WEBVTT
1
00:00:01.169 --> 00:00:04.790
How'd you like to listen to. NET Rocks with no ads? Easy.
2
00:00:05.370 --> 00:00:08.779
Become a patron. For just $ 5 a month, you get
3
00:00:08.890 --> 00:00:11.720
access to a private RSS feed where all the shows
4
00:00:11.800 --> 00:00:14.900
have no ads. $ 20 a month will get you that
5
00:00:15.119 --> 00:00:18.839
and a special. NET Rocks patron mug. Sign up now
6
00:00:18.899 --> 00:00:37.659
at patreon.dotnetrocks.com. Hey, welcome back to. NET Rocks. I'm Carl Franklin.
7
00:00:37.840 --> 00:00:42.229
And I'm Richard Campbell. This is episode 2022. It was
8
00:00:42.229 --> 00:00:45.630
22 episodes ago we were in that bar at the
9
00:00:45.950 --> 00:00:46.750
MVP Summit.
10
00:00:46.850 --> 00:00:47.590
Yeah.
11
00:00:47.890 --> 00:00:50.869
At Party with Palermo doing episode 2000. Yeah.
12
00:00:51.009 --> 00:00:55.299
Yeah. And now we're four away from the end of history. Yeah.
13
00:00:58.369 --> 00:01:02.200
It was only four years ago, 2022. So let's recap
14
00:01:02.219 --> 00:01:03.729
some of the things that happened that year.
15
00:01:03.759 --> 00:01:04.450
It's pretty current.
16
00:01:04.730 --> 00:01:07.359
A little thing called Russia invading Ukraine.
17
00:01:07.560 --> 00:01:08.799
Yeah. That's a little thing.
18
00:01:08.859 --> 00:01:10.140
Maybe you've heard of that little event.
19
00:01:10.239 --> 00:01:12.299
Remember when it was supposed to be three days, so
20
00:01:12.340 --> 00:01:14.359
they didn't bother packing a lot of food and said
21
00:01:14.379 --> 00:01:18.819
they packed their dress uniforms for the march down the boulevard?
22
00:01:18.879 --> 00:01:19.400
Yeah.
23
00:01:19.620 --> 00:01:20.079
How'd that go?
24
00:01:20.219 --> 00:01:20.379
Yeah.
25
00:01:20.640 --> 00:01:23.599
In other happy news, the Supreme Court overturned Roe v.
26
00:01:23.640 --> 00:01:24.560
Wade. Oh, geez.
27
00:01:24.760 --> 00:01:29.409
Queen Elizabeth II died. After 70 years on the throne.
28
00:01:29.849 --> 00:01:33.909
Second longest reigning monarch in history, something like that. Yeah. Yeah.
29
00:01:34.109 --> 00:01:36.640
She, from what I understand, she had quite a sense
30
00:01:36.659 --> 00:01:37.040
of humor.
31
00:01:37.280 --> 00:01:37.739
Very funny.
32
00:01:38.000 --> 00:01:38.239
Yeah.
33
00:01:38.340 --> 00:01:42.599
Very funny. Classically, like there's stories like Balmoral Castle is
34
00:01:42.659 --> 00:01:45.239
up in Scotland and she loved to take a walk
35
00:01:45.299 --> 00:01:47.200
and she'd never dressed up or anything. And she usually
36
00:01:47.219 --> 00:01:49.780
said a guard with her. So it's like two elderly
37
00:01:49.819 --> 00:01:53.019
people walking along the field and they bump into some
38
00:01:53.120 --> 00:01:58.319
tourists and they ask him, you know, I've heard the
39
00:01:58.599 --> 00:02:01.030
Windsor's Castle is around here somewhere. Oh, yeah, it's around
40
00:02:01.069 --> 00:02:03.209
here somewhere. Is it that one over there?
41
00:02:03.250 --> 00:02:04.269
He goes, I don't know.
42
00:02:04.329 --> 00:02:06.650
The little old lady there is a bit strange. They go, oh,
43
00:02:06.969 --> 00:02:08.750
thanks for your help. You have a good time. And
44
00:02:08.770 --> 00:02:11.650
they walk away. Hilarious.
45
00:02:11.770 --> 00:02:17.860
Yep, she was. Inflation, cost of living, crisis, not good.
46
00:02:18.020 --> 00:02:20.139
Well, we're still coming out of COVID, right?
47
00:02:20.180 --> 00:02:24.500
Yeah, we're coming out of COVID. That's right. Finally began retreating.
48
00:02:25.319 --> 00:02:29.659
as a global crisis in most countries, except most countries,
49
00:02:29.699 --> 00:02:30.419
except for China.
50
00:02:30.469 --> 00:02:30.550
Yeah.
51
00:02:30.560 --> 00:02:31.590
China stayed locked down.
52
00:02:31.669 --> 00:02:34.509
I don't know if that was a good call, but yeah. Um,
53
00:02:34.990 --> 00:02:41.050
Uvalde school shooting, 19 children. Uh, terrible, uh, historic us
54
00:02:41.110 --> 00:02:41.969
midterm elections.
55
00:02:42.810 --> 00:02:43.000
Yeah.
56
00:02:43.219 --> 00:02:49.599
Iran erupts in protest after the death of Masha Amini. Yeah.
57
00:02:50.099 --> 00:02:51.340
It's all seems so long ago now.
58
00:02:51.759 --> 00:02:57.189
Yeah, it does. Um, Humanity made major progress against preventable diseases.
59
00:02:57.270 --> 00:03:00.870
There were some remarkable public health victories that received comparatively
60
00:03:00.949 --> 00:03:13.580
little attention. Malawi and Vanuatu eliminated trachoma as a public
61
00:03:13.599 --> 00:03:16.699
health problem, protecting future generations from one of the world's
62
00:03:16.800 --> 00:03:21.770
leading infectious causes of blindness. Guinea worm disease also moved
63
00:03:21.889 --> 00:03:23.780
tantalizingly close to eradication.
64
00:03:23.810 --> 00:03:26.379
You know, you know who, uh, that was, uh, that
65
00:03:26.400 --> 00:03:28.039
was Jimmy Carter's Jimmy Carter.
66
00:03:28.060 --> 00:03:28.539
Yeah.
67
00:03:28.560 --> 00:03:28.860
Yeah.
68
00:03:29.060 --> 00:03:33.219
Yeah. Jimmy Carter basically taught these African countries how to
69
00:03:33.719 --> 00:03:36.860
get rid of Guinea worm by, um, what they would
70
00:03:36.900 --> 00:03:39.919
do is the Guinea worms would bury, bury themselves in your,
71
00:03:40.319 --> 00:03:43.500
or come out of your body through your legs and stuff.
72
00:03:43.539 --> 00:03:46.090
It was very painful. And what they would do was,
73
00:03:46.449 --> 00:03:51.689
you know, to soothe themselves is wash their, wounds in
74
00:03:52.050 --> 00:03:54.569
a pond or a lake, the same lake that they
75
00:03:54.590 --> 00:03:57.129
drank out of. So it was this perpetual thing. So
76
00:03:57.169 --> 00:04:00.189
he basically taught them to use like nylons, you know,
77
00:04:00.349 --> 00:04:04.310
pantyhose or whatever to strain the water. And yeah, that
78
00:04:04.349 --> 00:04:05.379
was the beginning of the end.
79
00:04:05.439 --> 00:04:06.810
Water filters just break the loop.
80
00:04:06.849 --> 00:04:07.009
Yeah.
81
00:04:07.030 --> 00:04:07.789
Brilliant.
82
00:04:08.129 --> 00:04:12.699
At the COP 15 biodiversity conference in Montreal, more than
83
00:04:12.699 --> 00:04:15.740
190 countries agreed to a framework aimed at protecting at
84
00:04:15.759 --> 00:04:20.029
least 30% of the world's land, inland waters, and oceans
85
00:04:20.089 --> 00:04:26.699
by 2030. Meanwhile, conservation programs produced tangible successes. Nepal's wild
86
00:04:26.800 --> 00:04:31.259
tiger population had nearly tripled, and several species believed lost
87
00:04:31.319 --> 00:04:34.000
or extinct were rediscovered. So I wanted to put a
88
00:04:34.019 --> 00:04:36.699
little good news at the end of that list. And
89
00:04:36.740 --> 00:04:40.050
there's more stuff, but what's happening in tech and space
90
00:04:40.110 --> 00:04:40.399
in 2022, Richard?
91
00:04:42.009 --> 00:04:46.470
All right, let's start in February with Ingenuity Flight 19.
92
00:04:46.449 --> 00:04:48.959
So this is the little helicopter that they put on
93
00:04:49.000 --> 00:04:50.879
Mars with the Perseverance rover. They had a little bit
94
00:04:50.899 --> 00:04:53.639
of payload left over. So let's do this experiment. You
95
00:04:53.660 --> 00:04:57.079
can actually make something fly in this crazy low pressure,
96
00:04:57.139 --> 00:05:03.050
like one-tenth of 1% of the atmosphere. So very high-speed
97
00:05:03.149 --> 00:05:05.990
counter-rotating blades. It was only supposed to work for five flights,
98
00:05:06.050 --> 00:05:09.670
just to test. And now it's on flight 19. So
99
00:05:09.689 --> 00:05:13.730
it's no longer experimental. It's actually a useful device for
100
00:05:13.829 --> 00:05:18.829
the Perseverance rover. It's doing scouting. It's looking for terrain problems.
101
00:05:18.910 --> 00:05:23.250
It's looking for potential scientifically interesting locations. So a huge
102
00:05:23.470 --> 00:05:25.430
step forward. It really has opened the door to a
103
00:05:25.470 --> 00:05:30.220
whole other way to explore other planetary bodies. In April...
104
00:05:31.189 --> 00:05:34.600
We hit the 120 successful Falcon 9 flights, which breaks
105
00:05:34.639 --> 00:05:38.790
the all-time record for everything. So SpaceX is just stretching
106
00:05:38.829 --> 00:05:40.470
out the lead. And one of those flights in April
107
00:05:40.589 --> 00:05:44.209
is the Axiom 1 flight to the International Space Station.
108
00:05:44.240 --> 00:05:48.189
This is a fully... private mission to the space station
109
00:05:48.449 --> 00:05:52.110
funded through Axiom Space as one of their fundraising strategies
110
00:05:52.129 --> 00:05:55.060
for building their own space station using Crew Dragon. Of course,
111
00:05:55.079 --> 00:05:57.240
the International Space Station is not the only space station.
112
00:05:57.279 --> 00:06:00.560
It's also the Tiangong Space Station. It's getting regular crew rotations,
113
00:06:00.660 --> 00:06:03.759
one in June and one in November. In July, we
114
00:06:03.800 --> 00:06:07.819
get our first images from James Webb. So launched in 2021,
115
00:06:07.819 --> 00:06:10.769
takes about six months to do that crazy deployment where
116
00:06:10.810 --> 00:06:12.430
it has to unfold the heat shield and all that
117
00:06:12.449 --> 00:06:15.040
cool stuff, get out to the L2 point. And we
118
00:06:15.060 --> 00:06:18.939
start getting our initial images back. And, of course, it's astonishing.
119
00:06:19.019 --> 00:06:22.459
Astonishing. And it's still astonishing. I think I saw an
120
00:06:22.519 --> 00:06:26.879
article where it just found an incredibly habitable planet.
121
00:06:27.000 --> 00:06:27.860
Possibly, yeah.
122
00:06:28.180 --> 00:06:31.000
It's some 14 light years away. Yeah.
123
00:06:31.120 --> 00:06:33.019
And we're getting to the point now where we can
124
00:06:33.040 --> 00:06:34.730
start to measure atmospheres.
125
00:06:35.170 --> 00:06:35.439
Yeah.
126
00:06:35.990 --> 00:06:39.250
And so– This is not what Star Trek told us
127
00:06:39.269 --> 00:06:40.810
would happen. You don't have to fly to the solar
128
00:06:40.850 --> 00:06:42.910
system to evaluate. We're going to have a map of
129
00:06:42.930 --> 00:06:46.250
all these planets long before we can get there. In September,
130
00:06:46.350 --> 00:06:49.250
the DART mission, so this is the Double Asteroid Redirect Test,
131
00:06:50.209 --> 00:06:56.060
called that because it's actually a binary asteroid system. Didymos
132
00:06:56.329 --> 00:06:59.860
is this asteroid that flies relatively close to the Earth.
133
00:06:59.920 --> 00:07:02.680
It's not huge, 780 meters across. You don't want it
134
00:07:02.720 --> 00:07:05.779
in your backyard or anything, but It's not a tremendously
135
00:07:05.800 --> 00:07:09.480
large rock, but it has a moon, which is called Dimorphos,
136
00:07:09.500 --> 00:07:12.759
although most people refer to it as Diddy Moon. Much smaller,
137
00:07:12.819 --> 00:07:17.220
about 160 meters across, and it orbits the larger asteroid
138
00:07:17.240 --> 00:07:19.980
about every 12 hours, 11 hours, 55 minutes. And so
139
00:07:20.720 --> 00:07:23.240
they picked this target because A, they could get to it. B,
140
00:07:23.279 --> 00:07:25.300
they could measure it, like they can actually see it
141
00:07:25.379 --> 00:07:28.139
well enough to know what's going on. And so they
142
00:07:28.189 --> 00:07:31.430
have this little impactor, and their goal was to alter
143
00:07:31.470 --> 00:07:34.790
the orbit by 73 seconds. So hit the moon so
144
00:07:34.810 --> 00:07:40.490
that it actually slows down and will orbit closer to Didymos.
145
00:07:41.920 --> 00:07:44.230
They overshoot the mark by a lot. It turns out
146
00:07:44.310 --> 00:07:46.930
that the moon is a rubble pile. It's just a
147
00:07:46.990 --> 00:07:49.709
bunch of little rocks all sort of clumped together with
148
00:07:49.750 --> 00:07:52.970
a weak gravity. And so the impact creates this huge
149
00:07:53.170 --> 00:07:57.639
spray of rocks behind, like kilometers long, and actually reduces
150
00:07:57.699 --> 00:08:00.420
the orbit by 33 minutes. And so the first time
151
00:08:00.500 --> 00:08:03.500
that humanity has demonstrated its ability to actually modify the
152
00:08:03.540 --> 00:08:06.790
behavior of an asteroid through an impact. There's another mission
153
00:08:07.149 --> 00:08:10.389
today on the way back to that binary moon system
154
00:08:10.470 --> 00:08:12.170
to do additional measurements, which is cool.
155
00:08:12.189 --> 00:08:12.889
Yep, I remember that.
156
00:08:12.910 --> 00:08:15.670
And finally on the space side in November, Artemis 1.
157
00:08:15.670 --> 00:08:19.230
So this was the first flight of the Orion spacecraft
158
00:08:19.470 --> 00:08:21.779
all up, although it didn't actually have a life support
159
00:08:21.800 --> 00:08:23.360
system on board, which is good because there's no people
160
00:08:23.420 --> 00:08:26.399
on it. But it did fly out to the moon
161
00:08:26.560 --> 00:08:29.540
and do a bunch of maneuvers we would never try
162
00:08:29.600 --> 00:08:33.019
with humans. You know, we hadn't flown this kind of
163
00:08:33.059 --> 00:08:36.330
spacecraft out around the moon before. in decades and we
164
00:08:36.389 --> 00:08:39.659
know a lot more about gravitation and grav and gravitational
165
00:08:39.679 --> 00:08:42.039
maneuvering now and so they basically did a set of
166
00:08:42.100 --> 00:08:45.360
experiments while it was unmanned to just show that they
167
00:08:45.419 --> 00:08:48.580
could do some really cool orbit changes and really energy
168
00:08:48.639 --> 00:08:52.490
efficient maneuvering and it'll come back in late november quite
169
00:08:52.509 --> 00:08:56.649
successfully de-orbit and return to the earth intact although the
170
00:08:56.690 --> 00:08:59.250
heat shield will take damage that makes them unhappy and
171
00:08:59.269 --> 00:09:02.980
they'll have to do some redesigns for artemis ii All right.
172
00:09:03.000 --> 00:09:03.879
Ready for some computing?
173
00:09:04.279 --> 00:09:04.879
Yeah.
174
00:09:04.919 --> 00:09:09.919
I'm going to start in March with this AI paper.
175
00:09:10.039 --> 00:09:15.299
It's called the Training Compute Optimizing Large Language Models. It's
176
00:09:15.350 --> 00:09:18.970
actually a paper out of DeepMinds. This is Dennis Hasselbiss
177
00:09:19.029 --> 00:09:21.509
and a bunch of folks working together. It's generally referred
178
00:09:21.549 --> 00:09:25.269
to as the chinchilla result. But what it really talks about,
179
00:09:25.289 --> 00:09:28.090
they had the computing scaling laws paper back in 2020.
180
00:09:28.090 --> 00:09:31.659
And this is sort of a counter to that, which says, actually,
181
00:09:32.220 --> 00:09:36.730
Most of these models are under-trained, that they don't have
182
00:09:36.799 --> 00:09:39.990
enough data, and that just making the bigger isn't going
183
00:09:40.009 --> 00:09:42.289
to be effective. You actually have to change, go down
184
00:09:42.309 --> 00:09:45.909
to fewer parameters with more training data to get actually
185
00:09:45.990 --> 00:09:48.690
more optimal results. And that will affect a bunch of
186
00:09:48.730 --> 00:09:50.269
things going forward. Now, remember, it's 2022.
187
00:09:50.250 --> 00:09:51.330
Right, right.
188
00:09:51.450 --> 00:09:53.029
So the crazy doesn't happen until the end of the year.
189
00:09:53.090 --> 00:09:53.750
Yeah.
190
00:09:53.769 --> 00:09:55.190
But there is a bunch of other cool stuff that
191
00:09:55.230 --> 00:09:58.360
happens this year on the, in the machine learning space,
192
00:09:58.379 --> 00:09:59.879
because this is the year, like in April, we get
193
00:09:59.919 --> 00:10:00.740
Dolly too.
194
00:10:01.159 --> 00:10:01.360
Right.
195
00:10:01.399 --> 00:10:03.919
And we get mid journey in July and we get
196
00:10:03.940 --> 00:10:07.919
stable division in August. So this is the era of
197
00:10:08.639 --> 00:10:11.960
the image generators. This is also when the Getty images
198
00:10:12.039 --> 00:10:13.159
crisis kicks off.
199
00:10:13.279 --> 00:10:13.679
Oh yes.
200
00:10:13.759 --> 00:10:14.120
I remember.
201
00:10:14.159 --> 00:10:16.850
Because all these images keep getting generated with mangled Getty
202
00:10:16.889 --> 00:10:17.690
logos on them.
203
00:10:18.940 --> 00:10:22.450
And they say, no, we're not using public domain.
204
00:10:22.490 --> 00:10:24.570
Copyrighted materials. I don't know what you're talking about.
205
00:10:24.649 --> 00:10:25.389
What are you talking about?
206
00:10:25.509 --> 00:10:27.320
I don't know where that logo came from. What?
207
00:10:29.730 --> 00:10:30.950
A lot of hand-waving.
208
00:10:31.509 --> 00:10:31.789
All right.
209
00:10:32.350 --> 00:10:34.210
On the Microsoft side, in May, we finally get the
210
00:10:34.250 --> 00:10:37.639
full release of MAUI. We were supposed to get MAUI with.
211
00:10:37.769 --> 00:10:41.230
NET 6 back in November 2021. Right. But for a
212
00:10:41.549 --> 00:10:44.970
variety of reasons, mostly a new dev environment and a
213
00:10:45.029 --> 00:10:47.500
new framework at the same time made that kind of impossible.
214
00:10:47.559 --> 00:10:49.759
So the feature set that they really wanted to get to,
215
00:10:49.799 --> 00:10:52.480
they get to in May. They'll get major updates by
216
00:10:52.580 --> 00:10:54.700
November when we get. NET 7 and Studio 2022 and
217
00:10:54.720 --> 00:11:01.190
C Sharp 11. All good stuff. Yes. And also in November,
218
00:11:01.250 --> 00:11:04.830
we get the collapse of FTX. Do you remember this?
219
00:11:05.690 --> 00:11:06.570
This is one of those.
220
00:11:06.590 --> 00:11:07.950
Is that a crypto thing?
221
00:11:08.169 --> 00:11:10.580
Yeah, that's right. And one would argue this is the
222
00:11:10.669 --> 00:11:16.320
moment that creates a crypto winter that arguably they have
223
00:11:16.379 --> 00:11:18.620
not recovered from. And I'm having a hard time being
224
00:11:18.679 --> 00:11:22.480
upset about it. Because also in November, everybody gets really,
225
00:11:23.139 --> 00:11:27.799
really very distracted. when OpenAI pushes out ChatGPT.
226
00:11:28.120 --> 00:11:31.669
I was working for a customer that was doing crypto
227
00:11:32.129 --> 00:11:34.629
exchanges or something. I'm not going to get into it,
228
00:11:34.700 --> 00:11:38.340
but The project abruptly stopped after that.
229
00:11:38.399 --> 00:11:39.320
Yeah, oddly enough. Yeah.
230
00:11:39.460 --> 00:11:41.940
Because the demand just shriveled up.
231
00:11:42.019 --> 00:11:44.659
Sure. And suddenly we have another thing to use GPUs for, right?
232
00:11:44.700 --> 00:11:48.000
Because we've been dancing around with these techs for a while.
233
00:11:48.080 --> 00:11:52.259
But ChatGPT is the, I would argue, accidental breakthrough product
234
00:11:52.539 --> 00:11:54.299
in the sense that- I would say that. They were
235
00:11:54.340 --> 00:11:56.539
trying to do second order training. They were running low
236
00:11:56.620 --> 00:11:59.220
on money. And so- putting it out in the public
237
00:11:59.240 --> 00:12:01.070
was another way to do second order training for free
238
00:12:01.610 --> 00:12:03.830
uh and then they signed up 100 million users in
239
00:12:04.110 --> 00:12:06.909
two months yep and before that the current the record
240
00:12:06.990 --> 00:12:09.549
was tiktok had gotten 100 million users in nine months
241
00:12:09.590 --> 00:12:11.509
so two months was kind of a big deal but
242
00:12:11.529 --> 00:12:14.950
that is a story for 2023 yeah sure is that's
243
00:12:14.970 --> 00:12:15.690
what i got dude.
244
00:12:15.690 --> 00:12:19.309
All right well uh did i say matthias carlson is
245
00:12:19.350 --> 00:12:22.129
here he's been very silent but matthias feel free to
246
00:12:22.169 --> 00:12:23.909
jump in if you want yeah i'm going to uh
247
00:12:24.620 --> 00:12:26.000
Carry on with Better Know a Framework.
248
00:12:26.019 --> 00:12:26.620
Awesome.
249
00:12:26.779 --> 00:12:35.980
All right, man, what do you got? Okay, so you
250
00:12:36.019 --> 00:12:38.990
know that I've been working with local models on a
251
00:12:39.350 --> 00:12:45.129
GPU machine, high GPU machine here at Pop Studios. And
252
00:12:45.149 --> 00:12:48.200
I've been going through, you know, first it was a Lama,
253
00:12:48.500 --> 00:12:52.179
and then I talked about Lama CPP, which was better.
254
00:12:53.120 --> 00:12:56.159
And working with Lama CPP and Quen 3.8, which I
255
00:12:56.200 --> 00:13:02.039
think is the state-of-the-art model right now for locals, it
256
00:13:02.059 --> 00:13:04.860
just would bonk. And, you know, it was great for
257
00:13:04.879 --> 00:13:07.279
the demo, but, you know, then I tried to use
258
00:13:07.320 --> 00:13:09.679
it afterwards, and it was bonking and bonking. And what
259
00:13:09.720 --> 00:13:12.440
I mean by that, it was, you know, just stopping.
