1 00:00:01,520 --> 00:00:03,520 Join us as we gather around the hedge, 2 00:00:03,520 --> 00:00:06,500 where we dig into technology, business, and culture 3 00:00:06,559 --> 00:00:08,900 with the finest minds in computer networking. 4 00:00:21,214 --> 00:00:24,654 Good morning, Tom. So you were saying you 5 00:00:24,654 --> 00:00:25,314 were saying 6 00:00:25,989 --> 00:00:27,910 that your wife has been pulling the dead 7 00:00:27,910 --> 00:00:29,850 leaves off the plant. I believe 8 00:00:30,710 --> 00:00:31,929 that is cheating. 9 00:00:32,630 --> 00:00:33,530 Cheating. No. 10 00:00:34,869 --> 00:00:36,710 No. This is grooming. What are you talking 11 00:00:36,710 --> 00:00:37,210 about? 12 00:00:37,590 --> 00:00:39,509 Cheating. Yeah. Is it cheating when you brush 13 00:00:39,509 --> 00:00:41,049 your teeth for us? Come on. 14 00:00:42,765 --> 00:00:45,744 Yeah. You can't say the plant looks beautiful, 15 00:00:46,685 --> 00:00:49,005 and all you're actually doing is taking the 16 00:00:49,005 --> 00:00:50,765 dead parts off. I don't I don't think 17 00:00:50,765 --> 00:00:53,164 it works that way. Well, the thing is 18 00:00:53,164 --> 00:00:55,164 great. I think the plant needs some I 19 00:00:55,164 --> 00:00:57,564 thought the plant needed some landscaping. I think 20 00:00:57,564 --> 00:00:57,780 this 21 00:01:00,100 --> 00:01:00,600 is 22 00:01:01,539 --> 00:01:02,840 okay. It did. 23 00:01:03,460 --> 00:01:04,280 Oh, boy. 24 00:01:04,579 --> 00:01:07,319 This is this is going downhill really fast. 25 00:01:09,939 --> 00:01:12,840 And, eventually, it will re reach 26 00:01:13,564 --> 00:01:15,885 George Michaelson's area of the world if it 27 00:01:15,885 --> 00:01:17,725 keeps going down. It goes down the far 28 00:01:17,725 --> 00:01:18,625 north. Hello. 29 00:01:19,725 --> 00:01:21,984 That's a long way. And we have biosecurity 30 00:01:22,444 --> 00:01:23,885 laws, so I will have to ask you 31 00:01:23,885 --> 00:01:24,545 some questions. 32 00:01:25,085 --> 00:01:27,105 Has that plant been near beekeeping 33 00:01:27,810 --> 00:01:29,909 equipment in the last three months? 34 00:01:32,450 --> 00:01:33,590 No. It has not. 35 00:01:35,650 --> 00:01:37,010 If we ever get this new piece of 36 00:01:37,010 --> 00:01:39,329 property, we're thinking about keeping bees. But, you 37 00:01:39,329 --> 00:01:41,269 know, we're trying to figure out good move. 38 00:01:41,625 --> 00:01:44,105 Yeah. We like we like bees. I like 39 00:01:44,105 --> 00:01:45,484 honey, so I like bees. 40 00:01:46,745 --> 00:01:50,444 Alright. Cool. So that's George. George is in 41 00:01:50,584 --> 00:01:51,084 Australia 42 00:01:51,385 --> 00:01:53,325 some place or another. Who knows where? 