31,920 --> 01:20:38,080
you're using is working based on kind of

2152
01:20:34,159 --> 01:20:40,639
the vibe and the wrong answers are funny

2153
01:20:38,080 --> 01:20:42,400
yeah and obviously even if your machine

2154
01:20:40,639 --> 01:20:44,639
learning model goes really wrong you're

2155
01:20:42,400 --> 01:20:46,960
probably not going to come up with this

2156
01:20:44,639 --> 01:20:48,960
they are a genuinely good band but the

2157
01:20:46,960 --> 01:20:50,719
name is kind of intentionally

2158
01:20:48,960 --> 01:20:51,920
unpronounceable

2159
01:20:50,719 --> 01:20:54,800
um

2160
01:20:51,920 --> 01:20:56,800
yeah so for my weekend projects in order

2161
01:20:54,800 --> 01:20:58,560
to actually get them done i try to keep

2162
01:20:56,800 --> 01:21:00,480
them simple and just

2163
01:20:58,560 --> 01:21:02,159
you know use what's already out there

2164
01:21:00,480 --> 01:21:03,199
rather than trying to do everything from

2165
01:21:02,159 --> 01:21:05,760
scratch

2166
01:21:03,199 --> 01:21:08,880
um so in this project you know there's

2167
01:21:05,760 --> 01:21:11,440
scraping the encyclopedia metallic

2168
01:21:08,880 --> 01:21:13,199
you can use scrape by or requests or

2169
01:21:11,440 --> 01:21:15,760
whatever else i think what ended up

2170
01:21:13,199 --> 01:21:17,840
working was just doing requests

2171
01:21:15,760 --> 01:21:20,719
it's really not a dev friendly site it

2172
01:21:17,840 --> 01:21:22,080
has no particularly nice api i sort of

2173
01:21:20,719 --> 01:21:24,880
had to look at scripts people had

2174
01:21:22,080 --> 01:21:26,800
already written and modify them a bit

2175
01:21:24,880 --> 01:21:29,600
and it was a bit of a nightmare

2176
01:21:26,800 --> 01:21:31,600
um and then yeah you've got numpy pandas

2177
01:21:29,600 --> 01:21:34,320
jupiter lab your kind of standard data

2178
01:21:31,600 --> 01:21:35,920
science tr um stack for

2179
01:21:34,320 --> 01:21:37,920
going through the initial data and kind

2180
01:21:35,920 --> 01:21:40,159
of figuring out what's there

2181
01:21:37,920 --> 01:21:41,840
um and then what really helped out this

2182
01:21:40,159 --> 01:21:43,760
off the ground is

2183
01:21:41,840 --> 01:21:46,320
someone's written this python package

2184
01:21:43,760 --> 01:21:48,800
called textgen rnn

2185
01:21:46,320 --> 01:21:50,639
and what it has is it has pre-trained

2186
01:21:48,800 --> 01:21:52,560
text generation models that have already

2187
01:21:50,639 --> 01:21:54,880
been trained on

2188
01:21:52,560 --> 01:21:57,520
huge batches of text so they sort of

2189
01:21:54,880 --> 01:21:59,760
already know how text works

2190
01:21:57,520 --> 01:22:01,520
yeah and training ml models from scratch

2191
01:21:59,760 --> 01:22:02,800
is not always fun it's not necessarily

2192
01:22:01,520 --> 01:22:04,239
something you want to be doing on your

2193
01:22:02,800 --> 01:22:05,840
weekends

2194
01:22:04,239 --> 01:22:06,960
and this gives you a really good base to

2195
01:22:05,840 --> 01:22:07,920
start from

2196
01:22:06,960 --> 01:22:10,159
um

2197
01:22:07,920 --> 01:22:11,920
and then the other thing that helped was

2198
01:22:10,159 --> 01:22:13,679
i wanted to put this up on the web so

2199
01:22:11,920 --> 01:22:15,840
people i could actually see it

2200
01:22:13,679 --> 01:22:18,560