260
00:13:13.500 --> 00:13:16.120
And we couldn't really figure out why.
261
00:13:16.139 --> 00:13:16.279
So....
262
00:13:18.720 --> 00:13:21.820
I landed on, and this is for episode 45 of
263
00:13:21.980 --> 00:13:24.940
Code It With AI, I landed on NINFR.
264
00:13:25.220 --> 00:13:25.899
What's a NINFR?
265
00:13:26.179 --> 00:13:29.289
It's a runtime for hosting your models, just like Alama
266
00:13:29.450 --> 00:13:34.649
is in Alama CPP. However, it's really, really good. And
267
00:13:34.950 --> 00:13:38.230
I've been using it now as of this recording for
268
00:13:38.269 --> 00:13:43.990
a week, and it's been wonderful. I'm using a 128K
269
00:13:44.029 --> 00:13:47.360
context model that also has vision so you can upload
270
00:13:47.389 --> 00:13:48.320
screenshots to it.
271
00:13:48.340 --> 00:13:48.840
Nice.
272
00:13:48.860 --> 00:13:53.580
And I've been running it steadily. I've had one terminal
273
00:13:53.700 --> 00:13:57.840
open for four days, and I'm just talking to it
274
00:13:58.200 --> 00:14:00.679
like I would talk to ChatGPT, and it's great.
275
00:14:00.919 --> 00:14:01.139
Wow.
276
00:14:01.220 --> 00:14:03.639
Of course, you have to do your own memory management
277
00:14:04.470 --> 00:14:08.529
to keep state about things that you do. And I
278
00:14:08.590 --> 00:14:11.470
solved that. I'm not sharing this, but I have a repo.
279
00:14:12.629 --> 00:14:16.899
that is private, that's just called agent memory. And I
280
00:14:16.960 --> 00:14:19.840
have a whole bunch of MD files in there, including
281
00:14:19.860 --> 00:14:22.960
a startup MD that I tell it what to do,
282
00:14:23.460 --> 00:14:26.899
all the rules that it needs to obey, including before
283
00:14:26.940 --> 00:14:31.000
it does a compaction. I tell it to go reread
284
00:14:31.159 --> 00:14:35.279
this particular very important MD file because it has rules
285
00:14:35.320 --> 00:14:38.940
that fall off after the compaction happens.
286
00:14:38.960 --> 00:14:40.659
Cool so you create your own harness there so you
287
00:14:40.679 --> 00:14:42.500
have an own harness system.
288
00:14:42.559 --> 00:14:45.039
Yeah i have my own harness system exactly and i'm
289
00:14:45.059 --> 00:14:48.559
using co-pilot cli so it has you know some it
290
00:14:48.620 --> 00:14:51.580
is a good harness but you know if you if
291
00:14:51.620 --> 00:14:56.159
you're dealing with these llms you really have to customize
292
00:14:56.419 --> 00:14:58.740
i have i think you have to customize the the
293
00:14:58.759 --> 00:15:01.440
memory structure and what it's supposed to do and i
294
00:15:01.559 --> 00:15:04.830
even have um a carl md file up there that
295
00:15:05.279 --> 00:15:09.000
has all the things that are about me, right? So
296
00:15:09.059 --> 00:15:11.279
if it learns that, oh, you play in a Steely
297
00:15:11.299 --> 00:15:14.919
Dan cover band, it puts that up there, right? And, oh,
298
00:15:15.639 --> 00:15:19.049
you know, Carl eats fish on Friday. Okay, well, put
299
00:15:19.110 --> 00:15:23.129
that up there. So anyway, I wanted to share this
300
00:15:23.210 --> 00:15:27.879
document that goes through the search for a stable local
301
00:15:28.120 --> 00:15:31.500
coding LLM and all of the different things that I
302
00:15:31.539 --> 00:15:38.909
tried I tried VLLM. I tried SGLang. Ollama got 128K
303
00:15:39.610 --> 00:15:45.029
context back, but the copilot sessions became pathological.
304
00:15:45.730 --> 00:15:46.000
Wow.
305
00:15:46.500 --> 00:15:48.100
And NINFR is the clear winner.
306
00:15:48.379 --> 00:15:48.840
Cool.
307
00:15:49.000 --> 00:15:50.139
I'll have to look at that.
308
00:15:50.340 --> 00:15:52.820
I think it should be pronounced NINFR.
309
00:15:53.000 --> 00:15:54.539
Yeah, I think you're right. NINFR.
310
00:15:54.639 --> 00:15:54.899
Yeah.
311
00:15:54.919 --> 00:15:55.340
That's right.
312
00:15:55.480 --> 00:16:00.899
And it wants 64-bit Linux, a 5090... This is what
313
00:16:00.940 --> 00:16:02.860
you did with that 5090 machine you built?
314
00:16:02.940 --> 00:16:03.360
That's right.
315
00:16:03.440 --> 00:16:03.740
Wow.
316
00:16:03.919 --> 00:16:08.759
Yeah, so it's basically made for this GPU and for
317
00:16:08.820 --> 00:16:10.600
this model and infer.
318
00:16:10.659 --> 00:16:13.059
So how does it handle concurrency and things? That's sometimes
319
00:16:13.100 --> 00:16:14.220
hard to know when they die.
320
00:16:14.320 --> 00:16:16.600
Well, I don't have more than one person using it
321
00:16:16.639 --> 00:16:17.080
at a time.
322
00:16:17.539 --> 00:16:20.450
I was talking about multiple agents and things.
323
00:16:21.429 --> 00:16:24.230
Yeah, if I did, then I probably would run into
324
00:16:25.429 --> 00:16:29.750
some bonking or some context. The context would be used up. more,
325
00:16:29.950 --> 00:16:32.850
you know, more quickly, but it's just for me. And
326
00:16:33.110 --> 00:16:36.840
I would also add that Copilot CLI has a slash
327
00:16:36.899 --> 00:16:41.059
remote option so that what you can do is pull
328
00:16:41.139 --> 00:16:46.090
up GitHub on your phone and go to Copilot and
329
00:16:46.120 --> 00:16:48.309
you see the sessions that are there. And if you
330
00:16:48.350 --> 00:16:50.509
click on the last session you were in, you can
331
00:16:50.669 --> 00:16:55.429
actually operate that Copilot CLI session from your phone. And
332
00:16:55.480 --> 00:16:57.519
it goes through GitHub. You don't need a VPN or
333
00:16:57.559 --> 00:17:00.159
anything like that. So you're authenticated on your phone. You're
334
00:17:00.200 --> 00:17:04.180
authenticated in the CLI. And you can even see what
335
00:17:04.220 --> 00:17:07.140
you type on the screen in the CLI terminal. Like
336
00:17:07.200 --> 00:17:09.700
it's really, really good. And it works with the local model.
337
00:17:09.920 --> 00:17:12.099
So that's what I got. Cool. Getting there. All right, man.
338
00:17:12.259 --> 00:17:12.500
All right.
339
00:17:12.759 --> 00:17:14.700
One more thing. Who's talking to us today, Richard?
340
00:17:14.839 --> 00:17:17.180
Grabbed a comment off the show 2017, the one we
341
00:17:17.220 --> 00:17:20.309
did with Billy Hollis back in August, which I just
342
00:17:20.349 --> 00:17:22.309
called arguing about AI because I think we did that.
343
00:17:22.410 --> 00:17:24.329
That's pretty much what we did.
344
00:17:24.390 --> 00:17:29.690
Yeah. And oddly enough, generated a lot of conversation. This
345
00:17:29.730 --> 00:17:33.259
particular one comes from Volkmar, who says, hi, good discussion. However,
346
00:17:33.299 --> 00:17:35.180
I think it was more from a developer point of
347
00:17:35.200 --> 00:17:37.900
view than from a software owner's point of view. I
348
00:17:37.940 --> 00:17:39.480
have been listening to your show for many years now,
349
00:17:39.660 --> 00:17:43.099
often sitting on a tractor in Germany as I'm a
350
00:17:43.220 --> 00:17:47.009
farmer and a software developer at the same time. So cool.
351
00:17:47.089 --> 00:17:47.589
So cool.
352
00:17:48.359 --> 00:17:50.339
I was working in the past also for a precision
353
00:17:50.420 --> 00:17:55.660
farming company designing web applications, starting with Silverlight. There were
354
00:17:56.220 --> 00:17:58.809
always so many ideas and so little resources to implement
355
00:17:58.829 --> 00:18:01.710
them back then. If you're a specialist in any field
356
00:18:02.089 --> 00:18:04.509
and a developer, it's like you get a genie in
357
00:18:04.529 --> 00:18:09.269
a bottle. Yeah, domain knowledge is valuable. In the end,
358
00:18:09.309 --> 00:18:11.470
it is the value for the user or customer of
359
00:18:11.490 --> 00:18:15.109
the software we can develop for real needs now, not
360
00:18:15.170 --> 00:18:17.829
only where the clients are making a good margin. I
361
00:18:17.869 --> 00:18:20.789
also believe that AI-generated software without an experienced developer in
362
00:18:20.809 --> 00:18:25.190
the driver's seat will not be successful. Keep it up. Thanks, Baltimore.
363
00:18:25.309 --> 00:18:25.609
Nice.
364
00:18:25.690 --> 00:18:26.099
There you go.
365
00:18:26.190 --> 00:18:27.579
Sounds like a Greenfield developer.
366
00:18:28.319 --> 00:18:31.460
That's really a Greenfield developer, my dears. You are right.
367
00:18:33.700 --> 00:18:35.799
How did I leave that on the table? Thank you
368
00:18:35.819 --> 00:18:36.819
for picking it up.
369
00:18:37.559 --> 00:18:40.269
Mateus is like, I can't believe any of these guys
370
00:18:40.289 --> 00:18:41.170
have thought of that yet.
371
00:18:41.269 --> 00:18:44.190
How do you leave that one alone? That's crazy. I know.
372
00:18:44.210 --> 00:18:48.970
I think he starts to spring on a brownfield, actually. Hey, Volkmar,
373
00:18:48.990 --> 00:18:51.599
thank you so much for your comment. Keep on farming
374
00:18:51.640 --> 00:18:53.779
and listen to us while you're driving around. And to
375
00:18:53.799 --> 00:18:55.460
give you some more to listen to, I'd like to
376
00:18:55.500 --> 00:18:57.099
send you a copy of Music to Codebyte. If you'd
377
00:18:57.140 --> 00:18:58.700
like a copy of Music to Codebyte, write a comment
378
00:18:58.720 --> 00:19:01.740
on the website at totnetrocks.com or on the Facebooks. We
379
00:19:01.789 --> 00:19:03.670
publish every show there and every comment there and everything
380
00:19:03.690 --> 00:19:05.230
on the show, we'll send you a copy of Music
381
00:19:05.250 --> 00:19:05.490
to Codebyte.
382
00:19:05.509 --> 00:19:06.769
Or if you just want to go get it, go
383
00:19:06.809 --> 00:19:12.170
to musictocodebyte.net and you can download the entire collection in MP3, WAV,
384
00:19:12.509 --> 00:19:13.390
and FLAC formats.
385
00:19:13.950 --> 00:19:14.119
Cool.
386
00:19:14.339 --> 00:19:15.519
All right, Matthias, you're up.
387
00:19:15.799 --> 00:19:16.440
What are you doing?
388
00:19:17.210 --> 00:19:21.170
This is not about cake. Let me read your bio,
389
00:19:21.230 --> 00:19:24.900
first of all. Matthias Carlsen is a partner and senior
390
00:19:24.940 --> 00:19:31.099
architect at WCOM AB in Gultenberg, Sweden, a Microsoft.net MVP
391
00:19:31.200 --> 00:19:35.480
and a passionate open source citizen. Matthias has been professionally
392
00:19:36.180 --> 00:19:40.490
in IT for almost 30 years, started coding already in
393
00:19:40.750 --> 00:19:44.690
the 1980s on a Commodore 128. But the true love
394
00:19:44.730 --> 00:19:49.269
for coding came with the Amiga. Outside of technology, Matthias
395
00:19:49.329 --> 00:19:51.519
is the father of two and the husband of one.
396
00:19:53.779 --> 00:19:55.500
All right, what are you doing these days, Matthias?
397
00:19:55.859 --> 00:19:59.559
Well, I'm a developer, mostly. I'm fielding AI like most
398
00:19:59.599 --> 00:20:02.180
people do, but I'm trying to do it in with
399
00:20:02.200 --> 00:20:05.880
some sense, because what I've seen is how we maintain
400
00:20:05.920 --> 00:20:08.059
things and stuff like that, but that's not what I'm
401
00:20:08.079 --> 00:20:12.059
going to talk about today. I think it's tech-related, but
402
00:20:12.400 --> 00:20:13.599
more soft in one way.
403
00:20:13.759 --> 00:20:16.839
I think I just heard thousands of people give a
404
00:20:16.880 --> 00:20:20.319
sigh of relief. Finally, those guys are talking about something
405
00:20:20.359 --> 00:20:20.799
else than AI.
406
00:20:21.019 --> 00:20:24.309
Yeah, it's even more important, I think. It's even important,
407
00:20:24.630 --> 00:20:28.390
because I got talking with Richard about... Well, I think
408
00:20:28.430 --> 00:20:33.180
we all here now in this room is... people with experience.
409
00:20:33.579 --> 00:20:36.180
I will talk about things like I'm not a lawyer,
410
00:20:36.359 --> 00:20:40.380
but I have learned things by just the earth spinning around.
411
00:20:42.119 --> 00:20:45.619
And one thing you notice when you get to a
412
00:20:45.619 --> 00:20:51.019
certain age is that, well, things like people that you
413
00:20:51.039 --> 00:20:56.309
thought wouldn't be there forever aren't. And you will notice
414
00:20:56.329 --> 00:21:01.529
that Well, us technologists, we are hoarders in some ways,
415
00:21:01.809 --> 00:21:08.250
both with our GitHub repositories, our technology, our retro equipment,
416
00:21:08.710 --> 00:21:09.410
or whatnot.
417
00:21:09.730 --> 00:21:11.500
Our drive folders full of old projects.
418
00:21:11.799 --> 00:21:16.039
Yeah. Before. I mean, I think what I reached out
419
00:21:16.059 --> 00:21:18.640
to Richard is, like, I'm not unique in this situation.
420
00:21:18.660 --> 00:21:19.240
I think we are.
421
00:21:19.259 --> 00:21:19.319
No.
422
00:21:19.319 --> 00:21:23.400
And there are many of us that live in what
423
00:21:23.460 --> 00:21:27.160
Hanselman said, like, mixed marriage, where... where there's one geek
424
00:21:27.200 --> 00:21:28.579
and one normal person.
425
00:21:28.380 --> 00:21:36.599
And that's kind of like and like with true experience
426
00:21:36.640 --> 00:21:41.380
I've had some well like collateral damage I lost friends
427
00:21:42.099 --> 00:21:45.589
like and things like that friends that have been.
428
00:21:47.690 --> 00:21:51.369
Like in different ways you can have like temporary with
429
00:21:51.410 --> 00:21:55.750
illness or it could be final terminal things and that's
430
00:21:55.769 --> 00:21:58.109
why I like had some ideas that I would like
431
00:21:58.470 --> 00:21:59.549
have a discussion.
432
00:21:59.170 --> 00:22:01.230
With you about. And I think some things that are
433
00:22:01.289 --> 00:22:03.170
important for us as innovators.
434
00:22:03.710 --> 00:22:05.460
And when you messaged me about this, I'm like, I
435
00:22:05.500 --> 00:22:08.059
had just gone through this with a friend who's passed away,
436
00:22:08.180 --> 00:22:10.380
whose spouse reached out to me and said, how do
437
00:22:10.400 --> 00:22:13.180
I do this? Like, how do I get access to
438
00:22:13.240 --> 00:22:16.279
his stuff? How do I clean up the accounts? How
439
00:22:16.319 --> 00:22:19.130
do I operate the house? He was a smart home guy. Like,
440
00:22:19.150 --> 00:22:22.079
and all of a sudden you're like, I'm trying to
441
00:22:22.180 --> 00:22:26.960
sysadmin for his place, uh, without him around. Like, and
442
00:22:26.980 --> 00:22:29.269
he ain't, he's not coming back. So I, you know,
443
00:22:29.289 --> 00:22:32.289
it turned into a very much forensic effort. What do
444
00:22:32.309 --> 00:22:32.910
we got to do?
445
00:22:33.150 --> 00:22:34.730
And that was still like the impotence for me also.
446
00:22:34.769 --> 00:22:36.869
Like I have actually a couple of years ago, I
447
00:22:36.869 --> 00:22:39.480
had a friend that like I sat with his wife
448
00:22:39.519 --> 00:22:42.279
when he passed away. And the whole thing there is was,
449
00:22:42.700 --> 00:22:45.569
and I, That's why I had this idea to talk
450
00:22:45.650 --> 00:22:47.619
about in draft for many years, because it was difficult
451
00:22:47.640 --> 00:22:52.500
when it happened. We spent hours going through. If someone's
452
00:22:52.519 --> 00:22:56.410
been around since the 70s, we gather things. You have
453
00:22:56.470 --> 00:23:02.309
virtual private servers running, source resources, online accounts, subscriptions. And
454
00:23:02.549 --> 00:23:03.490
like many of us.
455
00:23:03.509 --> 00:23:05.690
It's going to take somebody weeks to go through all
456
00:23:05.730 --> 00:23:06.109
my stuff.
457
00:23:06.190 --> 00:23:08.549
And like many also like to buy a domain a
458
00:23:08.569 --> 00:23:10.960
week and things like that. So it's, there's, there's a
459
00:23:10.960 --> 00:23:12.799
lot of things. And, uh, there are things that I
460
00:23:12.819 --> 00:23:16.039
think we need to, uh, it's, it's not like, I
461
00:23:16.079 --> 00:23:18.380
think to go a morbid way, it's just like, I
462
00:23:18.400 --> 00:23:22.160
think it matters to, to have some, like, I don't
463
00:23:22.240 --> 00:23:24.509
want to put my wife and kids in a situation
464
00:23:24.549 --> 00:23:28.069
that I've worked with my friend because it's And also
465
00:23:28.089 --> 00:23:33.130
subsigned the right home for the right thing. For a
466
00:23:33.730 --> 00:23:39.880
foreseeable future, I realized that even if I'm not here,
467
00:23:40.059 --> 00:23:41.259
the earth will still turn around.
468
00:23:41.279 --> 00:23:41.460
Yeah, sure.
469
00:23:42.980 --> 00:23:45.460
And that's why I like and I know some have
470
00:23:45.599 --> 00:23:48.079
seen the like the Swedes things like about death cleaning
471
00:23:48.140 --> 00:23:50.339
and things that like, how do we get our things
472
00:23:50.400 --> 00:23:52.500
in order? Or how do we have.
473
00:23:52.359 --> 00:23:55.480
Nobody told my Swedish mother about Swedish death cleaning. And
474
00:23:55.539 --> 00:23:56.660
so it all fell on me.
475
00:23:56.920 --> 00:23:59.960
Yeah, because we're like, we're, we're trying not to be
476
00:24:00.019 --> 00:24:02.329
too emotional, but we are very practical. I think we
477
00:24:02.349 --> 00:24:07.170
are Swedes. And like, it's not stoic. And also, it's
478
00:24:07.210 --> 00:24:12.359
not only about death is by things like hospital stays, burnout, accidents,
479
00:24:12.440 --> 00:24:15.279
and things that, things can happen where you have to
480
00:24:15.559 --> 00:24:18.220
like think, I think it's good to, just like you
481
00:24:18.240 --> 00:24:20.960
do with a contractor agreement, you have to do things
482
00:24:21.039 --> 00:24:24.210
sometimes before it happens to make it easier when it happens.
483
00:24:24.230 --> 00:24:25.250
Sure, yeah.
484
00:24:25.509 --> 00:24:30.750
I'm lucky because my wife is technical and she, I
485
00:24:30.769 --> 00:24:33.190
give her complete access to everything that I have. We
486
00:24:33.250 --> 00:24:38.799
share the password manager and all that stuff. So she
487
00:24:38.880 --> 00:24:42.309
goes and notices things that I've paid for that I'm
488
00:24:42.349 --> 00:24:45.369
no longer using and gives me a list like every
489
00:24:45.410 --> 00:24:47.910
couple of weeks. Hey, do you still use this? Do
490
00:24:47.930 --> 00:24:48.789
you still use that?
491
00:24:48.869 --> 00:24:48.990
Yeah.
492
00:24:49.029 --> 00:24:49.670
What is this?
493
00:24:50.109 --> 00:24:50.750
Yeah. Yeah.
494
00:24:50.789 --> 00:24:51.369
What is this?
495
00:24:51.390 --> 00:24:53.839
That's usually what I use a credit card expiry for
496
00:24:53.880 --> 00:24:56.599
because then like payment failed. And I was supposed to
497
00:24:56.640 --> 00:24:58.500
pay him for it.
498
00:24:58.660 --> 00:25:02.559
And we do have documentation online that talks about all
499
00:25:02.599 --> 00:25:05.990
the things that that I do that have to happen afterwards.
500
00:25:06.230 --> 00:25:12.579
And also, I have a great big honking NAS drive here.
501
00:25:13.579 --> 00:25:16.460
And it's got way more stuff on it than it
502
00:25:16.500 --> 00:25:17.480
should have.
503
00:25:17.779 --> 00:25:19.339
Yeah, they all have that.
504
00:25:19.960 --> 00:25:22.019
Yeah, because when I came to my work, what I
505
00:25:22.319 --> 00:25:26.509
started to ask myself is like, because some things aren't
506
00:25:26.549 --> 00:25:29.329
that important. They're important to me. So sometimes I decide
507
00:25:29.490 --> 00:25:31.569
what should die with me or what should live on,
508
00:25:31.690 --> 00:25:36.349
like the categories. And also things like who needs access
509
00:25:36.369 --> 00:25:39.890
to what. And things like domains.
510
00:25:40.410 --> 00:25:41.690
What is actually valuable?
511
00:25:43.049 --> 00:25:47.049
And for some open source projects, it's kind of the same.
512
00:25:47.089 --> 00:25:49.099
Because there are some things that are just important to me.
513
00:25:49.500 --> 00:25:52.980
There are some things that are mildly important. And there
514
00:25:53.019 --> 00:25:55.680
are some things that are successful and many depend on.
515
00:25:55.700 --> 00:25:58.880
So it could be different paths depending on what you
516
00:25:58.900 --> 00:25:59.259
have there.
517
00:25:59.279 --> 00:26:01.700
I mean, I got to think, especially if it's abrupt.
518
00:26:02.500 --> 00:26:06.130
It's the day-to-day first, like what bills are, you know,
519
00:26:06.170 --> 00:26:08.950
what charges are being incurred that are no longer relevant
520
00:26:09.000 --> 00:26:12.920
and how do we fix that? what things count on
521
00:26:12.980 --> 00:26:16.349
the email address of the deceased. I think number one is,
522
00:26:16.369 --> 00:26:18.829
can you get access to that email? And maybe an
523
00:26:18.890 --> 00:26:21.549
authenticator device, like the phone and the email account have
524
00:26:21.609 --> 00:26:24.109
got to be the starting point to deal with the
525
00:26:24.210 --> 00:26:27.450
day-to-day before we even get into the legacy.