43 00:01:54,745 --> 00:01:55,245 And 44 00:01:55,704 --> 00:01:57,250 I I don't know. George doesn't have his 45 00:01:57,250 --> 00:01:59,489 Christmas tree up yet, so I'm not really 46 00:01:59,569 --> 00:02:00,069 no. 47 00:02:00,849 --> 00:02:03,650 We just put our yeah. We we start 48 00:02:03,650 --> 00:02:04,150 now? 49 00:02:05,090 --> 00:02:07,250 Well, we're gonna start early because my wife's 50 00:02:07,250 --> 00:02:09,490 out of town pretty soon. So yeah. Okay. 51 00:02:09,490 --> 00:02:11,135 You You wanna surprise her with the treat 52 00:02:11,135 --> 00:02:12,895 when she comes home and pine needles in 53 00:02:12,895 --> 00:02:15,855 a circle underneath a bare wooden form? No. 54 00:02:15,855 --> 00:02:18,335 She wants no. She yeah. Like, the Charlie 55 00:02:18,335 --> 00:02:21,629 Brown Charlie Brown Christmas special. But, anyway, no. 56 00:02:21,790 --> 00:02:23,629 No. She she wants to help put it 57 00:02:23,629 --> 00:02:24,990 up. So if she wants to help put 58 00:02:24,990 --> 00:02:26,750 it up, it's gotta go up before she 59 00:02:26,750 --> 00:02:30,050 leaves, and she's leaving right around US Thanksgiving. 60 00:02:30,270 --> 00:02:30,770 So 61 00:02:31,069 --> 00:02:32,990 you know? Oh, wow. Being on the road 62 00:02:32,990 --> 00:02:33,730 at Thanksgiving, 63 00:02:34,110 --> 00:02:35,090 is she mad? 64 00:02:35,985 --> 00:02:38,384 Well, we have a granddaughter coming. Oh, well. 65 00:02:38,384 --> 00:02:40,305 And so she so she wants so she 66 00:02:40,305 --> 00:02:41,664 wants to be there for the 67 00:02:42,784 --> 00:02:45,344 Family rituals are important. It's good to do 68 00:02:45,344 --> 00:02:47,685 these things. Yeah. So that's fine. Alright. 69 00:02:47,985 --> 00:02:48,384 So 70 00:02:49,800 --> 00:02:52,199 oh my goodness. Am I that old? Alright. 71 00:02:52,199 --> 00:02:54,539 So let's talk about data centers and cost. 72 00:02:55,159 --> 00:02:55,659 Yep. 73 00:02:56,120 --> 00:02:56,620 So 74 00:02:57,479 --> 00:02:59,580 I kinda wanna come at this at three 75 00:03:00,199 --> 00:03:01,099 different tax, 76 00:03:01,560 --> 00:03:03,159 and they all got a good side and 77 00:03:03,159 --> 00:03:05,424 a bad side. Well, mostly a good side 78 00:03:05,424 --> 00:03:06,484 and a bad side. 79 00:03:06,864 --> 00:03:09,185 Mostly. And there's a techy and there's a 80 00:03:09,185 --> 00:03:10,084 techy part, 81 00:03:10,465 --> 00:03:13,104 and there's an economic part, and there's, like, 82 00:03:13,104 --> 00:03:14,245 a green part. 83 00:03:14,704 --> 00:03:15,204 Okay? 84 00:03:15,745 --> 00:03:17,824 Or and we could just say energy part 85 00:03:17,824 --> 00:03:19,444 because that might be a little more. 86 00:03:20,599 --> 00:03:21,900 But, you know, go ahead, sir. 87 00:03:22,439 --> 00:03:25,000 Energy part of water, but there's water as 88 00:03:25,000 --> 00:03:26,379 well. Yeah. And waste. 89 00:03:27,000 --> 00:03:28,860 Yeah. That's true. Downstream consequences. 90 00:03:29,159 --> 00:03:32,060 There's a bunch of stuff there. Okay. So 91 00:03:32,439 --> 00:03:32,939 DCs, 92 00:03:33,885 --> 00:03:36,044 I mean, they are amazing. Right? When you 93 00:03:36,044 --> 00:03:37,724 think back to the early days of the 94 00:03:37,724 --> 00:03:40,685 Internet, tin cans and string and the way 95 00:03:40,685 --> 00:03:43,564 we glued everything together, the idea We still 96 00:03:43,564 --> 00:03:44,865 had server rooms 97 00:03:45,245 --> 00:03:47,724 in our office buildings, and it was on 98 00:03:48,205 --> 00:03:50,689 and and, you know, the joke at BASF 99 00:03:50,750 --> 00:03:53,229 was that the server room was in the 100 00:03:53,229 --> 00:03:54,909 basement, so you had to put on your 101 00:03:54,909 --> 00:03:56,849 asbestos suit on your way down 102 00:03:57,150 --> 00:03:59,069 because, you know, you were going so deep 103 00:03:59,069 --> 00:04:01,169 into the earth that, like anyway. 