Bản chép lời đầy đủ của Controlling your Digital Legacy with Mattias Karlsson

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