526
00:26:27.650 --> 00:26:32.000
And the problem is, like, what actually happened on one
527
00:26:32.079 --> 00:26:36.039
process project where the maintainer died is that GitHub is really,
528
00:26:36.059 --> 00:26:38.349
as it should be, you need to go through the
529
00:26:38.369 --> 00:26:42.509
estate to get things. And that's why it's important to
530
00:26:42.609 --> 00:26:46.670
identify what should the successors be for this project. Or
531
00:26:47.210 --> 00:26:49.390
if it's a bigger project, you need co-maintainers so you
532
00:26:49.430 --> 00:26:50.569
have redundancy to things.
533
00:26:50.670 --> 00:26:53.450
Yeah, you could deal with this in advance.
534
00:26:53.710 --> 00:26:57.690
But also things like Carl mentioned, the credentials are important.
535
00:26:57.990 --> 00:27:02.250
What are the... What passwords are needed, who needs what,
536
00:27:02.309 --> 00:27:04.170
and then path for things like that.
537
00:27:04.269 --> 00:27:06.519
When I think about legacy now, I think about what
538
00:27:06.579 --> 00:27:08.960
most people can, you know, what do you carry out
539
00:27:09.000 --> 00:27:13.079
of the house? And it's photo. Like, if the deceased
540
00:27:13.140 --> 00:27:15.960
is the one with the big Google Photos account or
541
00:27:16.000 --> 00:27:19.809
wherever they put it, you know, who has access to that?
542
00:27:19.829 --> 00:27:23.230
Because you lose those, you've really lost something. You talk
543
00:27:23.269 --> 00:27:24.470
about loss of legacy.
544
00:27:24.670 --> 00:27:26.650
Yeah, and I think it's easier when you have the
545
00:27:26.690 --> 00:27:29.430
cloud services because you don't think about it. I like it.
546
00:27:30.470 --> 00:27:31.950
Yeah.
547
00:27:32.299 --> 00:27:35.539
Because I've been, like, when you did the cleaning after
548
00:27:35.579 --> 00:27:38.359
my grandparents passed away, you did the whole physical thing,
549
00:27:38.440 --> 00:27:40.819
like what furniture to keep and things like that. But
550
00:27:41.240 --> 00:27:45.160
it's much harder to know if you don't have this
551
00:27:45.240 --> 00:27:49.069
kind of, like, inventory of your digital, like, bill of
552
00:27:49.109 --> 00:27:53.450
materials of your digital life. Then it's really hard for
553
00:27:53.470 --> 00:27:55.970
someone else to know. And sometimes you find out when
554
00:27:56.019 --> 00:27:58.200
it's too late, when the domain expires or when things
555
00:27:58.240 --> 00:28:02.500
go inactive or whatever. And I think usually what I
556
00:28:02.539 --> 00:28:07.049
did is I did some categories of action. So what
557
00:28:07.109 --> 00:28:13.289
are personal ownerships? This is something that is only important
558
00:28:13.329 --> 00:28:17.200
to me. Then I will remain controlling indefinitely and it
559
00:28:17.220 --> 00:28:19.720
will die with me. It could be something that is
560
00:28:19.920 --> 00:28:24.380
like emergency succession where someone else can take over if
561
00:28:24.460 --> 00:28:28.390
I'm temporarily incapable. This product is someone who can take over.
562
00:28:28.569 --> 00:28:32.809
We're really talking about digital assets, right? Yeah, mostly. I mean,
563
00:28:32.829 --> 00:28:33.589
the physical ones.
564
00:28:33.710 --> 00:28:38.130
Or it could be things like my retro gaming collection is... Like,
565
00:28:38.170 --> 00:28:40.119
my wife has zero interest in that. But I could
566
00:28:40.240 --> 00:28:44.839
find someone... I have friends in the community that value those.
567
00:28:44.960 --> 00:28:45.359
Right, yeah.
568
00:28:45.380 --> 00:28:47.960
So there could be a plan for those, like... They
569
00:28:48.000 --> 00:28:50.519
have monetary value, perhaps, for my wife. Then someone that
570
00:28:50.579 --> 00:28:52.200
knows the stuff can sell it for her.
571
00:28:52.400 --> 00:28:53.369
Or it could be.
572
00:28:53.480 --> 00:28:54.710
So it could be physical things, too.
573
00:28:54.730 --> 00:28:55.829
Sure.
574
00:28:56.009 --> 00:28:56.230
Because.
575
00:28:56.950 --> 00:28:59.589
You can be a geek in many areas. It could
576
00:28:59.630 --> 00:29:03.150
be in bicycles or it could be whatever that you
577
00:29:03.170 --> 00:29:05.420
have a disconnect between. And it could be vice versa
578
00:29:05.440 --> 00:29:07.779
with the wife that she has some special interest that
579
00:29:07.819 --> 00:29:10.980
you don't. And then it can be like, who should
580
00:29:11.019 --> 00:29:14.420
take over this? But it also can be something like
581
00:29:14.920 --> 00:29:17.799
temporary delegation where you just hand over for a while.
582
00:29:18.180 --> 00:29:22.420
And it can also be some open succession where you
583
00:29:22.440 --> 00:29:26.539
essentially can post someone anyone that cares about the source
584
00:29:26.559 --> 00:29:28.480
party or something can put their hand up.
585
00:29:28.769 --> 00:29:30.650
Are we talking about a digital will?
586
00:29:31.029 --> 00:29:33.569
Yeah, it's actually like, well, I think like we have
587
00:29:33.710 --> 00:29:37.250
like the S-bomb for software building materials, but almost like
588
00:29:37.289 --> 00:29:41.869
have a digital building materials or your life's like, what
589
00:29:41.970 --> 00:29:44.619
are the things to have an inventory for things like
590
00:29:44.670 --> 00:29:48.519
services domains? I think it's good to have. And it's quite,
591
00:29:48.640 --> 00:29:50.539
I think that's something you could use AI for now
592
00:29:50.579 --> 00:29:52.880
where like, do absolutely. That's what I would go through
593
00:29:52.900 --> 00:29:57.680
your things. See what, I also like things like personally legacy,
594
00:29:57.700 --> 00:29:59.930
like I want them to remain and I want them
595
00:29:59.950 --> 00:30:01.930
to be active, then I can transfer it to someone
596
00:30:01.970 --> 00:30:06.670
else or things like that. So I think it's good
597
00:30:06.690 --> 00:30:11.269
to have those kind of discussions before you do it yourself,
598
00:30:11.509 --> 00:30:16.670
because As long as you're alive, you decide what should happen.
599
00:30:16.690 --> 00:30:20.000
But also you should go back to that list and see,
600
00:30:20.019 --> 00:30:22.220
because it could change. It could be like when my
601
00:30:22.240 --> 00:30:24.779
kids were five, they weren't interested in code, but it
602
00:30:24.819 --> 00:30:26.819
could be that when they're 20, they're interested in it.
603
00:30:26.859 --> 00:30:28.900
And so it could be that you do some... Same
604
00:30:28.940 --> 00:30:31.529
way you need to refresh any will. Reiterate, yeah.
605
00:30:31.720 --> 00:30:33.210
Hey, we should take a break. Yeah, we should.
606
00:30:33.430 --> 00:30:35.630
We'll be right back after these very important messages. Don't
607
00:30:35.650 --> 00:30:35.890
go away.
608
00:30:39.720 --> 00:30:42.000
And we're back. It's. NET Rocks. I'm Richard Campbell. That's
609
00:30:42.039 --> 00:30:43.289
Carl Franklin. Hey.
610
00:30:43.549 --> 00:30:44.009
We're back.
611
00:30:44.049 --> 00:30:46.609
We're spending some time with our friend, Matthias Carlson, about
612
00:30:46.650 --> 00:30:49.450
this idea of a digital legacy. Yeah, I'm still hung
613
00:30:49.509 --> 00:30:52.450
up on the fundamentals of who could get into my
614
00:30:52.529 --> 00:30:55.430
email account, who could get access to my phone for
615
00:30:55.450 --> 00:31:00.769
my authenticator. Because you're really jammed. And if you were
616
00:31:00.809 --> 00:31:04.680
only using face recognition... You don't have that option. You
617
00:31:04.700 --> 00:31:07.779
better make sure there's some backup strategies that are difficult
618
00:31:07.799 --> 00:31:10.359
to use but possible by someone else.
619
00:31:10.700 --> 00:31:13.359
You can't just count on your spouse or your partner
620
00:31:13.460 --> 00:31:17.619
because you two might die in a fiery crash.
621
00:31:17.779 --> 00:31:19.680
Might go together, the old plane crash thing.
622
00:31:19.700 --> 00:31:23.460
So have a child or somebody at the ready to
623
00:31:23.519 --> 00:31:23.900
do that.
624
00:31:24.380 --> 00:31:25.039
An executor.
625
00:31:25.059 --> 00:31:25.940
An executor, yeah.
626
00:31:26.359 --> 00:31:29.839
But also, because I've been through it, if you have
627
00:31:29.880 --> 00:31:33.700
someone near to you, You have stuff to arrange regardless.
628
00:31:34.500 --> 00:31:37.599
So you would like everything that happens. And that's kind
629
00:31:37.640 --> 00:31:39.569
of like, just to put the prayer, like, I think
630
00:31:39.609 --> 00:31:43.089
it's thoughtful to think of your, what can I do
631
00:31:43.230 --> 00:31:45.529
to make it easy? Like these kinds of things, I
632
00:31:45.569 --> 00:31:47.599
know this friend can take care of this. I can
633
00:31:47.630 --> 00:31:51.019
pass them to like a foundation or something and have
634
00:31:51.059 --> 00:31:53.920
that kind of thing. And that's kind of nice, like
635
00:31:53.960 --> 00:31:55.789
when you see something like the dotnet foundation where they
636
00:31:55.809 --> 00:31:58.190
have like one password and they have keyboards and things
637
00:31:58.210 --> 00:32:04.029
like that where something happens. And I think that's the
638
00:32:04.069 --> 00:32:07.829
discussion you need to have. You want to be thoughtful
639
00:32:07.910 --> 00:32:12.759
of the people around you. And also it's nice to
640
00:32:13.450 --> 00:32:16.140
leave that wish list because then they have one less
641
00:32:16.200 --> 00:32:19.900
thing to decide. But also, like I said, it could
642
00:32:19.920 --> 00:32:21.480
be temporary. It could be like if you have a
643
00:32:21.500 --> 00:32:25.349
stroke or if you have... You could survive the pain crash,
644
00:32:25.369 --> 00:32:26.970
but you couldn't manage the project.
645
00:32:26.990 --> 00:32:28.890
Well, and now you get into a situation like an
646
00:32:28.970 --> 00:32:32.809
executor is a legal construct that has specific rights. Yeah.
647
00:32:32.930 --> 00:32:36.089
But you have to be dead for that to activate.
648
00:32:36.150 --> 00:32:39.299
So what is the sort of living will but incapacitated
649
00:32:39.619 --> 00:32:40.519
model of that?
650
00:32:40.880 --> 00:32:44.269
And that's something I would like more... I think if
651
00:32:44.289 --> 00:32:46.670
we're a young industry, I would like more services to
652
00:32:46.730 --> 00:32:51.200
have this mindset. Yeah, because I know like GitHub added
653
00:32:51.220 --> 00:32:54.140
the successor model where you can actually add someone as
654
00:32:54.160 --> 00:32:58.380
a successor to you. And if you're become encapsulated, they
655
00:32:58.400 --> 00:33:01.519
could request access and get rid of the delay. But
656
00:33:01.539 --> 00:33:04.099
that's a big base for web, because that is everything.
657
00:33:04.119 --> 00:33:07.369
What they can do, essentially, then they can archive and
658
00:33:07.410 --> 00:33:10.890
transfer post stories that are public. They don't have login access,
659
00:33:10.930 --> 00:33:13.529
but they can still take over things that matter for others.
660
00:33:14.210 --> 00:33:16.430
And I know GitLab has the same, but then you
661
00:33:16.450 --> 00:33:20.039
get full, but others don't. So that's why If you
662
00:33:20.059 --> 00:33:22.920
have a bigger product, it's better to have a GitHub organization.
663
00:33:22.940 --> 00:33:26.279
Because then you can have multiple maintainers and owners and
664
00:33:26.299 --> 00:33:26.920
things like that.
665
00:33:27.059 --> 00:33:30.049
So things like that can be good to have.
666
00:33:30.190 --> 00:33:35.029
Yeah. I'm still digging through each of these different phases.
667
00:33:35.869 --> 00:33:37.950
I don't know that I want... I have been an executor.
668
00:33:38.529 --> 00:33:41.759
I did not enjoy the process. But being the technical
669
00:33:41.819 --> 00:33:45.859
advisor to the executor... Like, I generally recommend these days
670
00:33:45.880 --> 00:33:47.160
when people ask me, like, you should be my executor.
671
00:33:47.180 --> 00:33:49.680
It's like, you should probably have a lawyer. Preferably one
672
00:33:49.700 --> 00:33:53.119
you trust. But for us in the tech industry, being
673
00:33:53.180 --> 00:33:56.680
a friend to that person and being available to that
674
00:33:56.740 --> 00:33:59.490
executor who's probably not that technical to deal with the
675
00:33:59.529 --> 00:34:01.609
technical side of stuff is very useful.
676
00:34:01.910 --> 00:34:05.009
Yeah. And I think the whole thing there, I think it's...
677
00:34:06.589 --> 00:34:09.380
That's important to just give the mandate to the right
678
00:34:09.420 --> 00:34:13.730
person that can assist. And there are things you need to.
679
00:34:15.400 --> 00:34:18.300
To see like the whole thing when you have foresight
680
00:34:18.320 --> 00:34:22.440
to give someone attorney for you it's I think that's
681
00:34:23.000 --> 00:34:27.550
important and it's something and and even like they are
682
00:34:27.590 --> 00:34:30.690
getting better I like when I'm a grandmother passed away
683
00:34:30.730 --> 00:34:32.329
we went for funeral home and they know about things
684
00:34:32.389 --> 00:34:35.929
like Instagram and Facebook and closing those accounts down sure
685
00:34:35.949 --> 00:34:39.639
yeah so they're not it's getting better The social.
686
00:34:39.400 --> 00:34:41.260
Media side is one thing, but I get what you're
687
00:34:41.300 --> 00:34:43.679
talking about, which is like, I didn't know about GitHub Successor.
688
00:34:43.699 --> 00:34:44.920
I'll put that in the show links.
689
00:34:44.940 --> 00:34:46.380
That's a great tool.
690
00:34:47.199 --> 00:34:50.829
I think that's something I would wish all services had.
691
00:34:51.070 --> 00:34:51.889
Sure.
692
00:34:52.769 --> 00:34:58.269
Because it matters like my Netflix account Only my family
693
00:34:58.309 --> 00:35:01.000
wants that, but my GitHub account is a different thing
694
00:35:01.300 --> 00:35:04.320
where I actually had a peer of mine and I
695
00:35:04.320 --> 00:35:07.719
was the successor of him just because we are the
696
00:35:07.800 --> 00:35:09.699
ones that have the technical knowledge and know what to
697
00:35:09.739 --> 00:35:14.389
do with things and how to maintain it. And that's
698
00:35:14.429 --> 00:35:16.230
not a load that we want to place on a
699
00:35:16.269 --> 00:35:17.210
non-technical person.
700
00:35:17.369 --> 00:35:17.530
No.
701
00:35:18.000 --> 00:35:21.000
And in general, that's my thinking for all of this
702
00:35:21.039 --> 00:35:23.260
stuff is I do not want to burden my loved
703
00:35:23.380 --> 00:35:26.880
ones with this any more than is absolutely necessary. How
704
00:35:26.920 --> 00:35:28.480
do you make it as simple as possible?
705
00:35:28.679 --> 00:35:32.139
Yeah, but also some good advice like with password managers
706
00:35:32.179 --> 00:35:34.619
where you can have, many have where you can actually
707
00:35:34.659 --> 00:35:38.550
have a physical printed document which is still like the
708
00:35:38.639 --> 00:35:40.280
emergency reset thing.
709
00:35:40.780 --> 00:35:43.460
Or you can have like a USB drive.
710
00:35:43.500 --> 00:35:46.980
Setting up a shared password family password manager has been
711
00:35:47.019 --> 00:35:48.949
a big thing. Not only just because that we both
712
00:35:48.989 --> 00:35:51.110
have access to the home account. So if I was gone,
713
00:35:51.530 --> 00:35:54.369
she could just... has access to accounts anyway, but that
714
00:35:54.389 --> 00:35:56.719
you have a whole mechanism for, yeah, you can take
715
00:35:56.780 --> 00:35:58.880
over my account. Here's the successor system for that.
716
00:35:59.300 --> 00:36:01.559
And that's good to know. And the whole phone thing
717
00:36:01.579 --> 00:36:03.739
you mentioned is valid because phones die.
718
00:36:03.800 --> 00:36:04.260
They do.
719
00:36:04.340 --> 00:36:05.179
It doesn't have to be.
720
00:36:05.380 --> 00:36:08.050
You have a limited amount of time to figure this out. Yeah.
721
00:36:08.239 --> 00:36:10.650
Yeah. But also for yourself, like it could be that
722
00:36:10.690 --> 00:36:12.570
you drop your phone in the ocean and how do
723
00:36:12.610 --> 00:36:15.730
I get things? So it's, it's regardless, a good practice
724
00:36:15.750 --> 00:36:16.010
to have.
725
00:36:16.170 --> 00:36:19.690
Yeah, let's face it. If you're lost in an aircraft disaster,
726
00:36:20.090 --> 00:36:21.070
your phone was probably with you.
727
00:36:21.170 --> 00:36:24.300
Yeah, your computer's with you, your phone's with you. Like,
728
00:36:24.360 --> 00:36:24.780
it could be.
729
00:36:25.320 --> 00:36:25.659
Yeah.
730
00:36:26.719 --> 00:36:29.780
And that's why sometimes, like, it's good to have that peer-to-peer,
731
00:36:29.820 --> 00:36:34.320
to have that other technical friend that hopefully isn't on
732
00:36:34.340 --> 00:36:37.800
the same plane. But also to spread your risks, to have, like, see,
733
00:36:38.139 --> 00:36:41.449
for GitHub, it's this guy. For my retrocomputing, it's this guy.
734
00:36:41.469 --> 00:36:43.750
Or girl, whatever.
735
00:36:43.889 --> 00:36:48.210
And I think that could be then it's much easier
736
00:36:48.250 --> 00:36:52.659
for the executor to just go through things and like
737
00:36:52.699 --> 00:36:53.110
a checklist.
738
00:36:53.139 --> 00:36:55.440
Be able to work through that. I just pulled the
739
00:36:55.500 --> 00:36:59.400
notes for Bitwarden's emergency access resources. So now you've got
740
00:36:59.440 --> 00:37:01.760
me searching around the internet for what are the tools
741
00:37:01.840 --> 00:37:04.119
I use and what did they do in these circumstances?
742
00:37:05.539 --> 00:37:08.500
I don't think anybody looks at this and probably you should.
743
00:37:08.679 --> 00:37:12.210
And also it's complicated because we have something in Sweden
744
00:37:12.230 --> 00:37:14.610
where you can do like future power of attorney to
745
00:37:15.110 --> 00:37:19.070
delegate that to someone else. And that's something to notice
746
00:37:19.389 --> 00:37:21.469
when you work in open source is that the legal
747
00:37:21.510 --> 00:37:24.570
things can be different in different countries. So that's also
748
00:37:24.610 --> 00:37:29.840
something to have that in mind where there can be
749
00:37:29.880 --> 00:37:33.079
gaps or there could be where you notice like you
750
00:37:33.119 --> 00:37:35.119
can't just grant access or whatever.
751
00:37:36.000 --> 00:37:36.480
So that's.
752
00:37:38.070 --> 00:37:40.869
And sometimes, like I know where I handed over a domain,
753
00:37:40.969 --> 00:37:43.179
that needed to be notarized for some.
754
00:37:44.639 --> 00:37:45.360
Top-level things.
755
00:37:46.219 --> 00:37:47.679
So it could be different for different domains.
756
00:37:47.699 --> 00:37:50.719
Yeah, right. It's because if you don't do this right,
757
00:37:50.880 --> 00:37:53.000
now you're actually a thief, right? You're hacking.
758
00:37:54.000 --> 00:37:54.280
Yeah.
759
00:37:54.380 --> 00:37:57.070
So it's an interesting mix of things. And one of
760
00:37:57.110 --> 00:37:58.510
the reasons I jumped on this immediately when we were
761
00:37:58.530 --> 00:38:01.369
talking about it, Mattias, is I wanted the Europeans' perspective, too,
762
00:38:01.449 --> 00:38:04.210
because I think the rules are different in different countries,
763
00:38:04.230 --> 00:38:05.190
so how you approach these things.
764
00:38:05.769 --> 00:38:08.449
And that's also like Europeans are like, it could be
765
00:38:08.510 --> 00:38:12.630
different because there's different history in countries. In Sweden, we
766
00:38:12.670 --> 00:38:16.630
have very good registries of all people for, for, for.
767
00:38:18.190 --> 00:38:21.070
reasons that happen in the 40s there or after the
768
00:38:21.130 --> 00:38:23.769
40s they don't have that in germany sure that's a
769
00:38:23.809 --> 00:38:26.690
distributed system yeah with their.
770
00:38:26.769 --> 00:38:29.710
Legend well and anything in the eastern block you know
771
00:38:29.750 --> 00:38:31.969
like there's a lot of data problems when you get
772
00:38:32.010 --> 00:38:33.489
to places like poland.
773
00:38:33.230 --> 00:38:35.789
Yeah and then there are things like we're still um
774
00:38:36.309 --> 00:38:38.929
you need to be in person to do things um
775
00:38:39.389 --> 00:38:42.030
and so it's And that's sometimes like if you have
776
00:38:42.050 --> 00:38:44.349
a password manager, it's often much more easier because then
777
00:38:44.369 --> 00:38:45.429
you can do it yourself.
778
00:38:45.489 --> 00:38:45.670
Right.
779
00:38:45.730 --> 00:38:46.050
Fix it.
780
00:38:46.070 --> 00:38:49.369
Yeah. Or because Daniel, like you said, the email problem. Well,
781
00:38:49.409 --> 00:38:53.769
then you can have your someone else be the reset.
782
00:38:53.809 --> 00:38:56.050
But what I like when you have a good support
783
00:38:56.090 --> 00:38:58.530
for it, like would have with the success of something
784
00:38:58.699 --> 00:39:02.059
in in GitHub that someone requiring access and I can
785
00:39:02.139 --> 00:39:05.219
cancel it if it's not valid. But if I'm if
786
00:39:05.280 --> 00:39:09.889
I'm not there, the request won't be canceled. right so
787
00:39:09.909 --> 00:39:12.380
that's the thing where i like it's the or like
788
00:39:12.400 --> 00:39:16.179
the digital services should like if you design an erp
789
00:39:16.239 --> 00:39:17.960
system or something it would be nice if you add
790
00:39:17.980 --> 00:39:20.380
that for a customer support to have something to be
791
00:39:20.420 --> 00:39:24.269
able to request the family part like because there are
792
00:39:24.309 --> 00:39:27.510
services now that are been like even online for 20
793
00:39:27.510 --> 00:39:29.889
years so there's a lot of data and things that
794
00:39:30.050 --> 00:39:34.630
could be like photos or whatever that are and for
795
00:39:34.670 --> 00:39:37.449
good reasons there are privacy rules and privacy laws so
796
00:39:37.489 --> 00:39:40.090
that's why it's important to do this before it happens
797
00:39:40.150 --> 00:39:42.760
because it can be really hard to tell apple to
798
00:39:42.800 --> 00:39:46.840
do something if they don't want to or microsoft or
799
00:39:46.900 --> 00:39:50.719
one of the big players it's um And that's why
800
00:39:50.820 --> 00:39:53.719
I think it's important to have some forethought in there.