104 00:04:01,694 --> 00:04:03,935 So yeah. You've done duty in these places. 105 00:04:03,935 --> 00:04:05,555 You've done the machine room 106 00:04:05,935 --> 00:04:06,995 gig. Yes, sir. 107 00:04:07,455 --> 00:04:10,254 Yeah. They are scary places. There's a lot 108 00:04:10,254 --> 00:04:12,415 of industrial equipment. There's a lot of risk. 109 00:04:12,415 --> 00:04:13,715 There's a lot of noise. 110 00:04:14,550 --> 00:04:17,029 There's potential trip hazards. There are any number 111 00:04:17,029 --> 00:04:19,189 of times I've been in these places thinking, 112 00:04:19,189 --> 00:04:21,189 wow. Did I tell anyone I was in 113 00:04:21,189 --> 00:04:23,750 here? If one of these cabs fell over, 114 00:04:23,750 --> 00:04:24,649 would anybody 115 00:04:25,110 --> 00:04:25,610 know? 116 00:04:26,629 --> 00:04:27,129 But 117 00:04:28,115 --> 00:04:31,095 they exist. Right? So DCs, data centers, 118 00:04:31,555 --> 00:04:32,854 they are absolutely 119 00:04:33,394 --> 00:04:33,894 structurally 120 00:04:34,194 --> 00:04:34,694 central 121 00:04:35,074 --> 00:04:37,475 to constructing a network by point to point 122 00:04:37,475 --> 00:04:40,035 links. We needed them in order to make 123 00:04:40,035 --> 00:04:40,694 the clearinghouses 124 00:04:41,154 --> 00:04:43,189 where people could come into the room, put 125 00:04:43,189 --> 00:04:44,649 their communications equipment, 126 00:04:45,029 --> 00:04:47,610 and exchange data in the room, be peers, 127 00:04:47,910 --> 00:04:48,649 do stuff. 128 00:04:49,029 --> 00:04:50,569 Right? Store servers. 129 00:04:51,110 --> 00:04:53,110 Yeah. Have lots of compute, have lots of 130 00:04:53,110 --> 00:04:55,589 storage. Yeah. It's all very important. The physical 131 00:04:55,589 --> 00:04:56,569 space is important. 132 00:04:57,189 --> 00:05:00,235 So they kinda grew out of a technology 133 00:05:00,294 --> 00:05:01,814 that we were in for a very, very 134 00:05:01,814 --> 00:05:04,214 long time, which is essentially the infrastructure for 135 00:05:04,214 --> 00:05:06,314 telephone exchanges and for electricity 136 00:05:06,615 --> 00:05:07,115 distribution. 137 00:05:07,574 --> 00:05:09,675 Things like gas, water, 138 00:05:10,694 --> 00:05:11,514 heating, cooling, 139 00:05:12,089 --> 00:05:14,670 those things, they ran a slightly different system. 140 00:05:15,050 --> 00:05:17,290 They put equipment in the basement of every 141 00:05:17,290 --> 00:05:19,610 building. They ran trunks and pipes between them. 142 00:05:19,610 --> 00:05:21,050 And there was a grumpy old guy in 143 00:05:21,050 --> 00:05:22,970 a boiler suit with a corncob pipe in 144 00:05:22,970 --> 00:05:25,095 the corner who kind of fed the thing 145 00:05:25,095 --> 00:05:27,334 and kept it going. Mike Mulligan and his 146 00:05:27,334 --> 00:05:29,914 digger working it away. But for telecommunications, 147 00:05:31,014 --> 00:05:32,774 you had a small