and it turns out tensorflow models are

2201
01:22:15,840 --> 01:22:20,560
pretty easy to transfer from python um

2202
01:22:18,560 --> 01:22:23,199
once you've traded them up

2203
01:22:20,560 --> 01:22:25,440
and import them into tensorflow.js where

2204
01:22:23,199 --> 01:22:26,960
they can run in the browser

2205
01:22:25,440 --> 01:22:28,800
um

2206
01:22:26,960 --> 01:22:31,120
so what i actually ended up with for a

2207
01:22:28,800 --> 01:22:32,639
machine learning model that generates

2208
01:22:31,120 --> 01:22:34,880
metal band names

2209
01:22:32,639 --> 01:22:38,159
um it's mostly just using

2210
01:22:34,880 --> 01:22:40,400
texture and rnn's model architecture so

2211
01:22:38,159 --> 01:22:42,239
it's a recurrent neural network

2212
01:22:40,400 --> 01:22:43,440
using long short term memory for the

2213
01:22:42,239 --> 01:22:46,239
band names

2214
01:22:43,440 --> 01:22:49,120
um and then in order to kind of add in

2215
01:22:46,239 --> 01:22:51,280
this extra wrinkle of genre information

2216
01:22:49,120 --> 01:22:53,600
um we're feeding that in as kind of a

2217
01:22:51,280 --> 01:22:55,600
vector of binary values for each band

2218
01:22:53,600 --> 01:22:57,120
each band can have multiple different

2219
01:22:55,600 --> 01:22:58,639
genres

2220
01:22:57,120 --> 01:23:00,880
and that required some slight

2221
01:22:58,639 --> 01:23:03,120
modification of the text gen model

2222
01:23:00,880 --> 01:23:05,120
but that's sort of really one of what i

2223
01:23:03,120 --> 01:23:06,800
really wanted to focus on

2224
01:23:05,120 --> 01:23:09,679
so it's trained on all the bands that

2225
01:23:06,800 --> 01:23:11,520
were available in the encyclopedia motel

2226
01:23:09,679 --> 01:23:13,679
and it only needed five training epochs

2227
01:23:11,520 --> 01:23:15,840
to get good results

2228
01:23:13,679 --> 01:23:17,360
it's up on the web that was relatively

2229
01:23:15,840 --> 01:23:19,040
easy to do although i'm not great at

2230
01:23:17,360 --> 01:23:21,920
javascript

2231
01:23:19,040 --> 01:23:23,440
and what you end up with is some

2232
01:23:21,920 --> 01:23:25,679
code that generates potentially

2233
01:23:23,440 --> 01:23:27,040
offensive names so i'm not going to run

2234
01:23:25,679 --> 01:23:28,639
it live

2235
01:23:27,040 --> 01:23:30,480
but you should try it out for yourself

2236
01:23:28,639 --> 01:23:31,760
and it comes up with all these fun names

2237
01:23:30,480 --> 01:23:32,800
like this

2238
01:23:31,760 --> 01:23:35,199
um

2239
01:23:32,800 --> 01:23:37,360
yeah so that is playing around with

2240
01:23:35,199 --> 01:23:39,360
generative ml models at home and

2241
01:23:37,360 --> 01:23:40,800
something i'd encourage people to check

2242
01:23:39,360 --> 01:23:44,800
out

2243
01:23:40,800 --> 01:23:48,159
uh and thank you oh thank you marius

2244
01:23:44,800 --> 01:23:50,480
okay well that is all of the lightning

2245
01:23:48,159 --> 01:23:53,840
talks that we have this year

2246
01:23:50,480 --> 01:23:57,120
uh i'm as sad as all of you are as well

2247
01:23:53,840 --> 01:23:58,320
oh hi bobo hi i all know what you're

2248
01:23:57,120 --> 01:24:01,360
here for

2249
01:23:58,320 --> 01:24:03,440
hey hey hey hey hey hey what's this

2250
01:24:01,360 --> 01:24:06,159
what's this this over here over here

2251
01:24:03,440 --> 01:24:07,520