801
00:39:54.039 --> 00:39:57.860
Yeah, for sure. Yeah, I think the challenge here is
802
00:39:58.340 --> 00:40:00.860
just I feel like I just need to spend a day.
803
00:40:01.239 --> 00:40:04.320
Look at the landscape of what I've got and think
804
00:40:04.420 --> 00:40:07.039
through what that looks like. No different than updating a will.
805
00:40:08.099 --> 00:40:10.320
And wills need updating. There's no two ways about it.
806
00:40:10.320 --> 00:40:10.739
Yeah, they do.
807
00:40:10.900 --> 00:40:13.179
So what I did, I did a set of markdown
808
00:40:13.199 --> 00:40:15.159
documents in a private Git repository.
809
00:40:15.500 --> 00:40:16.199
You're such a geek.
810
00:40:18.099 --> 00:40:21.630
So what I mean, it's kind of easy because then
811
00:40:21.650 --> 00:40:24.230
I did my first step, like an inventory, a single
812
00:40:24.269 --> 00:40:28.469
document with the index of things. I did set up
813
00:40:28.590 --> 00:40:34.130
emergency access for password managers. So the one contact thing.
814
00:40:34.210 --> 00:40:36.320
An LLM would be very good at this with you.
815
00:40:36.340 --> 00:40:39.219
This would be a fine tool to chat with to
816
00:40:39.260 --> 00:40:39.920
go through this.
817
00:40:40.000 --> 00:40:42.539
Yeah, because I can just pour down my thoughts into it.
818
00:40:42.539 --> 00:40:46.519
And I'm also like, One good thing, like many services
819
00:40:46.559 --> 00:40:48.760
have actually, like you can print a Recover Kit or
820
00:40:48.800 --> 00:40:52.119
a physical thing or a USB drive. Things like Windows,
821
00:40:52.159 --> 00:40:54.639
you can have a BitLocker on a USB key and
822
00:40:54.659 --> 00:40:58.650
have that in your safe. But also like name successors
823
00:40:58.690 --> 00:41:01.590
in services that support it, like GitHub and GitLab support that.
824
00:41:03.110 --> 00:41:05.510
And have that note that I can easily go through.
825
00:41:06.829 --> 00:41:08.730
If my success becomes my enemy, I can just go
826
00:41:08.750 --> 00:41:11.650
through that document and see who I have given access to.
827
00:41:12.510 --> 00:41:15.000
But also things like it could be good to have
828
00:41:16.039 --> 00:41:18.760
enabled things like, so you have auto-renew, so you buy
829
00:41:18.820 --> 00:41:22.590
time for things like domains and things. So it's not
830
00:41:23.289 --> 00:41:26.349
an immediate threat for some things. Like if something's really important,
831
00:41:28.150 --> 00:41:31.010
and that's also like, you can go the organization route
832
00:41:31.050 --> 00:41:33.010
because it's much easier if it's a company or a
833
00:41:33.110 --> 00:41:37.030
GitHub organization or things like that, because that is an
834
00:41:37.420 --> 00:41:39.800
organization people can take over, even if you're not there.
835
00:41:41.110 --> 00:41:43.289
and also like pick a date at least once a
836
00:41:43.289 --> 00:41:46.110
year when you go through this list so that was
837
00:41:46.130 --> 00:41:48.030
like the thing that i set for myself.
838
00:41:48.269 --> 00:41:50.750
And yeah totally this is a i i want to
839
00:41:50.809 --> 00:41:54.050
put a prompt aside that i run like i said
840
00:41:54.179 --> 00:41:56.599
once a year that goes through all of this stuff
841
00:41:57.119 --> 00:42:00.079
and because those tools have access to what i have
842
00:42:00.099 --> 00:42:01.760
been doing will probably even help build the list of
843
00:42:01.820 --> 00:42:03.719
this is where you're out of date right.
844
00:42:03.639 --> 00:42:06.460
Yeah and i think the goal isn't to preserve everything
845
00:42:06.519 --> 00:42:10.730
it's It's make like a shutdown survival for the people
846
00:42:10.769 --> 00:42:12.269
you leave behind.
847
00:42:12.650 --> 00:42:15.250
And a path to what happens first.
848
00:42:15.389 --> 00:42:18.610
Yeah, and you don't want your family to inherit infrastructure.
849
00:42:18.639 --> 00:42:23.539
I think they inherit decisions about it, but they shouldn't
850
00:42:23.599 --> 00:42:27.900
like inherit your Azure web app or something like that.
851
00:42:28.340 --> 00:42:30.980
Getting back to this whole AI harness thing with memory,
852
00:42:31.119 --> 00:42:35.219
I mean, I know that my wife uses Windows Copilot
853
00:42:36.059 --> 00:42:39.139
to help her with her business stuff. And I would,
854
00:42:39.440 --> 00:42:41.619
you know, if something were to happen to her, I
855
00:42:41.639 --> 00:42:44.199
feel like I could just walk up to Copilot and say, okay,
856
00:42:44.239 --> 00:42:47.409
here's what happened. Where do I start? You know, that
857
00:42:47.449 --> 00:42:48.869
could help me with that process.
858
00:42:49.269 --> 00:42:52.250
But actually, that kind of state is a problem because
859
00:42:52.329 --> 00:42:54.170
with the cloud services, you don't know where that lives.
860
00:42:54.230 --> 00:42:57.570
There's data sometimes where, and that you can do, and
861
00:42:57.730 --> 00:43:00.570
that's kind of thanks to ERP, you can actually export
862
00:43:00.630 --> 00:43:03.760
data from services. So that could be one way to
863
00:43:03.820 --> 00:43:10.159
have backup. But think like it's uh it's doing like
864
00:43:10.480 --> 00:43:13.739
also doing things like what am i like you said
865
00:43:13.760 --> 00:43:16.159
your wife told you what you're not using and i
866
00:43:16.179 --> 00:43:17.489
think that's kind of when you do this kind of
867
00:43:17.510 --> 00:43:20.389
inventory you can see like does this domain that is
868
00:43:20.429 --> 00:43:24.789
for like discount sharkages.com that you sell like do i
869
00:43:24.809 --> 00:43:28.219
need it anymore or can i ask not renew it
870
00:43:28.599 --> 00:43:30.940
uh and that's also kind of nice because it's like
871
00:43:31.809 --> 00:43:34.289
the less technical depth you have, because that's actually what
872
00:43:34.309 --> 00:43:37.789
you have. The less you have there, the lighter you are,
873
00:43:37.829 --> 00:43:41.159
I think, in a way. And sometimes I'm the kind
874
00:43:41.179 --> 00:43:42.760
of person, if I just put it in a document,
875
00:43:42.800 --> 00:43:46.000
I can pop my stack and I feel lighter without
876
00:43:46.059 --> 00:43:49.820
having to dwell about it and think too much of it.
877
00:43:50.400 --> 00:43:55.429
And I mean, it's... Yeah, and I think it's okay
878
00:43:55.449 --> 00:43:58.909
for it to take time initially, like, because some things are,
879
00:43:59.510 --> 00:44:03.079
it's sometimes it's like killing your darlings is hard. Like,
880
00:44:04.159 --> 00:44:07.400
like I had the sometimes like, why do I have
881
00:44:07.420 --> 00:44:12.340
this broken, like the oldest Nokia phones from or whatever.
882
00:44:13.639 --> 00:44:15.690
Oh, no, it's your Nokia phone shrine.
883
00:44:15.969 --> 00:44:17.010
Yeah, it's like, I.
884
00:44:16.949 --> 00:44:17.570
Still have mine.
885
00:44:18.099 --> 00:44:22.170
Yeah, and it's like, it's hard to, and that's why
886
00:44:22.230 --> 00:44:24.789
I like to have to, my wife is really good.
887
00:44:24.849 --> 00:44:26.710
You know what? It's going to be easier for your
888
00:44:26.750 --> 00:44:27.250
loved ones.
889
00:44:27.650 --> 00:44:30.369
Yeah, but also like it's, it's sometimes good because my
890
00:44:30.409 --> 00:44:32.320
wife is really good at throwing things away. And that's
891
00:44:32.340 --> 00:44:35.420
kind of like, she has to go to iterations with this.
892
00:44:35.480 --> 00:44:37.380
I throw a little away each time she tells me
893
00:44:37.400 --> 00:44:38.159
to throw something away.
894
00:44:39.179 --> 00:44:41.340
Well, she, I suspect she's even better at throwing away
895
00:44:41.619 --> 00:44:42.179
your things.
896
00:44:42.219 --> 00:44:42.980
She's excellent at it.
897
00:44:43.139 --> 00:44:43.619
Well, yeah.
898
00:44:44.349 --> 00:44:47.630
Oh, you're, you're totally speaking my language, Richard. I wouldn't
899
00:44:47.690 --> 00:44:50.690
call myself a pack rat, but I did come from
900
00:44:50.750 --> 00:44:53.110
a pack rat, right? My mother was a pack rat.
901
00:44:53.610 --> 00:44:55.900
And so I can see, I have tendencies towards that,
902
00:44:55.920 --> 00:44:58.000
but my wife keeps me on the state and narrow. She's,
903
00:44:58.440 --> 00:45:01.519
and her mother is a chucker. Like get rid of this,
904
00:45:01.619 --> 00:45:03.460
get rid of that, get rid of this. Her house
905
00:45:03.500 --> 00:45:06.179
is always like pristine and clean and there's nothing that
906
00:45:06.219 --> 00:45:09.050
doesn't belong there. You know, my studio, I look around,
907
00:45:09.090 --> 00:45:11.150
I got a couple of boxes over here. I got
908
00:45:11.170 --> 00:45:14.750
some stuff over there. You know, and so if it
909
00:45:14.769 --> 00:45:17.909
weren't for my wife, it would be piled probably higher
910
00:45:17.949 --> 00:45:18.369
and deeper.
911
00:45:19.429 --> 00:45:19.809
Yeah.
912
00:45:19.889 --> 00:45:23.530
I mean, like, I think for me, like it, some
913
00:45:23.550 --> 00:45:25.289
things took me a couple of years to do because
914
00:45:25.309 --> 00:45:27.889
you both like you have to procrastinate a bit and
915
00:45:27.929 --> 00:45:30.570
you think like, and, and it's not the optimal system,
916
00:45:30.590 --> 00:45:33.230
but I think it's good to have at least have
917
00:45:33.250 --> 00:45:36.949
the mindset to go for it. And I think it's, um,
918
00:45:37.590 --> 00:45:41.840
it's an even, Sometimes it's so easy. Now I'm more
919
00:45:41.880 --> 00:45:44.119
like cloud storage, but I know when hard drives were.
920
00:45:44.409 --> 00:45:45.909
When I ran out, I just bought a new hard
921
00:45:45.929 --> 00:45:48.090
drive and moved that one aside. So I have a
922
00:45:48.690 --> 00:45:53.110
box of hard drives and things like that. So it's
923
00:45:53.289 --> 00:45:56.889
very easy. It was cheaper because Moore's Law to become
924
00:45:56.909 --> 00:45:57.530
a hoarder.
925
00:45:58.849 --> 00:45:59.010
Yeah.
926
00:46:00.409 --> 00:46:02.150
But I think it's kind of nice.
927
00:46:02.230 --> 00:46:06.090
Well, yeah. It's only recently we've been able to hoard petabytes.
928
00:46:06.630 --> 00:46:06.829
Yeah.
929
00:46:06.869 --> 00:46:09.550
Yeah. Well, I did the thing where I moved all
930
00:46:09.590 --> 00:46:13.630
my hard drive data to my NAS storage and then
931
00:46:14.090 --> 00:46:16.670
drilled a hole in every hard drive and then brought
932
00:46:16.730 --> 00:46:19.420
it to Staples because they can recycle them.
933
00:46:19.699 --> 00:46:20.440
Shred them.
934
00:46:20.480 --> 00:46:20.820
Yeah.
935
00:46:21.460 --> 00:46:23.300
And that's kind of good because then it's final because
936
00:46:23.360 --> 00:46:25.000
it's so easy when you have something stored that you
937
00:46:25.019 --> 00:46:29.059
don't know even if it works. And it probably doesn't.
938
00:46:29.219 --> 00:46:31.199
And that's kind of a problem. Like when you do retrocomputing,
939
00:46:31.219 --> 00:46:33.900
you learn it the hard way that disks don't last forever.
940
00:46:34.019 --> 00:46:38.530
And they are, you are storing on rust. What do
941
00:46:38.550 --> 00:46:39.269
you think? Yeah.
942
00:46:39.710 --> 00:46:43.829
And so I think sometimes it's, we were really good
943
00:46:43.949 --> 00:46:46.869
when we did like tape backups and things and dating myself,
944
00:46:46.929 --> 00:46:49.159
like we knew that things would break. So we moved
945
00:46:49.199 --> 00:46:52.179
things forward. So it's time we had a new backup
946
00:46:52.199 --> 00:46:54.460
strategy for this generation and we moved things forward.
947
00:46:54.480 --> 00:46:55.340
I think we need to do that.
948
00:46:55.380 --> 00:46:59.630
But, but, There was a time when I deliberately ordered
949
00:46:59.750 --> 00:47:03.909
backup tape drives left in a box for the prospect
950
00:47:03.929 --> 00:47:05.809
that that drive will fail and we'll need to migrate
951
00:47:05.849 --> 00:47:07.550
these tapes and we may not be able to buy
952
00:47:07.570 --> 00:47:08.650
the drive.
953
00:47:08.670 --> 00:47:11.780
It doesn't matter if you have that tape if you
954
00:47:11.800 --> 00:47:12.760
don't have a DAT reader.
955
00:47:12.980 --> 00:47:15.719
That's right. I probably can't get one now, so I
956
00:47:16.260 --> 00:47:19.079
keep a spare on the shelf. Then I learned to
957
00:47:19.119 --> 00:47:21.539
keep a spare on a shelf in a different building
958
00:47:21.579 --> 00:47:23.380
when the first building burned down.
959
00:47:24.739 --> 00:47:27.300
And that's why I learned something like I'm not a
960
00:47:27.320 --> 00:47:29.679
hoarder anymore because like the first death clean you did
961
00:47:29.699 --> 00:47:31.900
for someone else, you know, like, yeah, yeah. So much.
962
00:47:32.320 --> 00:47:35.460
And the conflicts that are a wake up call that
963
00:47:35.539 --> 00:47:38.000
are with inheritance and things like, like take care of
964
00:47:38.059 --> 00:47:41.340
that for your, for your loved ones before. And I
965
00:47:41.360 --> 00:47:44.010
think that's why I like to just, just do like
966
00:47:44.030 --> 00:47:48.230
inventory categorize who should have access and who's your successor.
967
00:47:48.650 --> 00:47:50.469
And if you have those kinds of things in, in checks,
968
00:47:50.530 --> 00:47:53.469
I think it's, it's good way. And it's kind of
969
00:47:53.550 --> 00:47:56.750
also like the whole thing with, open source is responsible
970
00:47:56.769 --> 00:48:00.449
because then there's a plan for things, how they will
971
00:48:00.550 --> 00:48:02.019
be maintained along the way for sure.
972
00:48:02.179 --> 00:48:03.820
Also, it's a good idea to do a little bit
973
00:48:03.860 --> 00:48:06.480
at a time, not sit down and think, okay, today
974
00:48:06.500 --> 00:48:09.019
I'm going to solve this problem. No, it's going to
975
00:48:09.039 --> 00:48:12.469
take many days and, you know, hours, an hour here,
976
00:48:12.510 --> 00:48:15.849
an hour there is such a better, way to go
977
00:48:15.869 --> 00:48:16.199
about it.
978
00:48:16.269 --> 00:48:18.190
Well, and you're going to have some hard ones you don't,
979
00:48:18.230 --> 00:48:20.670
you can't immediately decide either. So it's probably good to
980
00:48:20.690 --> 00:48:21.829
let that sit in the back of the head and
981
00:48:21.949 --> 00:48:22.849
digest for a while. Yeah.
982
00:48:23.300 --> 00:48:24.679
But that's the same thing you do in storage. So
983
00:48:24.699 --> 00:48:26.739
like you put a date on a box and if
984
00:48:26.760 --> 00:48:29.280
you haven't opened it by, and they can do the
985
00:48:29.320 --> 00:48:31.480
same digitally, like put this in a date folder.
986
00:48:31.500 --> 00:48:31.880
Yeah.
987
00:48:32.059 --> 00:48:35.420
And if you haven't opened that archive, you should probably
988
00:48:35.460 --> 00:48:35.820
throw it away.
989
00:48:35.840 --> 00:48:35.960
Right.
990
00:48:36.920 --> 00:48:39.570
Who is the guy that Hanselman interviewed? It was getting
991
00:48:39.610 --> 00:48:42.309
stuff done or it might have been getting done or whatever,
992
00:48:42.409 --> 00:48:45.079
but he had something like that. He had really good.
993
00:48:45.630 --> 00:48:49.010
strategies for not being overwhelmed and you know putting post-it
994
00:48:49.050 --> 00:48:50.570
notes on things and you know.
995
00:48:50.489 --> 00:48:52.820
Yeah well i think like the most like especially i
996
00:48:52.840 --> 00:48:55.059
don't know about others but i'm not good at multitasking
997
00:48:55.099 --> 00:48:57.400
i'm good at time slicing so i think that's that's
998
00:48:57.420 --> 00:48:59.840
how you like if you should divide things in things
999
00:48:59.860 --> 00:49:01.739
you can keep in your content like you said with
1000
00:49:01.860 --> 00:49:04.300
ai if you can keep it in your memory then you.
1001
00:49:04.260 --> 00:49:05.820
Will be efficient in it and you will do it
1002
00:49:06.019 --> 00:49:06.840
yeah and.
1003
00:49:06.980 --> 00:49:10.079
If it takes more than More than a Pomodoro time
1004
00:49:10.119 --> 00:49:12.659
to think about it, then it will be lost.
1005
00:49:13.219 --> 00:49:18.329
Yeah, that's right. So I feel like we're coming to
1006
00:49:18.349 --> 00:49:19.909
the end of this conversation. Is there more?
1007
00:49:20.019 --> 00:49:23.659
I think we also want to start a discussion. I
1008
00:49:23.699 --> 00:49:26.880
think we should reflect. And if there are good things,
1009
00:49:26.900 --> 00:49:30.820
because I'm happy to be educated here or be wrong
1010
00:49:30.900 --> 00:49:33.280
or things. I think here we as a community can
1011
00:49:33.320 --> 00:49:36.010
help each other. I think if you have some kind
1012
00:49:36.050 --> 00:49:38.659
of service, I think it's good to think about that
1013
00:49:39.090 --> 00:49:41.900
what's because you will you learn when you're like I
1014
00:49:41.920 --> 00:49:44.739
know I have had e-commerce services like you learn when
1015
00:49:44.760 --> 00:49:47.480
you get the first support ticket about something then when
1016
00:49:47.500 --> 00:49:50.300
you're like you learned about the missing features and usually
1017
00:49:50.320 --> 00:49:53.179
they're not in the beginning but I think this is
1018
00:49:53.219 --> 00:49:58.599
a feature you should think about how to handle one
1019
00:49:58.840 --> 00:50:02.519
because it can be as you say things happen in
1020
00:50:02.559 --> 00:50:06.079
life or things like I had a friend that had
1021
00:50:06.119 --> 00:50:11.289
a stroke which means He's still here, but he's not.
1022
00:50:11.949 --> 00:50:15.889
So there could be different reasons that you have to
1023
00:50:15.949 --> 00:50:21.269
be prepared. But also, I think it's a really thoughtful
1024
00:50:21.349 --> 00:50:23.869
and lovable thing you can do for everyone else.
1025
00:50:23.889 --> 00:50:24.610
I agree, sure.
1026
00:50:25.030 --> 00:50:26.989
And it's a responsible thing. And I think there's something
1027
00:50:27.010 --> 00:50:31.159
that the 20-year-old me wouldn't think of. But if you've
1028
00:50:31.219 --> 00:50:33.920
been through a couple of things in life, you notice
1029
00:50:33.940 --> 00:50:35.199
that it's important.
1030
00:50:35.389 --> 00:50:38.909
Well, when you're 20, you're omnipotent, right?
1031
00:50:38.969 --> 00:50:40.030
Yeah, you're like.
1032
00:50:40.110 --> 00:50:40.730
You're immortal.
1033
00:50:40.889 --> 00:50:45.070
Yeah. And suddenly, like, when you're around 50, you start
1034
00:50:45.090 --> 00:50:49.130
to get afraid of heights and things like that.
1035
00:50:49.210 --> 00:50:52.139
Hey, fall, take a long time to recover from now.
1036
00:50:52.159 --> 00:50:56.179
Nothing's easy. Yeah, that's... We've all been through cleaning up
1037
00:50:56.320 --> 00:50:58.219
after a loved one's passing.
1038
00:50:58.300 --> 00:50:58.440
Yeah.
1039
00:50:58.760 --> 00:51:00.699
And now we're much more aware looking around our own
1040
00:51:00.739 --> 00:51:04.059
rooms going, hmm. And some of those rooms are digital.
1041
00:51:04.219 --> 00:51:07.070
Yeah. And I mean... And that's why I'm like, yeah,
1042
00:51:07.090 --> 00:51:09.750
I think it's important. It's like, who's this important for?
1043
00:51:09.829 --> 00:51:12.289
If it's just me, then it's not important. It's important
1044
00:51:12.469 --> 00:51:14.730
as long as I care about it.
1045
00:51:14.750 --> 00:51:15.769
Yeah, totally.
1046
00:51:15.889 --> 00:51:20.389
But also like to reiterate and to talk to make
1047
00:51:20.429 --> 00:51:26.639
things easier because I think it's a technical debt that
1048
00:51:26.699 --> 00:51:29.019
will burden you or someone else. And I think it's
1049
00:51:29.159 --> 00:51:30.619
better to take it in pieces.
1050
00:51:30.699 --> 00:51:31.360
Nice.
1051
00:51:31.420 --> 00:51:35.019
Yeah. Don't propagate your technical debt on others unnecessarily.
1052
00:51:35.059 --> 00:51:35.739
Yeah, so true.
1053
00:51:36.019 --> 00:51:38.050
Yeah, I love it. Thanks, Matthias. All right.
1054
00:51:38.469 --> 00:51:42.599
Yeah, thank you, Matthias. What good stuff. And thanks for listening,
1055
00:51:42.639 --> 00:51:44.570
and we'll talk to you next week on.
1056
00:51:44.539 --> 00:51:44.989
NET Rocks.
1057
00:52:05.469 --> 00:52:08.170
NET Rocks is brought to you by Franklin's Net and
1058
00:52:08.230 --> 00:52:12.929
produced by Plop Studios, a full-service audio, video, and post-production
1059
00:52:12.989 --> 00:52:16.969
facility located physically in New London, Connecticut, and, of course,
1060
00:52:17.050 --> 00:52:20.650
in the cloud. Online at pwop.com.
1061
00:52:21.380 --> 00:52:29.280
Visit our website at dotnetrocks.com for RSS feeds, downloads, mobile apps, comments,
1062
00:52:29.699 --> 00:52:32.179
and access to the full archives going back to show
1063
00:52:32.219 --> 00:52:36.530
number one, recorded in September 2002 and make sure you
1064
00:52:36.570 --> 00:52:39.760
check out our sponsors they keep us in business now
1065
00:52:39.800 --> 00:52:41.960
go write some code see you next time
1
00:00:01.169 --> 00:00:04.790
How'd you like to listen to. NET Rocks with no ads? Easy.