amount of that stuff 148 00:05:32,774 --> 00:05:34,294 in the bottom of the building, but you 149 00:05:34,294 --> 00:05:35,995 had a lot of stuff in a central 150 00:05:36,055 --> 00:05:38,694 point, the data center. Right? And so the 151 00:05:38,694 --> 00:05:39,194 Internet 152 00:05:39,639 --> 00:05:41,960 coming on board, taking over the AT and 153 00:05:41,960 --> 00:05:44,540 T role, taking over Marbell's role, 154 00:05:45,400 --> 00:05:47,879 occupied that ground. Oh my god. They've got 155 00:05:47,879 --> 00:05:50,860 copper between buildings. Cool. Let's run two megabit 156 00:05:50,920 --> 00:05:51,420 services. 157 00:05:51,784 --> 00:05:54,604 Oh, they use Krone. Okay. We'll use Krone. 158 00:05:55,224 --> 00:05:57,625 So, you know, the technology that we're all 159 00:05:57,625 --> 00:05:59,324 used to using to build out 160 00:05:59,944 --> 00:06:01,724 grew out of the telecommunications 161 00:06:02,264 --> 00:06:04,444 industry right down to the glorious 162 00:06:04,985 --> 00:06:05,485 u, 163 00:06:06,160 --> 00:06:06,899 the unit. 164 00:06:07,199 --> 00:06:10,079 How many u's is your computer? How wide 165 00:06:10,079 --> 00:06:12,160 is a rack? 19 inches. Why is it 166 00:06:12,160 --> 00:06:13,939 19 inches? Because 167 00:06:16,319 --> 00:06:17,300 it just is. 168 00:06:17,600 --> 00:06:20,495 Yeah. What which telco standard should you follow 169 00:06:20,495 --> 00:06:22,275 for your power and your cabling? 170 00:06:22,814 --> 00:06:25,134 And should airflow be front to back or 171 00:06:25,134 --> 00:06:26,355 side to side or 172 00:06:26,895 --> 00:06:30,035 back to front? Hot aisle cola. Yeah. So 173 00:06:30,959 --> 00:06:34,480 these things are absolutely critical, but they were 174 00:06:34,480 --> 00:06:36,879 part of gluing together a point to point 175 00:06:36,879 --> 00:06:38,560 link where if I was dealing with a 176 00:06:38,560 --> 00:06:41,060 company that had my credit card information 177 00:06:41,600 --> 00:06:43,779 and they were an American banking corporation, 178 00:06:44,495 --> 00:06:46,495 we all kind of knew the packets had 179 00:06:46,495 --> 00:06:48,574 to flow from me in Australia down the 180 00:06:48,574 --> 00:06:49,074 transcontinental 181 00:06:49,615 --> 00:06:50,595 links into 182 00:06:51,134 --> 00:06:54,095 Bank of America's headquarter in Phoenix, Arizona and 183 00:06:54,095 --> 00:06:56,574 do stuff. That's how the world worked. Right? 184 00:06:56,574 --> 00:06:59,410 Data was close to the computing center. The 185 00:06:59,410 --> 00:07:01,410 trunk links ran through the DCs, and there 186 00:07:01,410 --> 00:07:02,870 was a mega DC 187 00:07:03,410 --> 00:07:03,910 somewhere 188 00:07:04,529 --> 00:07:06,449 where a bank had their computers. And there 189 00:07:06,449 --> 00:07:07,910 was another one where, 190 00:07:08,610 --> 00:07:11,009 I don't know, the weather forecasting people had 191 00:07:11,009 --> 00:07:13,865 their computers. And there were rumors about Cheyenne 192 00:07:13,925 --> 00:07:16,884 Mountain and enormous computing infrastructure down there to 193 00:07:16,884 --> 00:07:19,944 do missile defense. Everyone had their own right. 