over here over here oh my god

2252
01:24:06,159 --> 01:24:10,159
you're an adult it's right that you were

2253
01:24:07,520 --> 01:24:10,159
looking at it

2254
01:24:10,480 --> 01:24:14,560
it's everyone's favorite treat ignoring

2255
01:24:12,639 --> 01:24:16,400
horse

2256
01:24:14,560 --> 01:24:18,320
right uh so yeah that brings us to the

2257
01:24:16,400 --> 01:24:22,159
end of the lightning talks for this year

2258
01:24:18,320 --> 01:24:24,960
uh i'm sad uh i'm sure you're all sad as

2259
01:24:22,159 --> 01:24:26,960
well but uh you've got a whole year

2260
01:24:24,960 --> 01:24:28,800
hopefully not spent in lockdown to come

2261
01:24:26,960 --> 01:24:31,040
up with your next idea for lightning

2262
01:24:28,800 --> 01:24:32,000
talks which you can present at next

2263
01:24:31,040 --> 01:24:34,159
year's

2264
01:24:32,000 --> 01:24:36,400
pycon au

2265
01:24:34,159 --> 01:24:39,120
wherever that ends up being

2266
01:24:36,400 --> 01:24:40,159
either where you are right now or online

2267
01:24:39,120 --> 01:24:43,199
uh

2268
01:24:40,159 --> 01:24:45,360
as usual i would love to thank the great

2269
01:24:43,199 --> 01:24:47,679
av crew who make this thing work

2270
01:24:45,360 --> 01:24:49,360
including ryan who has been backstage

2271
01:24:47,679 --> 01:24:51,600
for the entire thing and won't let me

2272
01:24:49,360 --> 01:24:54,000
bring him on stage

2273
01:24:51,600 --> 01:24:55,520
jack who has been doing all the tech

2274
01:24:54,000 --> 01:24:59,360
checks to make sure that people actually

2275
01:24:55,520 --> 01:25:01,920
join stream yard without things breaking

2276
01:24:59,360 --> 01:25:03,760
and of course patrick the disembodied

2277
01:25:01,920 --> 01:25:06,080
clock

2278
01:25:03,760 --> 01:25:08,320
thank you disembodied clock

2279
01:25:06,080 --> 01:25:10,800
and of course thanks to libby berry who

2280
01:25:08,320 --> 01:25:12,719
has been wrangling all of the speakers

2281
01:25:10,800 --> 01:25:13,760
as well as bobo for today's lightning

2282
01:25:12,719 --> 01:25:16,320
talks

2283
01:25:13,760 --> 01:25:16,320
he helps

2284
01:25:16,560 --> 01:25:20,159
and is that what's going on right now oh

2285
01:25:18,639 --> 01:25:22,800
yeah

2286
01:25:20,159 --> 01:25:24,719
that is definitely helping okay so yeah

2287
01:25:22,800 --> 01:25:27,520
thank you libby and thank you to all of

2288
01:25:24,719 --> 01:25:30,639
our presenters uh i love these lightning

2289
01:25:27,520 --> 01:25:31,920
talks uh at every pike on eu and i'm

2290
01:25:30,639 --> 01:25:34,880
really looking forward to seeing what

2291
01:25:31,920 --> 01:25:36,719
you all have next year um

2292
01:25:34,880 --> 01:25:38,719
see you then and maybe see you in mario

2293
01:25:36,719 --> 01:25:40,400
kart in a bit if uh if you're playing

2294
01:25:38,719 --> 01:25:44,480
hey byron heck yeah

2295
01:25:40,400 --> 01:25:44,480
all right take care everyone stay safe

2296
01:25:46,239 --> 01:25:50,400
i presume how long are we waving for

2297
01:25:49,280 --> 01:25:52,840
are we off

2298
01:25:50,400 --> 01:25:54,960
i have no idea tell us

2299
01:25:52,840 --> 01:25:57,280
about hey bobo do you want to attack the

2300
01:25:54,960 --> 01:26:01,159
dead cat again that was a lot of fun do

2301
01:25:57,280 --> 01:26:01,159
it great thanks patrick