2
00:00:05.370 --> 00:00:08.779
Become a patron. For just $ 5 a month, you get
3
00:00:08.890 --> 00:00:11.720
access to a private RSS feed where all the shows
4
00:00:11.800 --> 00:00:14.900
have no ads. $ 20 a month will get you that
5
00:00:15.119 --> 00:00:18.839
and a special. NET Rocks patron mug. Sign up now
6
00:00:18.899 --> 00:00:37.659
at patreon.dotnetrocks.com. Hey, welcome back to. NET Rocks. I'm Carl Franklin.
7
00:00:37.840 --> 00:00:42.229
And I'm Richard Campbell. This is episode 2022. It was
8
00:00:42.229 --> 00:00:45.630
22 episodes ago we were in that bar at the
9
00:00:45.950 --> 00:00:46.750
MVP Summit.
10
00:00:46.850 --> 00:00:47.590
Yeah.
11
00:00:47.890 --> 00:00:50.869
At Party with Palermo doing episode 2000. Yeah.
12
00:00:51.009 --> 00:00:55.299
Yeah. And now we're four away from the end of history. Yeah.
13
00:00:58.369 --> 00:01:02.200
It was only four years ago, 2022. So let's recap
14
00:01:02.219 --> 00:01:03.729
some of the things that happened that year.
15
00:01:03.759 --> 00:01:04.450
It's pretty current.
16
00:01:04.730 --> 00:01:07.359
A little thing called Russia invading Ukraine.
17
00:01:07.560 --> 00:01:08.799
Yeah. That's a little thing.
18
00:01:08.859 --> 00:01:10.140
Maybe you've heard of that little event.
19
00:01:10.239 --> 00:01:12.299
Remember when it was supposed to be three days, so
20
00:01:12.340 --> 00:01:14.359
they didn't bother packing a lot of food and said
21
00:01:14.379 --> 00:01:18.819
they packed their dress uniforms for the march down the boulevard?
22
00:01:18.879 --> 00:01:19.400
Yeah.
23
00:01:19.620 --> 00:01:20.079
How'd that go?
24
00:01:20.219 --> 00:01:20.379
Yeah.
25
00:01:20.640 --> 00:01:23.599
In other happy news, the Supreme Court overturned Roe v.
26
00:01:23.640 --> 00:01:24.560
Wade. Oh, geez.
27
00:01:24.760 --> 00:01:29.409
Queen Elizabeth II died. After 70 years on the throne.
28
00:01:29.849 --> 00:01:33.909
Second longest reigning monarch in history, something like that. Yeah. Yeah.
29
00:01:34.109 --> 00:01:36.640
She, from what I understand, she had quite a sense
30
00:01:36.659 --> 00:01:37.040
of humor.
31
00:01:37.280 --> 00:01:37.739
Very funny.
32
00:01:38.000 --> 00:01:38.239
Yeah.
33
00:01:38.340 --> 00:01:42.599
Very funny. Classically, like there's stories like Balmoral Castle is
34
00:01:42.659 --> 00:01:45.239
up in Scotland and she loved to take a walk
35
00:01:45.299 --> 00:01:47.200
and she'd never dressed up or anything. And she usually
36
00:01:47.219 --> 00:01:49.780
said a guard with her. So it's like two elderly
37
00:01:49.819 --> 00:01:53.019
people walking along the field and they bump into some
38
00:01:53.120 --> 00:01:58.319
tourists and they ask him, you know, I've heard the
39
00:01:58.599 --> 00:02:01.030
Windsor's Castle is around here somewhere. Oh, yeah, it's around
40
00:02:01.069 --> 00:02:03.209
here somewhere. Is it that one over there?
41
00:02:03.250 --> 00:02:04.269
He goes, I don't know.
42
00:02:04.329 --> 00:02:06.650
The little old lady there is a bit strange. They go, oh,
43
00:02:06.969 --> 00:02:08.750
thanks for your help. You have a good time. And
44
00:02:08.770 --> 00:02:11.650
they walk away. Hilarious.
45
00:02:11.770 --> 00:02:17.860
Yep, she was. Inflation, cost of living, crisis, not good.
46
00:02:18.020 --> 00:02:20.139
Well, we're still coming out of COVID, right?
47
00:02:20.180 --> 00:02:24.500
Yeah, we're coming out of COVID. That's right. Finally began retreating.
48
00:02:25.319 --> 00:02:29.659
as a global crisis in most countries, except most countries,
49
00:02:29.699 --> 00:02:30.419
except for China.
50
00:02:30.469 --> 00:02:30.550
Yeah.
51
00:02:30.560 --> 00:02:31.590
China stayed locked down.
52
00:02:31.669 --> 00:02:34.509
I don't know if that was a good call, but yeah. Um,
53
00:02:34.990 --> 00:02:41.050
Uvalde school shooting, 19 children. Uh, terrible, uh, historic us
54
00:02:41.110 --> 00:02:41.969
midterm elections.
55
00:02:42.810 --> 00:02:43.000
Yeah.
56
00:02:43.219 --> 00:02:49.599
Iran erupts in protest after the death of Masha Amini. Yeah.
57
00:02:50.099 --> 00:02:51.340
It's all seems so long ago now.
58
00:02:51.759 --> 00:02:57.189
Yeah, it does. Um, Humanity made major progress against preventable diseases.
59
00:02:57.270 --> 00:03:00.870
There were some remarkable public health victories that received comparatively
60
00:03:00.949 --> 00:03:13.580
little attention. Malawi and Vanuatu eliminated trachoma as a public
61
00:03:13.599 --> 00:03:16.699
health problem, protecting future generations from one of the world's
62
00:03:16.800 --> 00:03:21.770
leading infectious causes of blindness. Guinea worm disease also moved
63
00:03:21.889 --> 00:03:23.780
tantalizingly close to eradication.
64
00:03:23.810 --> 00:03:26.379
You know, you know who, uh, that was, uh, that
65
00:03:26.400 --> 00:03:28.039
was Jimmy Carter's Jimmy Carter.
66
00:03:28.060 --> 00:03:28.539
Yeah.
67
00:03:28.560 --> 00:03:28.860
Yeah.
68
00:03:29.060 --> 00:03:33.219
Yeah. Jimmy Carter basically taught these African countries how to
69
00:03:33.719 --> 00:03:36.860
get rid of Guinea worm by, um, what they would
70
00:03:36.900 --> 00:03:39.919
do is the Guinea worms would bury, bury themselves in your,
71
00:03:40.319 --> 00:03:43.500
or come out of your body through your legs and stuff.
72
00:03:43.539 --> 00:03:46.090
It was very painful. And what they would do was,
73
00:03:46.449 --> 00:03:51.689
you know, to soothe themselves is wash their, wounds in
74
00:03:52.050 --> 00:03:54.569
a pond or a lake, the same lake that they
75
00:03:54.590 --> 00:03:57.129
drank out of. So it was this perpetual thing. So
76
00:03:57.169 --> 00:04:00.189
he basically taught them to use like nylons, you know,
77
00:04:00.349 --> 00:04:04.310
pantyhose or whatever to strain the water. And yeah, that
78
00:04:04.349 --> 00:04:05.379
was the beginning of the end.
79
00:04:05.439 --> 00:04:06.810
Water filters just break the loop.
80
00:04:06.849 --> 00:04:07.009
Yeah.
81
00:04:07.030 --> 00:04:07.789
Brilliant.
82
00:04:08.129 --> 00:04:12.699
At the COP 15 biodiversity conference in Montreal, more than
83
00:04:12.699 --> 00:04:15.740
190 countries agreed to a framework aimed at protecting at
84
00:04:15.759 --> 00:04:20.029
least 30% of the world's land, inland waters, and oceans
85
00:04:20.089 --> 00:04:26.699
by 2030. Meanwhile, conservation programs produced tangible successes. Nepal's wild
86
00:04:26.800 --> 00:04:31.259
tiger population had nearly tripled, and several species believed lost
87
00:04:31.319 --> 00:04:34.000
or extinct were rediscovered. So I wanted to put a
88
00:04:34.019 --> 00:04:36.699
little good news at the end of that list. And
89
00:04:36.740 --> 00:04:40.050
there's more stuff, but what's happening in tech and space
90
00:04:40.110 --> 00:04:40.399
in 2022, Richard?
91
00:04:42.009 --> 00:04:46.470
All right, let's start in February with Ingenuity Flight 19.
92
00:04:46.449 --> 00:04:48.959
So this is the little helicopter that they put on
93
00:04:49.000 --> 00:04:50.879
Mars with the Perseverance rover. They had a little bit
94
00:04:50.899 --> 00:04:53.639
of payload left over. So let's do this experiment. You
95
00:04:53.660 --> 00:04:57.079
can actually make something fly in this crazy low pressure,
96
00:04:57.139 --> 00:05:03.050
like one-tenth of 1% of the atmosphere. So very high-speed
97
00:05:03.149 --> 00:05:05.990
counter-rotating blades. It was only supposed to work for five flights,
98
00:05:06.050 --> 00:05:09.670
just to test. And now it's on flight 19. So
99
00:05:09.689 --> 00:05:13.730
it's no longer experimental. It's actually a useful device for
100
00:05:13.829 --> 00:05:18.829
the Perseverance rover. It's doing scouting. It's looking for terrain problems.
101
00:05:18.910 --> 00:05:23.250
It's looking for potential scientifically interesting locations. So a huge
102
00:05:23.470 --> 00:05:25.430
step forward. It really has opened the door to a
103
00:05:25.470 --> 00:05:30.220
whole other way to explore other planetary bodies. In April...
104
00:05:31.189 --> 00:05:34.600
We hit the 120 successful Falcon 9 flights, which breaks
105
00:05:34.639 --> 00:05:38.790
the all-time record for everything. So SpaceX is just stretching
106
00:05:38.829 --> 00:05:40.470
out the lead. And one of those flights in April
107
00:05:40.589 --> 00:05:44.209
is the Axiom 1 flight to the International Space Station.
108
00:05:44.240 --> 00:05:48.189
This is a fully... private mission to the space station
109
00:05:48.449 --> 00:05:52.110
funded through Axiom Space as one of their fundraising strategies
110
00:05:52.129 --> 00:05:55.060
for building their own space station using Crew Dragon. Of course,
111
00:05:55.079 --> 00:05:57.240
the International Space Station is not the only space station.
112
00:05:57.279 --> 00:06:00.560
It's also the Tiangong Space Station. It's getting regular crew rotations,
113
00:06:00.660 --> 00:06:03.759
one in June and one in November. In July, we
114
00:06:03.800 --> 00:06:07.819
get our first images from James Webb. So launched in 2021,
115
00:06:07.819 --> 00:06:10.769
takes about six months to do that crazy deployment where
116
00:06:10.810 --> 00:06:12.430
it has to unfold the heat shield and all that
117
00:06:12.449 --> 00:06:15.040
cool stuff, get out to the L2 point. And we
118
00:06:15.060 --> 00:06:18.939
start getting our initial images back. And, of course, it's astonishing.
119
00:06:19.019 --> 00:06:22.459
Astonishing. And it's still astonishing. I think I saw an
120
00:06:22.519 --> 00:06:26.879
article where it just found an incredibly habitable planet.
121
00:06:27.000 --> 00:06:27.860
Possibly, yeah.
122
00:06:28.180 --> 00:06:31.000
It's some 14 light years away. Yeah.
123
00:06:31.120 --> 00:06:33.019
And we're getting to the point now where we can
124
00:06:33.040 --> 00:06:34.730
start to measure atmospheres.
125
00:06:35.170 --> 00:06:35.439
Yeah.
126
00:06:35.990 --> 00:06:39.250
And so– This is not what Star Trek told us
127
00:06:39.269 --> 00:06:40.810
would happen. You don't have to fly to the solar
128
00:06:40.850 --> 00:06:42.910
system to evaluate. We're going to have a map of
129
00:06:42.930 --> 00:06:46.250
all these planets long before we can get there. In September,
130
00:06:46.350 --> 00:06:49.250
the DART mission, so this is the Double Asteroid Redirect Test,
131
00:06:50.209 --> 00:06:56.060
called that because it's actually a binary asteroid system. Didymos
132
00:06:56.329 --> 00:06:59.860
is this asteroid that flies relatively close to the Earth.
133
00:06:59.920 --> 00:07:02.680
It's not huge, 780 meters across. You don't want it
134
00:07:02.720 --> 00:07:05.779
in your backyard or anything, but It's not a tremendously
135
00:07:05.800 --> 00:07:09.480
large rock, but it has a moon, which is called Dimorphos,
136
00:07:09.500 --> 00:07:12.759
although most people refer to it as Diddy Moon. Much smaller,
137
00:07:12.819 --> 00:07:17.220
about 160 meters across, and it orbits the larger asteroid
138
00:07:17.240 --> 00:07:19.980
about every 12 hours, 11 hours, 55 minutes. And so
139
00:07:20.720 --> 00:07:23.240
they picked this target because A, they could get to it. B,
140
00:07:23.279 --> 00:07:25.300
they could measure it, like they can actually see it
141
00:07:25.379 --> 00:07:28.139
well enough to know what's going on. And so they
142
00:07:28.189 --> 00:07:31.430
have this little impactor, and their goal was to alter
143
00:07:31.470 --> 00:07:34.790
the orbit by 73 seconds. So hit the moon so
144
00:07:34.810 --> 00:07:40.490
that it actually slows down and will orbit closer to Didymos.
145
00:07:41.920 --> 00:07:44.230
They overshoot the mark by a lot. It turns out
146
00:07:44.310 --> 00:07:46.930
that the moon is a rubble pile. It's just a
147
00:07:46.990 --> 00:07:49.709
bunch of little rocks all sort of clumped together with
148
00:07:49.750 --> 00:07:52.970
a weak gravity. And so the impact creates this huge
149
00:07:53.170 --> 00:07:57.639
spray of rocks behind, like kilometers long, and actually reduces
150
00:07:57.699 --> 00:08:00.420
the orbit by 33 minutes. And so the first time
151
00:08:00.500 --> 00:08:03.500
that humanity has demonstrated its ability to actually modify the
152
00:08:03.540 --> 00:08:06.790
behavior of an asteroid through an impact. There's another mission
153
00:08:07.149 --> 00:08:10.389
today on the way back to that binary moon system
154
00:08:10.470 --> 00:08:12.170
to do additional measurements, which is cool.
155
00:08:12.189 --> 00:08:12.889
Yep, I remember that.
156
00:08:12.910 --> 00:08:15.670
And finally on the space side in November, Artemis 1.
157
00:08:15.670 --> 00:08:19.230
So this was the first flight of the Orion spacecraft
158
00:08:19.470 --> 00:08:21.779
all up, although it didn't actually have a life support
159
00:08:21.800 --> 00:08:23.360
system on board, which is good because there's no people
160
00:08:23.420 --> 00:08:26.399
on it. But it did fly out to the moon
161
00:08:26.560 --> 00:08:29.540
and do a bunch of maneuvers we would never try
162
00:08:29.600 --> 00:08:33.019
with humans. You know, we hadn't flown this kind of
163
00:08:33.059 --> 00:08:36.330
spacecraft out around the moon before. in decades and we
164
00:08:36.389 --> 00:08:39.659
know a lot more about gravitation and grav and gravitational
165
00:08:39.679 --> 00:08:42.039
maneuvering now and so they basically did a set of
166
00:08:42.100 --> 00:08:45.360
experiments while it was unmanned to just show that they
167
00:08:45.419 --> 00:08:48.580
could do some really cool orbit changes and really energy
168
00:08:48.639 --> 00:08:52.490
efficient maneuvering and it'll come back in late november quite
169
00:08:52.509 --> 00:08:56.649
successfully de-orbit and return to the earth intact although the
170
00:08:56.690 --> 00:08:59.250
heat shield will take damage that makes them unhappy and
171
00:08:59.269 --> 00:09:02.980
they'll have to do some redesigns for artemis ii All right.
172
00:09:03.000 --> 00:09:03.879
Ready for some computing?
173
00:09:04.279 --> 00:09:04.879
Yeah.
174
00:09:04.919 --> 00:09:09.919
I'm going to start in March with this AI paper.
175
00:09:10.039 --> 00:09:15.299
It's called the Training Compute Optimizing Large Language Models. It's
176
00:09:15.350 --> 00:09:18.970
actually a paper out of DeepMinds. This is Dennis Hasselbiss
177
00:09:19.029 --> 00:09:21.509
and a bunch of folks working together. It's generally referred
178
00:09:21.549 --> 00:09:25.269
to as the chinchilla result. But what it really talks about,
179
00:09:25.289 --> 00:09:28.090
they had the computing scaling laws paper back in 2020.
180
00:09:28.090 --> 00:09:31.659
And this is sort of a counter to that, which says, actually,
181
00:09:32.220 --> 00:09:36.730
Most of these models are under-trained, that they don't have
182
00:09:36.799 --> 00:09:39.990
enough data, and that just making the bigger isn't going
183
00:09:40.009 --> 00:09:42.289
to be effective. You actually have to change, go down
184
00:09:42.309 --> 00:09:45.909
to fewer parameters with more training data to get actually
185
00:09:45.990 --> 00:09:48.690
more optimal results. And that will affect a bunch of
186
00:09:48.730 --> 00:09:50.269
things going forward. Now, remember, it's 2022.
187
00:09:50.250 --> 00:09:51.330
Right, right.
188
00:09:51.450 --> 00:09:53.029
So the crazy doesn't happen until the end of the year.
189
00:09:53.090 --> 00:09:53.750
Yeah.
190
00:09:53.769 --> 00:09:55.190
But there is a bunch of other cool stuff that
191
00:09:55.230 --> 00:09:58.360
happens this year on the, in the machine learning space,
192
00:09:58.379 --> 00:09:59.879
because this is the year, like in April, we get
193
00:09:59.919 --> 00:10:00.740
Dolly too.
194
00:10:01.159 --> 00:10:01.360
Right.
195
00:10:01.399 --> 00:10:03.919
And we get mid journey in July and we get
196
00:10:03.940 --> 00:10:07.919
stable division in August. So this is the era of
197
00:10:08.639 --> 00:10:11.960
the image generators. This is also when the Getty images
198
00:10:12.039 --> 00:10:13.159
crisis kicks off.
199
00:10:13.279 --> 00:10:13.679
Oh yes.
200
00:10:13.759 --> 00:10:14.120
I remember.
201
00:10:14.159 --> 00:10:16.850
Because all these images keep getting generated with mangled Getty
202
00:10:16.889 --> 00:10:17.690
logos on them.
203
00:10:18.940 --> 00:10:22.450
And they say, no, we're not using public domain.
204
00:10:22.490 --> 00:10:24.570
Copyrighted materials. I don't know what you're talking about.
205
00:10:24.649 --> 00:10:25.389
What are you talking about?
206
00:10:25.509 --> 00:10:27.320
I don't know where that logo came from. What?
207
00:10:29.730 --> 00:10:30.950
A lot of hand-waving.
208
00:10:31.509 --> 00:10:31.789
All right.
209
00:10:32.350 --> 00:10:34.210
On the Microsoft side, in May, we finally get the
210
00:10:34.250 --> 00:10:37.639
full release of MAUI. We were supposed to get MAUI with.
211
00:10:37.769 --> 00:10:41.230
NET 6 back in November 2021. Right. But for a
212
00:10:41.549 --> 00:10:44.970
variety of reasons, mostly a new dev environment and a
213
00:10:45.029 --> 00:10:47.500
new framework at the same time made that kind of impossible.
214
00:10:47.559 --> 00:10:49.759
So the feature set that they really wanted to get to,
215
00:10:49.799 --> 00:10:52.480
they get to in May. They'll get major updates by
216
00:10:52.580 --> 00:10:54.700
November when we get. NET 7 and Studio 2022 and
217
00:10:54.720 --> 00:11:01.190
C Sharp 11. All good stuff. Yes. And also in November,
218
00:11:01.250 --> 00:11:04.830
we get the collapse of FTX. Do you remember this?
219
00:11:05.690 --> 00:11:06.570
This is one of those.
220
00:11:06.590 --> 00:11:07.950
Is that a crypto thing?
221
00:11:08.169 --> 00:11:10.580
Yeah, that's right. And one would argue this is the
222
00:11:10.669 --> 00:11:16.320
moment that creates a crypto winter that arguably they have
223
00:11:16.379 --> 00:11:18.620
not recovered from. And I'm having a hard time being
224
00:11:18.679 --> 00:11:22.480
upset about it. Because also in November, everybody gets really,
225
00:11:23.139 --> 00:11:27.799
really very distracted. when OpenAI pushes out ChatGPT.
226
00:11:28.120 --> 00:11:31.669
I was working for a customer that was doing crypto
227
00:11:32.129 --> 00:11:34.629
exchanges or something. I'm not going to get into it,
228
00:11:34.700 --> 00:11:38.340
but The project abruptly stopped after that.
229
00:11:38.399 --> 00:11:39.320
Yeah, oddly enough. Yeah.
230
00:11:39.460 --> 00:11:41.940
Because the demand just shriveled up.
231
00:11:42.019 --> 00:11:44.659
Sure. And suddenly we have another thing to use GPUs for, right?
232
00:11:44.700 --> 00:11:48.000
Because we've been dancing around with these techs for a while.
233
00:11:48.080 --> 00:11:52.259
But ChatGPT is the, I would argue, accidental breakthrough product
234
00:11:52.539 --> 00:11:54.299
in the sense that- I would say that. They were
235
00:11:54.340 --> 00:11:56.539
trying to do second order training. They were running low
236
00:11:56.620 --> 00:11:59.220
on money. And so- putting it out in the public
237
00:11:59.240 --> 00:12:01.070
was another way to do second order training for free
238
00:12:01.610 --> 00:12:03.830
uh and then they signed up 100 million users in
239
00:12:04.110 --> 00:12:06.909
two months yep and before that the current the record
240
00:12:06.990 --> 00:12:09.549
was tiktok had gotten 100 million users in nine months
241
00:12:09.590 --> 00:12:11.509
so two months was kind of a big deal but
242
00:12:11.529 --> 00:12:14.950
that is a story for 2023 yeah sure is that's
243
00:12:14.970 --> 00:12:15.690
what i got dude.
244
00:12:15.690 --> 00:12:19.309
All right well uh did i say matthias carlson is
245
00:12:19.350 --> 00:12:22.129
here he's been very silent but matthias feel free to
246
00:12:22.169 --> 00:12:23.909
jump in if you want yeah i'm going to uh
247
00:12:24.620 --> 00:12:26.000
Carry on with Better Know a Framework.
248
00:12:26.019 --> 00:12:26.620
Awesome.
249
00:12:26.779 --> 00:12:35.980
All right, man, what do you got? Okay, so you
250
00:12:36.019 --> 00:12:38.990
know that I've been working with local models on a
251
00:12:39.350 --> 00:12:45.129
GPU machine, high GPU machine here at Pop Studios. And
252
00:12:45.149 --> 00:12:48.200
I've been going through, you know, first it was a Lama,
253
00:12:48.500 --> 00:12:52.179
and then I talked about Lama CPP, which was better.
254
00:12:53.120 --> 00:12:56.159
And working with Lama CPP and Quen 3.8, which I
255
00:12:56.200 --> 00:13:02.039
think is the state-of-the-art model right now for locals, it
256
00:13:02.059 --> 00:13:04.860
just would bonk. And, you know, it was great for
257
00:13:04.879 --> 00:13:07.279
the demo, but, you know, then I tried to use
258
00:13:07.320 --> 00:13:09.679
it afterwards, and it was bonking and bonking. And what
259
00:13:09.720 --> 00:13:12.440
I mean by that, it was, you know, just stopping.