194 00:07:21,044 --> 00:07:22,805 That's not what we're really here to talk 195 00:07:22,805 --> 00:07:23,305 about. 196 00:07:24,084 --> 00:07:26,665 We're here to talk about the idea that 197 00:07:27,079 --> 00:07:27,899 a DC 198 00:07:28,279 --> 00:07:29,180 is now 199 00:07:30,040 --> 00:07:30,860 a commodity 200 00:07:31,480 --> 00:07:32,699 construction activity, 201 00:07:33,399 --> 00:07:36,439 and you are probably, for any value of 202 00:07:36,439 --> 00:07:38,060 you, you are probably sitting 203 00:07:38,839 --> 00:07:41,044 at the central point of three of them 204 00:07:41,044 --> 00:07:43,764 within an incredibly short distance of where you 205 00:07:43,764 --> 00:07:45,524 are. The number of people in the world 206 00:07:45,524 --> 00:07:48,324 who live rural and who necessarily are a 207 00:07:48,324 --> 00:07:50,324 long way from a DC is, of course, 208 00:07:50,324 --> 00:07:52,564 enormous. But in the modern Internet, in the 209 00:07:52,564 --> 00:07:53,544 online world, 210 00:07:54,060 --> 00:07:56,139 there actually aren't that many of the Internet 211 00:07:56,139 --> 00:07:58,379 users who are not really close to a 212 00:07:58,379 --> 00:08:01,420 DC. I know that I'm within about 10 213 00:08:01,420 --> 00:08:04,480 kilometers of three right now, and I suspect 214 00:08:05,180 --> 00:08:07,839 almost all of the content I'm consuming 215 00:08:08,300 --> 00:08:10,794 comes out of them. And that has been 216 00:08:10,794 --> 00:08:12,175 a massive shift 217 00:08:12,475 --> 00:08:13,935 in the nature of the network. 218 00:08:14,314 --> 00:08:16,394 We no longer think we're sending our bank 219 00:08:16,394 --> 00:08:16,894 transactions 220 00:08:17,435 --> 00:08:18,334 over to Phoenix. 221 00:08:18,634 --> 00:08:21,035 We think we're sending them to a local 222 00:08:21,035 --> 00:08:23,879 computer running in a DC close by. Right? 223 00:08:23,879 --> 00:08:25,479 Yeah. Yeah. Well, as a matter of fact 224 00:08:25,639 --> 00:08:27,240 and and part of the reason of this 225 00:08:27,240 --> 00:08:29,180 is economics. It used to be 226 00:08:30,600 --> 00:08:31,979 you go back in time 227 00:08:32,600 --> 00:08:35,320 and New York City and Atlanta, for instance, 228 00:08:35,320 --> 00:08:36,539 in Washington DC 229 00:08:36,919 --> 00:08:39,339 had lots and lots of connectivity. Why? 230 00:08:40,064 --> 00:08:41,824 Because they had lots and lots of train 231 00:08:41,824 --> 00:08:42,324 tracks. 232 00:08:44,064 --> 00:08:47,264 Because that was where you laid fiber and 233 00:08:47,504 --> 00:08:50,164 Because rights of way Mhmm. Were expensive. 234 00:08:50,944 --> 00:08:53,570 Exactly. And so you built in places where 235 00:08:53,570 --> 00:08:55,970 there were a lot of train tracks because 236 00:08:55,970 --> 00:08:57,429 you get a lot of connectivity 237 00:08:58,450 --> 00:08:59,269 very quickly 238 00:08:59,970 --> 00:09:01,669 onto those train tracks 239 00:09:02,210 --> 00:09:04,450 now or through those train tracks and through 240 00:09:04,450 --> 00:09:05,669 that existing infrastructure. 241 00:09:06,365 --> 00:09:08,065 And now things have changed 242 00:09:08,764 --> 00:09:09,504 such that 243 00:09:10,524 --> 00:09:13,264 running fiber is cheaper than finding power. 244 00:09:13,565 --> 00:09:16,204 And that's why things in water. Yep. And 245 00:09:16,204 --> 00:09:19,745 that's why data centers themselves are being distributed. 246 00:09:20,444 --> 00:09:22,490 So people think, oh, I live 247 00:09:23,190 --> 00:09:24,250 in Sydney, Nebraska. 248 00:09:24,629 --> 00:09:27,429 Nobody there's no data center around here. We're 249 00:09:27,429 --> 00:09:30,389 too poor. Oh, yes. There is. Because Yep. 250 00:09:30,470 --> 00:09:32,330 Because you have water and power. 251 00:09:33,269 --> 00:09:34,570 So I think the reason 252 00:09:35,164 --> 00:09:37,004 I I think we should make a distinction 253 00:09:37,004 --> 00:09:40,144 though because there are facilities, colocation facilities, 254 00:09:40,605 --> 00:09:41,824 that are more like 255 00:09:42,764 --> 00:09:44,284 they'll be offended by me saying this, but 256 00:09:44,284 --> 00:09:46,044 that's okay. They're more like the the thing 257 00:09:46,044 --> 00:09:47,564 we talked about at the beginning, the the 258 00:09:47,564 --> 00:09:48,924 thing in the in the basement that you 259 00:09:48,924 --> 00:09:50,524 put your your couple of servers in Oh, 260 00:09:50,524 --> 00:09:53,799 yeah. Than mega scale data centers run by 261 00:09:54,019 --> 00:09:56,019 the people who run all of the Internet 262 00:09:56,019 --> 00:09:56,840 now, basically. 263 00:09:57,220 --> 00:09:59,480 Like, you know They still exist. 264 00:10:00,019 --> 00:10:01,379 Yeah. Yeah. You can't you can't put your 265 00:10:01,379 --> 00:10:03,795 server in AWS data center, But but I 266 00:10:03,795 --> 00:10:05,394 think they're they're still there, and I think 267 00:10:05,394 --> 00:10:07,075 we should at least, you know, make the 268 00:10:07,075 --> 00:10:09,315 distinction between the two personnel. Should, and we 269 00:10:09,315 --> 00:10:11,495 should acknowledge them. That localization, 270 00:10:12,355 --> 00:10:14,754 you're right about the power. You're right about 271 00:10:14,754 --> 00:10:16,240 the the cooling, and we're gonna come back 272 00:10:16,240 --> 00:10:18,320 and talk about that. But the other magic 273 00:10:18,320 --> 00:10:20,100 piece here is that we became 274 00:10:20,399 --> 00:10:22,500 incredibly intolerant of delay. 275 00:10:23,120 --> 00:10:24,179 And in a network, 276 00:10:25,039 --> 00:10:27,919 distance is just unavoidable delay. There's no getting 277 00:10:27,919 --> 00:10:29,600 around it. The further you send them, the 278 00:10:29,600 --> 00:10:32,105 longer it takes. And things become trains of 279 00:10:32,105 --> 00:10:35,144 packets. You're building into states where the behavior 280 00:10:35,144 --> 00:10:35,884 of the protocol 281 00:10:36,264 --> 00:10:39,245 is bound up in the delay bandwidth product. 282 00:10:39,304 --> 00:10:41,144 And if you bring all the data that 283 00:10:41,144 --> 00:10:41,884 we're consuming 284 00:10:42,424 --> 00:10:43,804 close to you as a consumer, 285 00:10:45,064 --> 00:10:46,044 life is sweet. 