260
00:13:13.500 --> 00:13:16.120
And we couldn't really figure out why.
261
00:13:16.139 --> 00:13:16.279
So....
262
00:13:18.720 --> 00:13:21.820
I landed on, and this is for episode 45 of
263
00:13:21.980 --> 00:13:24.940
Code It With AI, I landed on NINFR.
264
00:13:25.220 --> 00:13:25.899
What's a NINFR?
265
00:13:26.179 --> 00:13:29.289
It's a runtime for hosting your models, just like Alama
266
00:13:29.450 --> 00:13:34.649
is in Alama CPP. However, it's really, really good. And
267
00:13:34.950 --> 00:13:38.230
I've been using it now as of this recording for
268
00:13:38.269 --> 00:13:43.990
a week, and it's been wonderful. I'm using a 128K
269
00:13:44.029 --> 00:13:47.360
context model that also has vision so you can upload
270
00:13:47.389 --> 00:13:48.320
screenshots to it.
271
00:13:48.340 --> 00:13:48.840
Nice.
272
00:13:48.860 --> 00:13:53.580
And I've been running it steadily. I've had one terminal
273
00:13:53.700 --> 00:13:57.840
open for four days, and I'm just talking to it
274
00:13:58.200 --> 00:14:00.679
like I would talk to ChatGPT, and it's great.
275
00:14:00.919 --> 00:14:01.139
Wow.
276
00:14:01.220 --> 00:14:03.639
Of course, you have to do your own memory management
277
00:14:04.470 --> 00:14:08.529
to keep state about things that you do. And I
278
00:14:08.590 --> 00:14:11.470
solved that. I'm not sharing this, but I have a repo.
279
00:14:12.629 --> 00:14:16.899
that is private, that's just called agent memory. And I
280
00:14:16.960 --> 00:14:19.840
have a whole bunch of MD files in there, including
281
00:14:19.860 --> 00:14:22.960
a startup MD that I tell it what to do,
282
00:14:23.460 --> 00:14:26.899
all the rules that it needs to obey, including before
283
00:14:26.940 --> 00:14:31.000
it does a compaction. I tell it to go reread
284
00:14:31.159 --> 00:14:35.279
this particular very important MD file because it has rules
285
00:14:35.320 --> 00:14:38.940
that fall off after the compaction happens.
286
00:14:38.960 --> 00:14:40.659
Cool so you create your own harness there so you
287
00:14:40.679 --> 00:14:42.500
have an own harness system.
288
00:14:42.559 --> 00:14:45.039
Yeah i have my own harness system exactly and i'm
289
00:14:45.059 --> 00:14:48.559
using co-pilot cli so it has you know some it
290
00:14:48.620 --> 00:14:51.580
is a good harness but you know if you if
291
00:14:51.620 --> 00:14:56.159
you're dealing with these llms you really have to customize
292
00:14:56.419 --> 00:14:58.740
i have i think you have to customize the the
293
00:14:58.759 --> 00:15:01.440
memory structure and what it's supposed to do and i
294
00:15:01.559 --> 00:15:04.830
even have um a carl md file up there that
295
00:15:05.279 --> 00:15:09.000
has all the things that are about me, right? So
296
00:15:09.059 --> 00:15:11.279
if it learns that, oh, you play in a Steely
297
00:15:11.299 --> 00:15:14.919
Dan cover band, it puts that up there, right? And, oh,
298
00:15:15.639 --> 00:15:19.049
you know, Carl eats fish on Friday. Okay, well, put
299
00:15:19.110 --> 00:15:23.129
that up there. So anyway, I wanted to share this
300
00:15:23.210 --> 00:15:27.879
document that goes through the search for a stable local
301
00:15:28.120 --> 00:15:31.500
coding LLM and all of the different things that I
302
00:15:31.539 --> 00:15:38.909
tried I tried VLLM. I tried SGLang. Ollama got 128K
303
00:15:39.610 --> 00:15:45.029
context back, but the copilot sessions became pathological.
304
00:15:45.730 --> 00:15:46.000
Wow.
305
00:15:46.500 --> 00:15:48.100
And NINFR is the clear winner.
306
00:15:48.379 --> 00:15:48.840
Cool.
307
00:15:49.000 --> 00:15:50.139
I'll have to look at that.
308
00:15:50.340 --> 00:15:52.820
I think it should be pronounced NINFR.
309
00:15:53.000 --> 00:15:54.539
Yeah, I think you're right. NINFR.
310
00:15:54.639 --> 00:15:54.899
Yeah.
311
00:15:54.919 --> 00:15:55.340
That's right.
312
00:15:55.480 --> 00:16:00.899
And it wants 64-bit Linux, a 5090... This is what
313
00:16:00.940 --> 00:16:02.860
you did with that 5090 machine you built?
314
00:16:02.940 --> 00:16:03.360
That's right.
315
00:16:03.440 --> 00:16:03.740
Wow.
316
00:16:03.919 --> 00:16:08.759
Yeah, so it's basically made for this GPU and for
317
00:16:08.820 --> 00:16:10.600
this model and infer.
318
00:16:10.659 --> 00:16:13.059
So how does it handle concurrency and things? That's sometimes
319
00:16:13.100 --> 00:16:14.220
hard to know when they die.
320
00:16:14.320 --> 00:16:16.600
Well, I don't have more than one person using it
321
00:16:16.639 --> 00:16:17.080
at a time.
322
00:16:17.539 --> 00:16:20.450
I was talking about multiple agents and things.
323
00:16:21.429 --> 00:16:24.230
Yeah, if I did, then I probably would run into
324
00:16:25.429 --> 00:16:29.750
some bonking or some context. The context would be used up. more,
325
00:16:29.950 --> 00:16:32.850
you know, more quickly, but it's just for me. And
326
00:16:33.110 --> 00:16:36.840
I would also add that Copilot CLI has a slash
327
00:16:36.899 --> 00:16:41.059
remote option so that what you can do is pull
328
00:16:41.139 --> 00:16:46.090
up GitHub on your phone and go to Copilot and
329
00:16:46.120 --> 00:16:48.309
you see the sessions that are there. And if you
330
00:16:48.350 --> 00:16:50.509
click on the last session you were in, you can
331
00:16:50.669 --> 00:16:55.429
actually operate that Copilot CLI session from your phone. And
332
00:16:55.480 --> 00:16:57.519
it goes through GitHub. You don't need a VPN or
333
00:16:57.559 --> 00:17:00.159
anything like that. So you're authenticated on your phone. You're
334
00:17:00.200 --> 00:17:04.180
authenticated in the CLI. And you can even see what
335
00:17:04.220 --> 00:17:07.140
you type on the screen in the CLI terminal. Like
336
00:17:07.200 --> 00:17:09.700
it's really, really good. And it works with the local model.
337
00:17:09.920 --> 00:17:12.099
So that's what I got. Cool. Getting there. All right, man.
338
00:17:12.259 --> 00:17:12.500
All right.
339
00:17:12.759 --> 00:17:14.700
One more thing. Who's talking to us today, Richard?
340
00:17:14.839 --> 00:17:17.180
Grabbed a comment off the show 2017, the one we
341
00:17:17.220 --> 00:17:20.309
did with Billy Hollis back in August, which I just
342
00:17:20.349 --> 00:17:22.309
called arguing about AI because I think we did that.
343
00:17:22.410 --> 00:17:24.329
That's pretty much what we did.
344
00:17:24.390 --> 00:17:29.690
Yeah. And oddly enough, generated a lot of conversation. This
345
00:17:29.730 --> 00:17:33.259
particular one comes from Volkmar, who says, hi, good discussion. However,
346
00:17:33.299 --> 00:17:35.180
I think it was more from a developer point of
347
00:17:35.200 --> 00:17:37.900
view than from a software owner's point of view. I
348
00:17:37.940 --> 00:17:39.480
have been listening to your show for many years now,
349
00:17:39.660 --> 00:17:43.099
often sitting on a tractor in Germany as I'm a
350
00:17:43.220 --> 00:17:47.009
farmer and a software developer at the same time. So cool.
351
00:17:47.089 --> 00:17:47.589
So cool.
352
00:17:48.359 --> 00:17:50.339
I was working in the past also for a precision
353
00:17:50.420 --> 00:17:55.660
farming company designing web applications, starting with Silverlight. There were
354
00:17:56.220 --> 00:17:58.809
always so many ideas and so little resources to implement
355
00:17:58.829 --> 00:18:01.710
them back then. If you're a specialist in any field
356
00:18:02.089 --> 00:18:04.509
and a developer, it's like you get a genie in
357
00:18:04.529 --> 00:18:09.269
a bottle. Yeah, domain knowledge is valuable. In the end,
358
00:18:09.309 --> 00:18:11.470
it is the value for the user or customer of
359
00:18:11.490 --> 00:18:15.109
the software we can develop for real needs now, not
360
00:18:15.170 --> 00:18:17.829
only where the clients are making a good margin. I
361
00:18:17.869 --> 00:18:20.789
also believe that AI-generated software without an experienced developer in
362
00:18:20.809 --> 00:18:25.190
the driver's seat will not be successful. Keep it up. Thanks, Baltimore.
363
00:18:25.309 --> 00:18:25.609
Nice.
364
00:18:25.690 --> 00:18:26.099
There you go.
365
00:18:26.190 --> 00:18:27.579
Sounds like a Greenfield developer.
366
00:18:28.319 --> 00:18:31.460
That's really a Greenfield developer, my dears. You are right.
367
00:18:33.700 --> 00:18:35.799
How did I leave that on the table? Thank you
368
00:18:35.819 --> 00:18:36.819
for picking it up.
369
00:18:37.559 --> 00:18:40.269
Mateus is like, I can't believe any of these guys
370
00:18:40.289 --> 00:18:41.170
have thought of that yet.
371
00:18:41.269 --> 00:18:44.190
How do you leave that one alone? That's crazy. I know.
372
00:18:44.210 --> 00:18:48.970
I think he starts to spring on a brownfield, actually. Hey, Volkmar,
373
00:18:48.990 --> 00:18:51.599
thank you so much for your comment. Keep on farming
374
00:18:51.640 --> 00:18:53.779
and listen to us while you're driving around. And to
375
00:18:53.799 --> 00:18:55.460
give you some more to listen to, I'd like to
376
00:18:55.500 --> 00:18:57.099
send you a copy of Music to Codebyte. If you'd
377
00:18:57.140 --> 00:18:58.700
like a copy of Music to Codebyte, write a comment
378
00:18:58.720 --> 00:19:01.740
on the website at totnetrocks.com or on the Facebooks. We
379
00:19:01.789 --> 00:19:03.670
publish every show there and every comment there and everything
380
00:19:03.690 --> 00:19:05.230
on the show, we'll send you a copy of Music
381
00:19:05.250 --> 00:19:05.490
to Codebyte.
382
00:19:05.509 --> 00:19:06.769
Or if you just want to go get it, go
383
00:19:06.809 --> 00:19:12.170
to musictocodebyte.net and you can download the entire collection in MP3, WAV,
384
00:19:12.509 --> 00:19:13.390
and FLAC formats.
385
00:19:13.950 --> 00:19:14.119
Cool.
386
00:19:14.339 --> 00:19:15.519
All right, Matthias, you're up.
387
00:19:15.799 --> 00:19:16.440
What are you doing?
388
00:19:17.210 --> 00:19:21.170
This is not about cake. Let me read your bio,
389
00:19:21.230 --> 00:19:24.900
first of all. Matthias Carlsen is a partner and senior
390
00:19:24.940 --> 00:19:31.099
architect at WCOM AB in Gultenberg, Sweden, a Microsoft.net MVP
391
00:19:31.200 --> 00:19:35.480
and a passionate open source citizen. Matthias has been professionally
392
00:19:36.180 --> 00:19:40.490
in IT for almost 30 years, started coding already in
393
00:19:40.750 --> 00:19:44.690
the 1980s on a Commodore 128. But the true love
394
00:19:44.730 --> 00:19:49.269
for coding came with the Amiga. Outside of technology, Matthias
395
00:19:49.329 --> 00:19:51.519
is the father of two and the husband of one.
396
00:19:53.779 --> 00:19:55.500
All right, what are you doing these days, Matthias?
397
00:19:55.859 --> 00:19:59.559
Well, I'm a developer, mostly. I'm fielding AI like most
398
00:19:59.599 --> 00:20:02.180
people do, but I'm trying to do it in with
399
00:20:02.200 --> 00:20:05.880
some sense, because what I've seen is how we maintain
400
00:20:05.920 --> 00:20:08.059
things and stuff like that, but that's not what I'm
401
00:20:08.079 --> 00:20:12.059
going to talk about today. I think it's tech-related, but
402
00:20:12.400 --> 00:20:13.599
more soft in one way.
403
00:20:13.759 --> 00:20:16.839
I think I just heard thousands of people give a
404
00:20:16.880 --> 00:20:20.319
sigh of relief. Finally, those guys are talking about something
405
00:20:20.359 --> 00:20:20.799
else than AI.
406
00:20:21.019 --> 00:20:24.309
Yeah, it's even more important, I think. It's even important,
407
00:20:24.630 --> 00:20:28.390
because I got talking with Richard about... Well, I think
408
00:20:28.430 --> 00:20:33.180
we all here now in this room is... people with experience.
409
00:20:33.579 --> 00:20:36.180
I will talk about things like I'm not a lawyer,
410
00:20:36.359 --> 00:20:40.380
but I have learned things by just the earth spinning around.
411
00:20:42.119 --> 00:20:45.619
And one thing you notice when you get to a
412
00:20:45.619 --> 00:20:51.019
certain age is that, well, things like people that you
413
00:20:51.039 --> 00:20:56.309
thought wouldn't be there forever aren't. And you will notice
414
00:20:56.329 --> 00:21:01.529
that Well, us technologists, we are hoarders in some ways,
415
00:21:01.809 --> 00:21:08.250
both with our GitHub repositories, our technology, our retro equipment,
416
00:21:08.710 --> 00:21:09.410
or whatnot.
417
00:21:09.730 --> 00:21:11.500
Our drive folders full of old projects.
418
00:21:11.799 --> 00:21:16.039
Yeah. Before. I mean, I think what I reached out
419
00:21:16.059 --> 00:21:18.640
to Richard is, like, I'm not unique in this situation.
420
00:21:18.660 --> 00:21:19.240
I think we are.
421
00:21:19.259 --> 00:21:19.319
No.
422
00:21:19.319 --> 00:21:23.400
And there are many of us that live in what
423
00:21:23.460 --> 00:21:27.160
Hanselman said, like, mixed marriage, where... where there's one geek
424
00:21:27.200 --> 00:21:28.579
and one normal person.
425
00:21:28.380 --> 00:21:36.599
And that's kind of like and like with true experience
426
00:21:36.640 --> 00:21:41.380
I've had some well like collateral damage I lost friends
427
00:21:42.099 --> 00:21:45.589
like and things like that friends that have been.
428
00:21:47.690 --> 00:21:51.369
Like in different ways you can have like temporary with
429
00:21:51.410 --> 00:21:55.750
illness or it could be final terminal things and that's
430
00:21:55.769 --> 00:21:58.109
why I like had some ideas that I would like
431
00:21:58.470 --> 00:21:59.549
have a discussion.
432
00:21:59.170 --> 00:22:01.230
With you about. And I think some things that are
433
00:22:01.289 --> 00:22:03.170
important for us as innovators.
434
00:22:03.710 --> 00:22:05.460
And when you messaged me about this, I'm like, I
435
00:22:05.500 --> 00:22:08.059
had just gone through this with a friend who's passed away,
436
00:22:08.180 --> 00:22:10.380
whose spouse reached out to me and said, how do
437
00:22:10.400 --> 00:22:13.180
I do this? Like, how do I get access to
438
00:22:13.240 --> 00:22:16.279
his stuff? How do I clean up the accounts? How
439
00:22:16.319 --> 00:22:19.130
do I operate the house? He was a smart home guy. Like,
440
00:22:19.150 --> 00:22:22.079
and all of a sudden you're like, I'm trying to
441
00:22:22.180 --> 00:22:26.960
sysadmin for his place, uh, without him around. Like, and
442
00:22:26.980 --> 00:22:29.269
he ain't, he's not coming back. So I, you know,
443
00:22:29.289 --> 00:22:32.289
it turned into a very much forensic effort. What do
444
00:22:32.309 --> 00:22:32.910
we got to do?
445
00:22:33.150 --> 00:22:34.730
And that was still like the impotence for me also.
446
00:22:34.769 --> 00:22:36.869
Like I have actually a couple of years ago, I
447
00:22:36.869 --> 00:22:39.480
had a friend that like I sat with his wife
448
00:22:39.519 --> 00:22:42.279
when he passed away. And the whole thing there is was,
449
00:22:42.700 --> 00:22:45.569
and I, That's why I had this idea to talk
450
00:22:45.650 --> 00:22:47.619
about in draft for many years, because it was difficult
451
00:22:47.640 --> 00:22:52.500
when it happened. We spent hours going through. If someone's
452
00:22:52.519 --> 00:22:56.410
been around since the 70s, we gather things. You have
453
00:22:56.470 --> 00:23:02.309
virtual private servers running, source resources, online accounts, subscriptions. And
454
00:23:02.549 --> 00:23:03.490
like many of us.
455
00:23:03.509 --> 00:23:05.690
It's going to take somebody weeks to go through all
456
00:23:05.730 --> 00:23:06.109
my stuff.
457
00:23:06.190 --> 00:23:08.549
And like many also like to buy a domain a
458
00:23:08.569 --> 00:23:10.960
week and things like that. So it's, there's, there's a
459
00:23:10.960 --> 00:23:12.799
lot of things. And, uh, there are things that I
460
00:23:12.819 --> 00:23:16.039
think we need to, uh, it's, it's not like, I
461
00:23:16.079 --> 00:23:18.380
think to go a morbid way, it's just like, I
462
00:23:18.400 --> 00:23:22.160
think it matters to, to have some, like, I don't
463
00:23:22.240 --> 00:23:24.509
want to put my wife and kids in a situation
464
00:23:24.549 --> 00:23:28.069
that I've worked with my friend because it's And also
465
00:23:28.089 --> 00:23:33.130
subsigned the right home for the right thing. For a
466
00:23:33.730 --> 00:23:39.880
foreseeable future, I realized that even if I'm not here,
467
00:23:40.059 --> 00:23:41.259
the earth will still turn around.
468
00:23:41.279 --> 00:23:41.460
Yeah, sure.
469
00:23:42.980 --> 00:23:45.460
And that's why I like and I know some have
470
00:23:45.599 --> 00:23:48.079
seen the like the Swedes things like about death cleaning
471
00:23:48.140 --> 00:23:50.339
and things that like, how do we get our things
472
00:23:50.400 --> 00:23:52.500
in order? Or how do we have.
473
00:23:52.359 --> 00:23:55.480
Nobody told my Swedish mother about Swedish death cleaning. And
474
00:23:55.539 --> 00:23:56.660
so it all fell on me.
475
00:23:56.920 --> 00:23:59.960
Yeah, because we're like, we're, we're trying not to be
476
00:24:00.019 --> 00:24:02.329
too emotional, but we are very practical. I think we
477
00:24:02.349 --> 00:24:07.170
are Swedes. And like, it's not stoic. And also, it's
478
00:24:07.210 --> 00:24:12.359
not only about death is by things like hospital stays, burnout, accidents,
479
00:24:12.440 --> 00:24:15.279
and things that, things can happen where you have to
480
00:24:15.559 --> 00:24:18.220
like think, I think it's good to, just like you
481
00:24:18.240 --> 00:24:20.960
do with a contractor agreement, you have to do things
482
00:24:21.039 --> 00:24:24.210
sometimes before it happens to make it easier when it happens.
483
00:24:24.230 --> 00:24:25.250
Sure, yeah.
484
00:24:25.509 --> 00:24:30.750
I'm lucky because my wife is technical and she, I
485
00:24:30.769 --> 00:24:33.190
give her complete access to everything that I have. We
486
00:24:33.250 --> 00:24:38.799
share the password manager and all that stuff. So she
487
00:24:38.880 --> 00:24:42.309
goes and notices things that I've paid for that I'm
488
00:24:42.349 --> 00:24:45.369
no longer using and gives me a list like every
489
00:24:45.410 --> 00:24:47.910
couple of weeks. Hey, do you still use this? Do
490
00:24:47.930 --> 00:24:48.789
you still use that?
491
00:24:48.869 --> 00:24:48.990
Yeah.
492
00:24:49.029 --> 00:24:49.670
What is this?
493
00:24:50.109 --> 00:24:50.750
Yeah. Yeah.
494
00:24:50.789 --> 00:24:51.369
What is this?
495
00:24:51.390 --> 00:24:53.839
That's usually what I use a credit card expiry for
496
00:24:53.880 --> 00:24:56.599
because then like payment failed. And I was supposed to
497
00:24:56.640 --> 00:24:58.500
pay him for it.
498
00:24:58.660 --> 00:25:02.559
And we do have documentation online that talks about all
499
00:25:02.599 --> 00:25:05.990
the things that that I do that have to happen afterwards.
500
00:25:06.230 --> 00:25:12.579
And also, I have a great big honking NAS drive here.
501
00:25:13.579 --> 00:25:16.460
And it's got way more stuff on it than it
502
00:25:16.500 --> 00:25:17.480
should have.
503
00:25:17.779 --> 00:25:19.339
Yeah, they all have that.
504
00:25:19.960 --> 00:25:22.019
Yeah, because when I came to my work, what I
505
00:25:22.319 --> 00:25:26.509
started to ask myself is like, because some things aren't
506
00:25:26.549 --> 00:25:29.329
that important. They're important to me. So sometimes I decide
507
00:25:29.490 --> 00:25:31.569
what should die with me or what should live on,
508
00:25:31.690 --> 00:25:36.349
like the categories. And also things like who needs access
509
00:25:36.369 --> 00:25:39.890
to what. And things like domains.
510
00:25:40.410 --> 00:25:41.690
What is actually valuable?
511
00:25:43.049 --> 00:25:47.049
And for some open source projects, it's kind of the same.
512
00:25:47.089 --> 00:25:49.099
Because there are some things that are just important to me.
513
00:25:49.500 --> 00:25:52.980
There are some things that are mildly important. And there
514
00:25:53.019 --> 00:25:55.680
are some things that are successful and many depend on.
515
00:25:55.700 --> 00:25:58.880
So it could be different paths depending on what you
516
00:25:58.900 --> 00:25:59.259
have there.
517
00:25:59.279 --> 00:26:01.700
I mean, I got to think, especially if it's abrupt.
518
00:26:02.500 --> 00:26:06.130
It's the day-to-day first, like what bills are, you know,
519
00:26:06.170 --> 00:26:08.950
what charges are being incurred that are no longer relevant
520
00:26:09.000 --> 00:26:12.920
and how do we fix that? what things count on
521
00:26:12.980 --> 00:26:16.349
the email address of the deceased. I think number one is,
522
00:26:16.369 --> 00:26:18.829
can you get access to that email? And maybe an
523
00:26:18.890 --> 00:26:21.549
authenticator device, like the phone and the email account have
524
00:26:21.609 --> 00:26:24.109
got to be the starting point to deal with the
525
00:26:24.210 --> 00:26:27.450
day-to-day before we even get into the legacy.