286 00:10:46,610 --> 00:10:48,209 Yeah. And as a matter of fact, we 287 00:10:48,209 --> 00:10:50,129 don't think about it in terms of you 288 00:10:50,129 --> 00:10:50,949 don't just 289 00:10:51,490 --> 00:10:53,669 when you when you hit a website nowadays, 290 00:10:53,730 --> 00:10:56,149 you're not getting that stuff from that server. 291 00:10:56,929 --> 00:10:59,115 You're getting stuff from that server. That's fine. 292 00:10:59,595 --> 00:11:01,514 But it's not just caches. And we always 293 00:11:01,514 --> 00:11:03,195 think of CDNs when we think about this, 294 00:11:03,195 --> 00:11:06,815 but there's also entire advertising networks behind this 295 00:11:07,195 --> 00:11:08,014 that are doing 296 00:11:08,634 --> 00:11:09,134 microsecond 297 00:11:09,834 --> 00:11:10,334 bidding. 298 00:11:11,399 --> 00:11:15,079 And and we're pulling typefaces and fonts and 299 00:11:15,079 --> 00:11:15,579 rendering 300 00:11:15,959 --> 00:11:18,839 them from some website some some server across 301 00:11:18,839 --> 00:11:21,319 the world. And, like, this stuff is all 302 00:11:21,319 --> 00:11:23,319 interconnected now, and it's it's Yeah. And the 303 00:11:23,319 --> 00:11:25,899 delay, like you said, is really bad. 304 00:11:26,945 --> 00:11:28,004 Politics aside, 305 00:11:28,384 --> 00:11:30,865 that's like the good side of the data 306 00:11:30,865 --> 00:11:33,284 industry moving to localized. It's quick. 307 00:11:34,144 --> 00:11:35,125 It's onshore. 308 00:11:36,465 --> 00:11:38,384 You kind of get a warm fuzzy glow 309 00:11:38,384 --> 00:11:40,379 that you're getting things like your neighbors at 310 00:11:40,460 --> 00:11:42,860 the same quality. It's quite equalizing in some 311 00:11:42,860 --> 00:11:45,440 ways. It can be personalized to suit you. 312 00:11:46,139 --> 00:11:48,399 And, Tom, you're right. There are the bespokes 313 00:11:48,460 --> 00:11:51,100 who are letting people put small industry servers, 314 00:11:51,100 --> 00:11:52,860 one or two u of rack in a 315 00:11:52,860 --> 00:11:55,519 facility and run it locally to supply service. 316 00:11:55,884 --> 00:11:57,504 But also the majors 317 00:11:57,884 --> 00:11:59,264 are looking at us saying, 318 00:11:59,964 --> 00:12:01,884 you don't want to own nuts and bolts. 319 00:12:01,884 --> 00:12:03,324 You don't want to have to the hassle 320 00:12:03,324 --> 00:12:05,485 of owning your own. Look. We can build 321 00:12:05,485 --> 00:12:07,644 out an instance of Amazon for you close 322 00:12:07,644 --> 00:12:10,125 by. It's economically viable for us now. We'd 323 00:12:10,125 --> 00:12:11,500 like to do that. How would you feel 324 00:12:11,500 --> 00:12:13,980 about rehoming in one of ours? How do 325 00:12:13,980 --> 00:12:16,299 you feel about sharing a CPU? And you 326 00:12:16,299 --> 00:12:19,820 know what? Everybody's doing it. So there's kind 327 00:12:19,820 --> 00:12:22,000 of a warm fuzzy glow on that. 328 00:12:22,379 --> 00:12:24,700 Especially if your compute is in US East 329 00:12:24,700 --> 00:12:25,200 1. 330 00:12:25,945 --> 00:12:26,445 Well, 331 00:12:26,904 --> 00:12:28,745 I was kind of I was kind of 332 00:12:28,745 --> 00:12:30,985 gonna mention that we've recently had a feeling 333 00:12:30,985 --> 00:12:33,225 that maybe the warm fuzzy glows here aren't 334 00:12:33,225 --> 00:12:34,764 so fuzzy and warm. Right? 