526
00:26:27.650 --> 00:26:32.000
And the problem is, like, what actually happened on one
527
00:26:32.079 --> 00:26:36.039
process project where the maintainer died is that GitHub is really,
528
00:26:36.059 --> 00:26:38.349
as it should be, you need to go through the
529
00:26:38.369 --> 00:26:42.509
estate to get things. And that's why it's important to
530
00:26:42.609 --> 00:26:46.670
identify what should the successors be for this project. Or
531
00:26:47.210 --> 00:26:49.390
if it's a bigger project, you need co-maintainers so you
532
00:26:49.430 --> 00:26:50.569
have redundancy to things.
533
00:26:50.670 --> 00:26:53.450
Yeah, you could deal with this in advance.
534
00:26:53.710 --> 00:26:57.690
But also things like Carl mentioned, the credentials are important.
535
00:26:57.990 --> 00:27:02.250
What are the... What passwords are needed, who needs what,
536
00:27:02.309 --> 00:27:04.170
and then path for things like that.
537
00:27:04.269 --> 00:27:06.519
When I think about legacy now, I think about what
538
00:27:06.579 --> 00:27:08.960
most people can, you know, what do you carry out
539
00:27:09.000 --> 00:27:13.079
of the house? And it's photo. Like, if the deceased
540
00:27:13.140 --> 00:27:15.960
is the one with the big Google Photos account or
541
00:27:16.000 --> 00:27:19.809
wherever they put it, you know, who has access to that?
542
00:27:19.829 --> 00:27:23.230
Because you lose those, you've really lost something. You talk
543
00:27:23.269 --> 00:27:24.470
about loss of legacy.
544
00:27:24.670 --> 00:27:26.650
Yeah, and I think it's easier when you have the
545
00:27:26.690 --> 00:27:29.430
cloud services because you don't think about it. I like it.
546
00:27:30.470 --> 00:27:31.950
Yeah.
547
00:27:32.299 --> 00:27:35.539
Because I've been, like, when you did the cleaning after
548
00:27:35.579 --> 00:27:38.359
my grandparents passed away, you did the whole physical thing,
549
00:27:38.440 --> 00:27:40.819
like what furniture to keep and things like that. But
550
00:27:41.240 --> 00:27:45.160
it's much harder to know if you don't have this
551
00:27:45.240 --> 00:27:49.069
kind of, like, inventory of your digital, like, bill of
552
00:27:49.109 --> 00:27:53.450
materials of your digital life. Then it's really hard for
553
00:27:53.470 --> 00:27:55.970
someone else to know. And sometimes you find out when
554
00:27:56.019 --> 00:27:58.200
it's too late, when the domain expires or when things
555
00:27:58.240 --> 00:28:02.500
go inactive or whatever. And I think usually what I
556
00:28:02.539 --> 00:28:07.049
did is I did some categories of action. So what
557
00:28:07.109 --> 00:28:13.289
are personal ownerships? This is something that is only important
558
00:28:13.329 --> 00:28:17.200
to me. Then I will remain controlling indefinitely and it
559
00:28:17.220 --> 00:28:19.720
will die with me. It could be something that is
560
00:28:19.920 --> 00:28:24.380
like emergency succession where someone else can take over if
561
00:28:24.460 --> 00:28:28.390
I'm temporarily incapable. This product is someone who can take over.
562
00:28:28.569 --> 00:28:32.809
We're really talking about digital assets, right? Yeah, mostly. I mean,
563
00:28:32.829 --> 00:28:33.589
the physical ones.
564
00:28:33.710 --> 00:28:38.130
Or it could be things like my retro gaming collection is... Like,
565
00:28:38.170 --> 00:28:40.119
my wife has zero interest in that. But I could
566
00:28:40.240 --> 00:28:44.839
find someone... I have friends in the community that value those.
567
00:28:44.960 --> 00:28:45.359
Right, yeah.
568
00:28:45.380 --> 00:28:47.960
So there could be a plan for those, like... They
569
00:28:48.000 --> 00:28:50.519
have monetary value, perhaps, for my wife. Then someone that
570
00:28:50.579 --> 00:28:52.200
knows the stuff can sell it for her.
571
00:28:52.400 --> 00:28:53.369
Or it could be.
572
00:28:53.480 --> 00:28:54.710
So it could be physical things, too.
573
00:28:54.730 --> 00:28:55.829
Sure.
574
00:28:56.009 --> 00:28:56.230
Because.
575
00:28:56.950 --> 00:28:59.589
You can be a geek in many areas. It could
576
00:28:59.630 --> 00:29:03.150
be in bicycles or it could be whatever that you
577
00:29:03.170 --> 00:29:05.420
have a disconnect between. And it could be vice versa
578
00:29:05.440 --> 00:29:07.779
with the wife that she has some special interest that
579
00:29:07.819 --> 00:29:10.980
you don't. And then it can be like, who should
580
00:29:11.019 --> 00:29:14.420
take over this? But it also can be something like
581
00:29:14.920 --> 00:29:17.799
temporary delegation where you just hand over for a while.
582
00:29:18.180 --> 00:29:22.420
And it can also be some open succession where you
583
00:29:22.440 --> 00:29:26.539
essentially can post someone anyone that cares about the source
584
00:29:26.559 --> 00:29:28.480
party or something can put their hand up.
585
00:29:28.769 --> 00:29:30.650
Are we talking about a digital will?
586
00:29:31.029 --> 00:29:33.569
Yeah, it's actually like, well, I think like we have
587
00:29:33.710 --> 00:29:37.250
like the S-bomb for software building materials, but almost like
588
00:29:37.289 --> 00:29:41.869
have a digital building materials or your life's like, what
589
00:29:41.970 --> 00:29:44.619
are the things to have an inventory for things like
590
00:29:44.670 --> 00:29:48.519
services domains? I think it's good to have. And it's quite,
591
00:29:48.640 --> 00:29:50.539
I think that's something you could use AI for now
592
00:29:50.579 --> 00:29:52.880
where like, do absolutely. That's what I would go through
593
00:29:52.900 --> 00:29:57.680
your things. See what, I also like things like personally legacy,
594
00:29:57.700 --> 00:29:59.930
like I want them to remain and I want them
595
00:29:59.950 --> 00:30:01.930
to be active, then I can transfer it to someone
596
00:30:01.970 --> 00:30:06.670
else or things like that. So I think it's good
597
00:30:06.690 --> 00:30:11.269
to have those kind of discussions before you do it yourself,
598
00:30:11.509 --> 00:30:16.670
because As long as you're alive, you decide what should happen.
599
00:30:16.690 --> 00:30:20.000
But also you should go back to that list and see,
600
00:30:20.019 --> 00:30:22.220
because it could change. It could be like when my
601
00:30:22.240 --> 00:30:24.779
kids were five, they weren't interested in code, but it
602
00:30:24.819 --> 00:30:26.819
could be that when they're 20, they're interested in it.
603
00:30:26.859 --> 00:30:28.900
And so it could be that you do some... Same
604
00:30:28.940 --> 00:30:31.529
way you need to refresh any will. Reiterate, yeah.
605
00:30:31.720 --> 00:30:33.210
Hey, we should take a break. Yeah, we should.
606
00:30:33.430 --> 00:30:35.630
We'll be right back after these very important messages. Don't
607
00:30:35.650 --> 00:30:35.890
go away.
608
00:30:39.720 --> 00:30:42.000
And we're back. It's. NET Rocks. I'm Richard Campbell. That's
609
00:30:42.039 --> 00:30:43.289
Carl Franklin. Hey.
610
00:30:43.549 --> 00:30:44.009
We're back.
611
00:30:44.049 --> 00:30:46.609
We're spending some time with our friend, Matthias Carlson, about
612
00:30:46.650 --> 00:30:49.450
this idea of a digital legacy. Yeah, I'm still hung
613
00:30:49.509 --> 00:30:52.450
up on the fundamentals of who could get into my
614
00:30:52.529 --> 00:30:55.430
email account, who could get access to my phone for
615
00:30:55.450 --> 00:31:00.769
my authenticator. Because you're really jammed. And if you were
616
00:31:00.809 --> 00:31:04.680
only using face recognition... You don't have that option. You
617
00:31:04.700 --> 00:31:07.779
better make sure there's some backup strategies that are difficult
618
00:31:07.799 --> 00:31:10.359
to use but possible by someone else.
619
00:31:10.700 --> 00:31:13.359
You can't just count on your spouse or your partner
620
00:31:13.460 --> 00:31:17.619
because you two might die in a fiery crash.
621
00:31:17.779 --> 00:31:19.680
Might go together, the old plane crash thing.
622
00:31:19.700 --> 00:31:23.460
So have a child or somebody at the ready to
623
00:31:23.519 --> 00:31:23.900
do that.
624
00:31:24.380 --> 00:31:25.039
An executor.
625
00:31:25.059 --> 00:31:25.940
An executor, yeah.
626
00:31:26.359 --> 00:31:29.839
But also, because I've been through it, if you have
627
00:31:29.880 --> 00:31:33.700
someone near to you, You have stuff to arrange regardless.
628
00:31:34.500 --> 00:31:37.599
So you would like everything that happens. And that's kind
629
00:31:37.640 --> 00:31:39.569
of like, just to put the prayer, like, I think
630
00:31:39.609 --> 00:31:43.089
it's thoughtful to think of your, what can I do
631
00:31:43.230 --> 00:31:45.529
to make it easy? Like these kinds of things, I
632
00:31:45.569 --> 00:31:47.599
know this friend can take care of this. I can
633
00:31:47.630 --> 00:31:51.019
pass them to like a foundation or something and have
634
00:31:51.059 --> 00:31:53.920
that kind of thing. And that's kind of nice, like
635
00:31:53.960 --> 00:31:55.789
when you see something like the dotnet foundation where they
636
00:31:55.809 --> 00:31:58.190
have like one password and they have keyboards and things
637
00:31:58.210 --> 00:32:04.029
like that where something happens. And I think that's the
638
00:32:04.069 --> 00:32:07.829
discussion you need to have. You want to be thoughtful
639
00:32:07.910 --> 00:32:12.759
of the people around you. And also it's nice to
640
00:32:13.450 --> 00:32:16.140
leave that wish list because then they have one less
641
00:32:16.200 --> 00:32:19.900
thing to decide. But also, like I said, it could
642
00:32:19.920 --> 00:32:21.480
be temporary. It could be like if you have a
643
00:32:21.500 --> 00:32:25.349
stroke or if you have... You could survive the pain crash,
644
00:32:25.369 --> 00:32:26.970
but you couldn't manage the project.
645
00:32:26.990 --> 00:32:28.890
Well, and now you get into a situation like an
646
00:32:28.970 --> 00:32:32.809
executor is a legal construct that has specific rights. Yeah.
647
00:32:32.930 --> 00:32:36.089
But you have to be dead for that to activate.
648
00:32:36.150 --> 00:32:39.299
So what is the sort of living will but incapacitated
649
00:32:39.619 --> 00:32:40.519
model of that?
650
00:32:40.880 --> 00:32:44.269
And that's something I would like more... I think if
651
00:32:44.289 --> 00:32:46.670
we're a young industry, I would like more services to
652
00:32:46.730 --> 00:32:51.200
have this mindset. Yeah, because I know like GitHub added
653
00:32:51.220 --> 00:32:54.140
the successor model where you can actually add someone as
654
00:32:54.160 --> 00:32:58.380
a successor to you. And if you're become encapsulated, they
655
00:32:58.400 --> 00:33:01.519
could request access and get rid of the delay. But
656
00:33:01.539 --> 00:33:04.099
that's a big base for web, because that is everything.
657
00:33:04.119 --> 00:33:07.369
What they can do, essentially, then they can archive and
658
00:33:07.410 --> 00:33:10.890
transfer post stories that are public. They don't have login access,
659
00:33:10.930 --> 00:33:13.529
but they can still take over things that matter for others.
660
00:33:14.210 --> 00:33:16.430
And I know GitLab has the same, but then you
661
00:33:16.450 --> 00:33:20.039
get full, but others don't. So that's why If you
662
00:33:20.059 --> 00:33:22.920
have a bigger product, it's better to have a GitHub organization.
663
00:33:22.940 --> 00:33:26.279
Because then you can have multiple maintainers and owners and
664
00:33:26.299 --> 00:33:26.920
things like that.
665
00:33:27.059 --> 00:33:30.049
So things like that can be good to have.
666
00:33:30.190 --> 00:33:35.029
Yeah. I'm still digging through each of these different phases.
667
00:33:35.869 --> 00:33:37.950
I don't know that I want... I have been an executor.
668
00:33:38.529 --> 00:33:41.759
I did not enjoy the process. But being the technical
669
00:33:41.819 --> 00:33:45.859
advisor to the executor... Like, I generally recommend these days
670
00:33:45.880 --> 00:33:47.160
when people ask me, like, you should be my executor.
671
00:33:47.180 --> 00:33:49.680
It's like, you should probably have a lawyer. Preferably one
672
00:33:49.700 --> 00:33:53.119
you trust. But for us in the tech industry, being
673
00:33:53.180 --> 00:33:56.680
a friend to that person and being available to that
674
00:33:56.740 --> 00:33:59.490
executor who's probably not that technical to deal with the
675
00:33:59.529 --> 00:34:01.609
technical side of stuff is very useful.
676
00:34:01.910 --> 00:34:05.009
Yeah. And I think the whole thing there, I think it's...
677
00:34:06.589 --> 00:34:09.380
That's important to just give the mandate to the right
678
00:34:09.420 --> 00:34:13.730
person that can assist. And there are things you need to.
679
00:34:15.400 --> 00:34:18.300
To see like the whole thing when you have foresight
680
00:34:18.320 --> 00:34:22.440
to give someone attorney for you it's I think that's
681
00:34:23.000 --> 00:34:27.550
important and it's something and and even like they are
682
00:34:27.590 --> 00:34:30.690
getting better I like when I'm a grandmother passed away
683
00:34:30.730 --> 00:34:32.329
we went for funeral home and they know about things
684
00:34:32.389 --> 00:34:35.929
like Instagram and Facebook and closing those accounts down sure
685
00:34:35.949 --> 00:34:39.639
yeah so they're not it's getting better The social.
686
00:34:39.400 --> 00:34:41.260
Media side is one thing, but I get what you're
687
00:34:41.300 --> 00:34:43.679
talking about, which is like, I didn't know about GitHub Successor.
688
00:34:43.699 --> 00:34:44.920
I'll put that in the show links.
689
00:34:44.940 --> 00:34:46.380
That's a great tool.
690
00:34:47.199 --> 00:34:50.829
I think that's something I would wish all services had.
691
00:34:51.070 --> 00:34:51.889
Sure.
692
00:34:52.769 --> 00:34:58.269
Because it matters like my Netflix account Only my family
693
00:34:58.309 --> 00:35:01.000
wants that, but my GitHub account is a different thing
694
00:35:01.300 --> 00:35:04.320
where I actually had a peer of mine and I
695
00:35:04.320 --> 00:35:07.719
was the successor of him just because we are the
696
00:35:07.800 --> 00:35:09.699
ones that have the technical knowledge and know what to
697
00:35:09.739 --> 00:35:14.389
do with things and how to maintain it. And that's
698
00:35:14.429 --> 00:35:16.230
not a load that we want to place on a
699
00:35:16.269 --> 00:35:17.210
non-technical person.
700
00:35:17.369 --> 00:35:17.530
No.
701
00:35:18.000 --> 00:35:21.000
And in general, that's my thinking for all of this
702
00:35:21.039 --> 00:35:23.260
stuff is I do not want to burden my loved
703
00:35:23.380 --> 00:35:26.880
ones with this any more than is absolutely necessary. How
704
00:35:26.920 --> 00:35:28.480
do you make it as simple as possible?
705
00:35:28.679 --> 00:35:32.139
Yeah, but also some good advice like with password managers
706
00:35:32.179 --> 00:35:34.619
where you can have, many have where you can actually
707
00:35:34.659 --> 00:35:38.550
have a physical printed document which is still like the
708
00:35:38.639 --> 00:35:40.280
emergency reset thing.
709
00:35:40.780 --> 00:35:43.460
Or you can have like a USB drive.
710
00:35:43.500 --> 00:35:46.980
Setting up a shared password family password manager has been
711
00:35:47.019 --> 00:35:48.949
a big thing. Not only just because that we both
712
00:35:48.989 --> 00:35:51.110
have access to the home account. So if I was gone,
713
00:35:51.530 --> 00:35:54.369
she could just... has access to accounts anyway, but that
714
00:35:54.389 --> 00:35:56.719
you have a whole mechanism for, yeah, you can take
715
00:35:56.780 --> 00:35:58.880
over my account. Here's the successor system for that.
716
00:35:59.300 --> 00:36:01.559
And that's good to know. And the whole phone thing
717
00:36:01.579 --> 00:36:03.739
you mentioned is valid because phones die.
718
00:36:03.800 --> 00:36:04.260
They do.
719
00:36:04.340 --> 00:36:05.179
It doesn't have to be.
720
00:36:05.380 --> 00:36:08.050
You have a limited amount of time to figure this out. Yeah.
721
00:36:08.239 --> 00:36:10.650
Yeah. But also for yourself, like it could be that
722
00:36:10.690 --> 00:36:12.570
you drop your phone in the ocean and how do
723
00:36:12.610 --> 00:36:15.730
I get things? So it's, it's regardless, a good practice
724
00:36:15.750 --> 00:36:16.010
to have.
725
00:36:16.170 --> 00:36:19.690
Yeah, let's face it. If you're lost in an aircraft disaster,
726
00:36:20.090 --> 00:36:21.070
your phone was probably with you.
727
00:36:21.170 --> 00:36:24.300
Yeah, your computer's with you, your phone's with you. Like,
728
00:36:24.360 --> 00:36:24.780
it could be.
729
00:36:25.320 --> 00:36:25.659
Yeah.
730
00:36:26.719 --> 00:36:29.780
And that's why sometimes, like, it's good to have that peer-to-peer,
731
00:36:29.820 --> 00:36:34.320
to have that other technical friend that hopefully isn't on
732
00:36:34.340 --> 00:36:37.800
the same plane. But also to spread your risks, to have, like, see,
733
00:36:38.139 --> 00:36:41.449
for GitHub, it's this guy. For my retrocomputing, it's this guy.
734
00:36:41.469 --> 00:36:43.750
Or girl, whatever.
735
00:36:43.889 --> 00:36:48.210
And I think that could be then it's much easier
736
00:36:48.250 --> 00:36:52.659
for the executor to just go through things and like
737
00:36:52.699 --> 00:36:53.110
a checklist.
738
00:36:53.139 --> 00:36:55.440
Be able to work through that. I just pulled the
739
00:36:55.500 --> 00:36:59.400
notes for Bitwarden's emergency access resources. So now you've got
740
00:36:59.440 --> 00:37:01.760
me searching around the internet for what are the tools
741
00:37:01.840 --> 00:37:04.119
I use and what did they do in these circumstances?
742
00:37:05.539 --> 00:37:08.500
I don't think anybody looks at this and probably you should.
743
00:37:08.679 --> 00:37:12.210
And also it's complicated because we have something in Sweden
744
00:37:12.230 --> 00:37:14.610
where you can do like future power of attorney to
745
00:37:15.110 --> 00:37:19.070
delegate that to someone else. And that's something to notice
746
00:37:19.389 --> 00:37:21.469
when you work in open source is that the legal
747
00:37:21.510 --> 00:37:24.570
things can be different in different countries. So that's also
748
00:37:24.610 --> 00:37:29.840
something to have that in mind where there can be
749
00:37:29.880 --> 00:37:33.079
gaps or there could be where you notice like you
750
00:37:33.119 --> 00:37:35.119
can't just grant access or whatever.
751
00:37:36.000 --> 00:37:36.480
So that's.
752
00:37:38.070 --> 00:37:40.869
And sometimes, like I know where I handed over a domain,
753
00:37:40.969 --> 00:37:43.179
that needed to be notarized for some.
754
00:37:44.639 --> 00:37:45.360
Top-level things.
755
00:37:46.219 --> 00:37:47.679
So it could be different for different domains.
756
00:37:47.699 --> 00:37:50.719
Yeah, right. It's because if you don't do this right,
757
00:37:50.880 --> 00:37:53.000
now you're actually a thief, right? You're hacking.
758
00:37:54.000 --> 00:37:54.280
Yeah.
759
00:37:54.380 --> 00:37:57.070
So it's an interesting mix of things. And one of
760
00:37:57.110 --> 00:37:58.510
the reasons I jumped on this immediately when we were
761
00:37:58.530 --> 00:38:01.369
talking about it, Mattias, is I wanted the Europeans' perspective, too,
762
00:38:01.449 --> 00:38:04.210
because I think the rules are different in different countries,
763
00:38:04.230 --> 00:38:05.190
so how you approach these things.
764
00:38:05.769 --> 00:38:08.449
And that's also like Europeans are like, it could be
765
00:38:08.510 --> 00:38:12.630
different because there's different history in countries. In Sweden, we
766
00:38:12.670 --> 00:38:16.630
have very good registries of all people for, for, for.
767
00:38:18.190 --> 00:38:21.070
reasons that happen in the 40s there or after the
768
00:38:21.130 --> 00:38:23.769
40s they don't have that in germany sure that's a
769
00:38:23.809 --> 00:38:26.690
distributed system yeah with their.
770
00:38:26.769 --> 00:38:29.710
Legend well and anything in the eastern block you know
771
00:38:29.750 --> 00:38:31.969
like there's a lot of data problems when you get
772
00:38:32.010 --> 00:38:33.489
to places like poland.
773
00:38:33.230 --> 00:38:35.789
Yeah and then there are things like we're still um
774
00:38:36.309 --> 00:38:38.929
you need to be in person to do things um
775
00:38:39.389 --> 00:38:42.030
and so it's And that's sometimes like if you have
776
00:38:42.050 --> 00:38:44.349
a password manager, it's often much more easier because then
777
00:38:44.369 --> 00:38:45.429
you can do it yourself.
778
00:38:45.489 --> 00:38:45.670
Right.
779
00:38:45.730 --> 00:38:46.050
Fix it.
780
00:38:46.070 --> 00:38:49.369
Yeah. Or because Daniel, like you said, the email problem. Well,
781
00:38:49.409 --> 00:38:53.769
then you can have your someone else be the reset.
782
00:38:53.809 --> 00:38:56.050
But what I like when you have a good support
783
00:38:56.090 --> 00:38:58.530
for it, like would have with the success of something
784
00:38:58.699 --> 00:39:02.059
in in GitHub that someone requiring access and I can
785
00:39:02.139 --> 00:39:05.219
cancel it if it's not valid. But if I'm if
786
00:39:05.280 --> 00:39:09.889
I'm not there, the request won't be canceled. right so
787
00:39:09.909 --> 00:39:12.380
that's the thing where i like it's the or like
788
00:39:12.400 --> 00:39:16.179
the digital services should like if you design an erp
789
00:39:16.239 --> 00:39:17.960
system or something it would be nice if you add
790
00:39:17.980 --> 00:39:20.380
that for a customer support to have something to be
791
00:39:20.420 --> 00:39:24.269
able to request the family part like because there are
792
00:39:24.309 --> 00:39:27.510
services now that are been like even online for 20
793
00:39:27.510 --> 00:39:29.889
years so there's a lot of data and things that
794
00:39:30.050 --> 00:39:34.630
could be like photos or whatever that are and for
795
00:39:34.670 --> 00:39:37.449
good reasons there are privacy rules and privacy laws so
796
00:39:37.489 --> 00:39:40.090
that's why it's important to do this before it happens
797
00:39:40.150 --> 00:39:42.760
because it can be really hard to tell apple to
798
00:39:42.800 --> 00:39:46.840
do something if they don't want to or microsoft or
799
00:39:46.900 --> 00:39:50.719
one of the big players it's um And that's why
800
00:39:50.820 --> 00:39:53.719
I think it's important to have some forethought in there.
801
00:39:54.039 --> 00:39:57.860
Yeah, for sure. Yeah, I think the challenge here is
802
00:39:58.340 --> 00:40:00.860
just I feel like I just need to spend a day.
803
00:40:01.239 --> 00:40:04.320
Look at the landscape of what I've got and think
804
00:40:04.420 --> 00:40:07.039
through what that looks like. No different than updating a will.
805
00:40:08.099 --> 00:40:10.320
And wills need updating. There's no two ways about it.
806
00:40:10.320 --> 00:40:10.739
Yeah, they do.
807
00:40:10.900 --> 00:40:13.179
So what I did, I did a set of markdown
808
00:40:13.199 --> 00:40:15.159
documents in a private Git repository.
809
00:40:15.500 --> 00:40:16.199
You're such a geek.
810
00:40:18.099 --> 00:40:21.630
So what I mean, it's kind of easy because then
811
00:40:21.650 --> 00:40:24.230
I did my first step, like an inventory, a single
812
00:40:24.269 --> 00:40:28.469
document with the index of things. I did set up
813
00:40:28.590 --> 00:40:34.130
emergency access for password managers. So the one contact thing.
814
00:40:34.210 --> 00:40:36.320
An LLM would be very good at this with you.
815
00:40:36.340 --> 00:40:39.219
This would be a fine tool to chat with to
816
00:40:39.260 --> 00:40:39.920
go through this.
817
00:40:40.000 --> 00:40:42.539
Yeah, because I can just pour down my thoughts into it.
818
00:40:42.539 --> 00:40:46.519
And I'm also like, One good thing, like many services
819
00:40:46.559 --> 00:40:48.760
have actually, like you can print a Recover Kit or
820
00:40:48.800 --> 00:40:52.119
a physical thing or a USB drive. Things like Windows,
821
00:40:52.159 --> 00:40:54.639
you can have a BitLocker on a USB key and
822
00:40:54.659 --> 00:40:58.650
have that in your safe. But also like name successors
823
00:40:58.690 --> 00:41:01.590
in services that support it, like GitHub and GitLab support that.
824
00:41:03.110 --> 00:41:05.510
And have that note that I can easily go through.
825
00:41:06.829 --> 00:41:08.730
If my success becomes my enemy, I can just go
826
00:41:08.750 --> 00:41:11.650
through that document and see who I have given access to.
827
00:41:12.510 --> 00:41:15.000
But also things like it could be good to have
828
00:41:16.039 --> 00:41:18.760
enabled things like, so you have auto-renew, so you buy
829
00:41:18.820 --> 00:41:22.590
time for things like domains and things. So it's not
830
00:41:23.289 --> 00:41:26.349
an immediate threat for some things. Like if something's really important,
831
00:41:28.150 --> 00:41:31.010
and that's also like, you can go the organization route
832
00:41:31.050 --> 00:41:33.010
because it's much easier if it's a company or a
833
00:41:33.110 --> 00:41:37.030
GitHub organization or things like that, because that is an
834
00:41:37.420 --> 00:41:39.800
organization people can take over, even if you're not there.
835
00:41:41.110 --> 00:41:43.289
and also like pick a date at least once a
836
00:41:43.289 --> 00:41:46.110
year when you go through this list so that was
837
00:41:46.130 --> 00:41:48.030
like the thing that i set for myself.
838
00:41:48.269 --> 00:41:50.750
And yeah totally this is a i i want to
839
00:41:50.809 --> 00:41:54.050
put a prompt aside that i run like i said
840
00:41:54.179 --> 00:41:56.599
once a year that goes through all of this stuff
841
00:41:57.119 --> 00:42:00.079
and because those tools have access to what i have
842
00:42:00.099 --> 00:42:01.760
been doing will probably even help build the list of
843
00:42:01.820 --> 00:42:03.719
this is where you're out of date right.
844
00:42:03.639 --> 00:42:06.460
Yeah and i think the goal isn't to preserve everything
845
00:42:06.519 --> 00:42:10.730
it's It's make like a shutdown survival for the people
846
00:42:10.769 --> 00:42:12.269
you leave behind.
847
00:42:12.650 --> 00:42:15.250
And a path to what happens first.
848
00:42:15.389 --> 00:42:18.610
Yeah, and you don't want your family to inherit infrastructure.
849
00:42:18.639 --> 00:42:23.539
I think they inherit decisions about it, but they shouldn't
850
00:42:23.599 --> 00:42:27.900
like inherit your Azure web app or something like that.
851
00:42:28.340 --> 00:42:30.980
Getting back to this whole AI harness thing with memory,
852
00:42:31.119 --> 00:42:35.219
I mean, I know that my wife uses Windows Copilot
853
00:42:36.059 --> 00:42:39.139
to help her with her business stuff. And I would,
854
00:42:39.440 --> 00:42:41.619
you know, if something were to happen to her, I
855
00:42:41.639 --> 00:42:44.199
feel like I could just walk up to Copilot and say, okay,
856
00:42:44.239 --> 00:42:47.409
here's what happened. Where do I start? You know, that
857
00:42:47.449 --> 00:42:48.869
could help me with that process.
858
00:42:49.269 --> 00:42:52.250
But actually, that kind of state is a problem because
859
00:42:52.329 --> 00:42:54.170
with the cloud services, you don't know where that lives.
860
00:42:54.230 --> 00:42:57.570
There's data sometimes where, and that you can do, and
861
00:42:57.730 --> 00:43:00.570
that's kind of thanks to ERP, you can actually export
862
00:43:00.630 --> 00:43:03.760
data from services. So that could be one way to
863
00:43:03.820 --> 00:43:10.159
have backup. But think like it's uh it's doing like
864
00:43:10.480 --> 00:43:13.739
also doing things like what am i like you said
865
00:43:13.760 --> 00:43:16.159
your wife told you what you're not using and i
866
00:43:16.179 --> 00:43:17.489
think that's kind of when you do this kind of
867
00:43:17.510 --> 00:43:20.389
inventory you can see like does this domain that is
868
00:43:20.429 --> 00:43:24.789
for like discount sharkages.com that you sell like do i
869
00:43:24.809 --> 00:43:28.219
need it anymore or can i ask not renew it
870
00:43:28.599 --> 00:43:30.940
uh and that's also kind of nice because it's like
871
00:43:31.809 --> 00:43:34.289
the less technical depth you have, because that's actually what
872
00:43:34.309 --> 00:43:37.789
you have. The less you have there, the lighter you are,
873
00:43:37.829 --> 00:43:41.159
I think, in a way. And sometimes I'm the kind
874
00:43:41.179 --> 00:43:42.760
of person, if I just put it in a document,
875
00:43:42.800 --> 00:43:46.000
I can pop my stack and I feel lighter without
876
00:43:46.059 --> 00:43:49.820
having to dwell about it and think too much of it.
877
00:43:50.400 --> 00:43:55.429
And I mean, it's... Yeah, and I think it's okay
878
00:43:55.449 --> 00:43:58.909
for it to take time initially, like, because some things are,
879
00:43:59.510 --> 00:44:03.079
it's sometimes it's like killing your darlings is hard. Like,
880
00:44:04.159 --> 00:44:07.400
like I had the sometimes like, why do I have
881
00:44:07.420 --> 00:44:12.340
this broken, like the oldest Nokia phones from or whatever.
882
00:44:13.639 --> 00:44:15.690
Oh, no, it's your Nokia phone shrine.
883
00:44:15.969 --> 00:44:17.010
Yeah, it's like, I.
884
00:44:16.949 --> 00:44:17.570
Still have mine.
885
00:44:18.099 --> 00:44:22.170
Yeah, and it's like, it's hard to, and that's why
886
00:44:22.230 --> 00:44:24.789
I like to have to, my wife is really good.
887
00:44:24.849 --> 00:44:26.710
You know what? It's going to be easier for your
888
00:44:26.750 --> 00:44:27.250
loved ones.
889
00:44:27.650 --> 00:44:30.369
Yeah, but also like it's, it's sometimes good because my
890
00:44:30.409 --> 00:44:32.320
wife is really good at throwing things away. And that's
891
00:44:32.340 --> 00:44:35.420
kind of like, she has to go to iterations with this.
892
00:44:35.480 --> 00:44:37.380
I throw a little away each time she tells me
893
00:44:37.400 --> 00:44:38.159
to throw something away.
894
00:44:39.179 --> 00:44:41.340
Well, she, I suspect she's even better at throwing away
895
00:44:41.619 --> 00:44:42.179
your things.
896
00:44:42.219 --> 00:44:42.980
She's excellent at it.
897
00:44:43.139 --> 00:44:43.619
Well, yeah.
898
00:44:44.349 --> 00:44:47.630
Oh, you're, you're totally speaking my language, Richard. I wouldn't
899
00:44:47.690 --> 00:44:50.690
call myself a pack rat, but I did come from
900
00:44:50.750 --> 00:44:53.110
a pack rat, right? My mother was a pack rat.
901
00:44:53.610 --> 00:44:55.900
And so I can see, I have tendencies towards that,
902
00:44:55.920 --> 00:44:58.000
but my wife keeps me on the state and narrow. She's,
903
00:44:58.440 --> 00:45:01.519
and her mother is a chucker. Like get rid of this,
904
00:45:01.619 --> 00:45:03.460
get rid of that, get rid of this. Her house
905
00:45:03.500 --> 00:45:06.179
is always like pristine and clean and there's nothing that
906
00:45:06.219 --> 00:45:09.050
doesn't belong there. You know, my studio, I look around,
907
00:45:09.090 --> 00:45:11.150
I got a couple of boxes over here. I got
908
00:45:11.170 --> 00:45:14.750
some stuff over there. You know, and so if it
909
00:45:14.769 --> 00:45:17.909
weren't for my wife, it would be piled probably higher
910
00:45:17.949 --> 00:45:18.369
and deeper.
911
00:45:19.429 --> 00:45:19.809
Yeah.
912
00:45:19.889 --> 00:45:23.530
I mean, like, I think for me, like it, some
913
00:45:23.550 --> 00:45:25.289
things took me a couple of years to do because
914
00:45:25.309 --> 00:45:27.889
you both like you have to procrastinate a bit and
915
00:45:27.929 --> 00:45:30.570
you think like, and, and it's not the optimal system,
916
00:45:30.590 --> 00:45:33.230
but I think it's good to have at least have
917
00:45:33.250 --> 00:45:36.949
the mindset to go for it. And I think it's, um,
918
00:45:37.590 --> 00:45:41.840
it's an even, Sometimes it's so easy. Now I'm more
919
00:45:41.880 --> 00:45:44.119
like cloud storage, but I know when hard drives were.
920
00:45:44.409 --> 00:45:45.909
When I ran out, I just bought a new hard
921
00:45:45.929 --> 00:45:48.090
drive and moved that one aside. So I have a
922
00:45:48.690 --> 00:45:53.110
box of hard drives and things like that. So it's
923
00:45:53.289 --> 00:45:56.889
very easy. It was cheaper because Moore's Law to become
924
00:45:56.909 --> 00:45:57.530
a hoarder.
925
00:45:58.849 --> 00:45:59.010
Yeah.
926
00:46:00.409 --> 00:46:02.150
But I think it's kind of nice.
927
00:46:02.230 --> 00:46:06.090
Well, yeah. It's only recently we've been able to hoard petabytes.
928
00:46:06.630 --> 00:46:06.829
Yeah.
929
00:46:06.869 --> 00:46:09.550
Yeah. Well, I did the thing where I moved all
930
00:46:09.590 --> 00:46:13.630
my hard drive data to my NAS storage and then
931
00:46:14.090 --> 00:46:16.670
drilled a hole in every hard drive and then brought
932
00:46:16.730 --> 00:46:19.420
it to Staples because they can recycle them.
933
00:46:19.699 --> 00:46:20.440
Shred them.
934
00:46:20.480 --> 00:46:20.820
Yeah.
935
00:46:21.460 --> 00:46:23.300
And that's kind of good because then it's final because
936
00:46:23.360 --> 00:46:25.000
it's so easy when you have something stored that you
937
00:46:25.019 --> 00:46:29.059
don't know even if it works. And it probably doesn't.
938
00:46:29.219 --> 00:46:31.199
And that's kind of a problem. Like when you do retrocomputing,
939
00:46:31.219 --> 00:46:33.900
you learn it the hard way that disks don't last forever.
940
00:46:34.019 --> 00:46:38.530
And they are, you are storing on rust. What do
941
00:46:38.550 --> 00:46:39.269
you think? Yeah.
942
00:46:39.710 --> 00:46:43.829
And so I think sometimes it's, we were really good
943
00:46:43.949 --> 00:46:46.869
when we did like tape backups and things and dating myself,
944
00:46:46.929 --> 00:46:49.159
like we knew that things would break. So we moved
945
00:46:49.199 --> 00:46:52.179
things forward. So it's time we had a new backup
946
00:46:52.199 --> 00:46:54.460
strategy for this generation and we moved things forward.
947
00:46:54.480 --> 00:46:55.340
I think we need to do that.
948
00:46:55.380 --> 00:46:59.630
But, but, There was a time when I deliberately ordered
949
00:46:59.750 --> 00:47:03.909
backup tape drives left in a box for the prospect
950
00:47:03.929 --> 00:47:05.809
that that drive will fail and we'll need to migrate
951
00:47:05.849 --> 00:47:07.550
these tapes and we may not be able to buy
952
00:47:07.570 --> 00:47:08.650
the drive.
953
00:47:08.670 --> 00:47:11.780
It doesn't matter if you have that tape if you
954
00:47:11.800 --> 00:47:12.760
don't have a DAT reader.
955
00:47:12.980 --> 00:47:15.719
That's right. I probably can't get one now, so I
956
00:47:16.260 --> 00:47:19.079
keep a spare on the shelf. Then I learned to
957
00:47:19.119 --> 00:47:21.539
keep a spare on a shelf in a different building
958
00:47:21.579 --> 00:47:23.380
when the first building burned down.
959
00:47:24.739 --> 00:47:27.300
And that's why I learned something like I'm not a
960
00:47:27.320 --> 00:47:29.679
hoarder anymore because like the first death clean you did
961
00:47:29.699 --> 00:47:31.900
for someone else, you know, like, yeah, yeah. So much.
962
00:47:32.320 --> 00:47:35.460
And the conflicts that are a wake up call that
963
00:47:35.539 --> 00:47:38.000
are with inheritance and things like, like take care of
964
00:47:38.059 --> 00:47:41.340
that for your, for your loved ones before. And I
965
00:47:41.360 --> 00:47:44.010
think that's why I like to just, just do like
966
00:47:44.030 --> 00:47:48.230
inventory categorize who should have access and who's your successor.
967
00:47:48.650 --> 00:47:50.469
And if you have those kinds of things in, in checks,
968
00:47:50.530 --> 00:47:53.469
I think it's, it's good way. And it's kind of
969
00:47:53.550 --> 00:47:56.750
also like the whole thing with, open source is responsible
970
00:47:56.769 --> 00:48:00.449
because then there's a plan for things, how they will
971
00:48:00.550 --> 00:48:02.019
be maintained along the way for sure.
972
00:48:02.179 --> 00:48:03.820
Also, it's a good idea to do a little bit
973
00:48:03.860 --> 00:48:06.480
at a time, not sit down and think, okay, today
974
00:48:06.500 --> 00:48:09.019
I'm going to solve this problem. No, it's going to
975
00:48:09.039 --> 00:48:12.469
take many days and, you know, hours, an hour here,
976
00:48:12.510 --> 00:48:15.849
an hour there is such a better, way to go
977
00:48:15.869 --> 00:48:16.199
about it.
978
00:48:16.269 --> 00:48:18.190
Well, and you're going to have some hard ones you don't,
979
00:48:18.230 --> 00:48:20.670
you can't immediately decide either. So it's probably good to
980
00:48:20.690 --> 00:48:21.829
let that sit in the back of the head and
981
00:48:21.949 --> 00:48:22.849
digest for a while. Yeah.
982
00:48:23.300 --> 00:48:24.679
But that's the same thing you do in storage. So
983
00:48:24.699 --> 00:48:26.739
like you put a date on a box and if
984
00:48:26.760 --> 00:48:29.280
you haven't opened it by, and they can do the
985
00:48:29.320 --> 00:48:31.480
same digitally, like put this in a date folder.
986
00:48:31.500 --> 00:48:31.880
Yeah.
987
00:48:32.059 --> 00:48:35.420
And if you haven't opened that archive, you should probably
988
00:48:35.460 --> 00:48:35.820
throw it away.
989
00:48:35.840 --> 00:48:35.960
Right.
990
00:48:36.920 --> 00:48:39.570
Who is the guy that Hanselman interviewed? It was getting
991
00:48:39.610 --> 00:48:42.309
stuff done or it might have been getting done or whatever,
992
00:48:42.409 --> 00:48:45.079
but he had something like that. He had really good.
993
00:48:45.630 --> 00:48:49.010
strategies for not being overwhelmed and you know putting post-it
994
00:48:49.050 --> 00:48:50.570
notes on things and you know.
995
00:48:50.489 --> 00:48:52.820
Yeah well i think like the most like especially i
996
00:48:52.840 --> 00:48:55.059
don't know about others but i'm not good at multitasking
997
00:48:55.099 --> 00:48:57.400
i'm good at time slicing so i think that's that's
998
00:48:57.420 --> 00:48:59.840
how you like if you should divide things in things
999
00:48:59.860 --> 00:49:01.739
you can keep in your content like you said with
1000
00:49:01.860 --> 00:49:04.300
ai if you can keep it in your memory then you.
1001
00:49:04.260 --> 00:49:05.820
Will be efficient in it and you will do it
1002
00:49:06.019 --> 00:49:06.840
yeah and.
1003
00:49:06.980 --> 00:49:10.079
If it takes more than More than a Pomodoro time
1004
00:49:10.119 --> 00:49:12.659
to think about it, then it will be lost.
1005
00:49:13.219 --> 00:49:18.329
Yeah, that's right. So I feel like we're coming to
1006
00:49:18.349 --> 00:49:19.909
the end of this conversation. Is there more?
1007
00:49:20.019 --> 00:49:23.659
I think we also want to start a discussion. I
1008
00:49:23.699 --> 00:49:26.880
think we should reflect. And if there are good things,
1009
00:49:26.900 --> 00:49:30.820
because I'm happy to be educated here or be wrong
1010
00:49:30.900 --> 00:49:33.280
or things. I think here we as a community can
1011
00:49:33.320 --> 00:49:36.010
help each other. I think if you have some kind
1012
00:49:36.050 --> 00:49:38.659
of service, I think it's good to think about that
1013
00:49:39.090 --> 00:49:41.900
what's because you will you learn when you're like I
1014
00:49:41.920 --> 00:49:44.739
know I have had e-commerce services like you learn when
1015
00:49:44.760 --> 00:49:47.480
you get the first support ticket about something then when
1016
00:49:47.500 --> 00:49:50.300
you're like you learned about the missing features and usually
1017
00:49:50.320 --> 00:49:53.179
they're not in the beginning but I think this is
1018
00:49:53.219 --> 00:49:58.599
a feature you should think about how to handle one
1019
00:49:58.840 --> 00:50:02.519
because it can be as you say things happen in
1020
00:50:02.559 --> 00:50:06.079
life or things like I had a friend that had
1021
00:50:06.119 --> 00:50:11.289
a stroke which means He's still here, but he's not.
1022
00:50:11.949 --> 00:50:15.889
So there could be different reasons that you have to
1023
00:50:15.949 --> 00:50:21.269
be prepared. But also, I think it's a really thoughtful
1024
00:50:21.349 --> 00:50:23.869
and lovable thing you can do for everyone else.
1025
00:50:23.889 --> 00:50:24.610
I agree, sure.
1026
00:50:25.030 --> 00:50:26.989
And it's a responsible thing. And I think there's something
1027
00:50:27.010 --> 00:50:31.159
that the 20-year-old me wouldn't think of. But if you've
1028
00:50:31.219 --> 00:50:33.920
been through a couple of things in life, you notice
1029
00:50:33.940 --> 00:50:35.199
that it's important.
1030
00:50:35.389 --> 00:50:38.909
Well, when you're 20, you're omnipotent, right?
1031
00:50:38.969 --> 00:50:40.030
Yeah, you're like.
1032
00:50:40.110 --> 00:50:40.730
You're immortal.
1033
00:50:40.889 --> 00:50:45.070
Yeah. And suddenly, like, when you're around 50, you start
1034
00:50:45.090 --> 00:50:49.130
to get afraid of heights and things like that.
1035
00:50:49.210 --> 00:50:52.139
Hey, fall, take a long time to recover from now.
1036
00:50:52.159 --> 00:50:56.179
Nothing's easy. Yeah, that's... We've all been through cleaning up
1037
00:50:56.320 --> 00:50:58.219
after a loved one's passing.
1038
00:50:58.300 --> 00:50:58.440
Yeah.
1039
00:50:58.760 --> 00:51:00.699
And now we're much more aware looking around our own
1040
00:51:00.739 --> 00:51:04.059
rooms going, hmm. And some of those rooms are digital.
1041
00:51:04.219 --> 00:51:07.070
Yeah. And I mean... And that's why I'm like, yeah,
1042
00:51:07.090 --> 00:51:09.750
I think it's important. It's like, who's this important for?
1043
00:51:09.829 --> 00:51:12.289
If it's just me, then it's not important. It's important
1044
00:51:12.469 --> 00:51:14.730
as long as I care about it.
1045
00:51:14.750 --> 00:51:15.769
Yeah, totally.
1046
00:51:15.889 --> 00:51:20.389
But also like to reiterate and to talk to make
1047
00:51:20.429 --> 00:51:26.639
things easier because I think it's a technical debt that
1048
00:51:26.699 --> 00:51:29.019
will burden you or someone else. And I think it's
1049
00:51:29.159 --> 00:51:30.619
better to take it in pieces.
1050
00:51:30.699 --> 00:51:31.360
Nice.
1051
00:51:31.420 --> 00:51:35.019
Yeah. Don't propagate your technical debt on others unnecessarily.
1052
00:51:35.059 --> 00:51:35.739
Yeah, so true.
1053
00:51:36.019 --> 00:51:38.050
Yeah, I love it. Thanks, Matthias. All right.
1054
00:51:38.469 --> 00:51:42.599
Yeah, thank you, Matthias. What good stuff. And thanks for listening,
1055
00:51:42.639 --> 00:51:44.570
and we'll talk to you next week on.
1056
00:51:44.539 --> 00:51:44.989
NET Rocks.
1057
00:52:05.469 --> 00:52:08.170
NET Rocks is brought to you by Franklin's Net and
1058
00:52:08.230 --> 00:52:12.929
produced by Plop Studios, a full-service audio, video, and post-production
1059
00:52:12.989 --> 00:52:16.969
facility located physically in New London, Connecticut, and, of course,
1060
00:52:17.050 --> 00:52:20.650
in the cloud. Online at pwop.com.
1061
00:52:21.380 --> 00:52:29.280
Visit our website at dotnetrocks.com for RSS feeds, downloads, mobile apps, comments,
1062
00:52:29.699 --> 00:52:32.179
and access to the full archives going back to show
1063
00:52:32.219 --> 00:52:36.530
number one, recorded in September 2002 and make sure you
1064
00:52:36.570 --> 00:52:39.760
check out our sponsors they keep us in business now
1065
00:52:39.800 --> 00:52:41.960
go write some code see you next time