335 00:12:36,105 --> 00:12:38,445 Because kinda cold and blue, aren't they? 336 00:12:39,065 --> 00:12:39,725 They are. 337 00:12:40,049 --> 00:12:41,509 Because in parallel 338 00:12:41,970 --> 00:12:42,470 with 339 00:12:43,169 --> 00:12:46,230 everything comes local because the DC came local, 340 00:12:46,529 --> 00:12:49,330 everything went super scale because the big guys 341 00:12:49,330 --> 00:12:50,950 took over the niche market. 342 00:12:51,649 --> 00:12:53,544 Running a DC, as we're probably going to 343 00:12:53,544 --> 00:12:55,544 talk about later, is not a mom and 344 00:12:55,544 --> 00:12:58,365 pop business. Right? It's a major investment. 345 00:12:58,825 --> 00:13:00,764 And the economic scale here 346 00:13:01,225 --> 00:13:03,705 ran to, let's really only have three or 347 00:13:03,705 --> 00:13:06,220 four providers who do this gig. And the 348 00:13:06,220 --> 00:13:06,720 problem 349 00:13:07,980 --> 00:13:09,360 is supply chains, 350 00:13:10,059 --> 00:13:12,059 which you wanna be short. Right? You don't 351 00:13:12,059 --> 00:13:15,740 wanna have 16 intermediary middlemen handling every bit 352 00:13:15,740 --> 00:13:18,059 of your service delivery. You kinda wanna keep 353 00:13:18,059 --> 00:13:18,799 that tight. 354 00:13:19,345 --> 00:13:20,725 Got a little too tight. 355 00:13:21,024 --> 00:13:23,664 And there's dogfooding in the story too. People 356 00:13:23,664 --> 00:13:25,044 like Amazon necessarily 357 00:13:25,664 --> 00:13:27,985 use their own equipment to run the service 358 00:13:27,985 --> 00:13:29,985 that they're selling to you and me and 359 00:13:29,985 --> 00:13:32,370 the companies we work for. Well, part of 360 00:13:32,370 --> 00:13:34,769 their boxes didn't work right because they had 361 00:13:34,769 --> 00:13:37,250 an upgrade moment. And that want of a 362 00:13:37,250 --> 00:13:38,710 nail loss of service 363 00:13:39,170 --> 00:13:42,070 didn't just take down an internal fragment of 364 00:13:42,290 --> 00:13:42,790 AWS. 365 00:13:43,330 --> 00:13:45,590 It took down customer facing infrastructure, 366 00:13:46,365 --> 00:13:46,865 and 367 00:13:47,245 --> 00:13:49,325 a lot of people lost service. Now here 368 00:13:49,325 --> 00:13:51,804 we are a few days later. Oh my 369 00:13:51,804 --> 00:13:52,304 goodness. 370 00:13:52,764 --> 00:13:55,345 Azure have just had a major outage. 371 00:13:56,205 --> 00:13:57,504 And suddenly, 372 00:13:57,965 --> 00:14:00,605 a whole different cohort of the known world 373 00:14:00,605 --> 00:14:03,570 is going, I can't do Okta. 374 00:14:04,029 --> 00:14:05,089 I can't do 375 00:14:05,709 --> 00:14:08,690 three six five. I can't do my email. 376 00:14:09,230 --> 00:14:11,649 And people who've built hybrid, unfortunately, 377 00:14:12,110 --> 00:14:14,449 have now gotten the worst of both worlds. 378 00:14:15,065 --> 00:14:18,264 Right. You think you're getting resilience by having 379 00:14:18,264 --> 00:14:19,084 two provider 380 00:14:19,384 --> 00:14:21,704 construction, but what you actually have done is 381 00:14:21,704 --> 00:14:24,825 you've doubled your surface of exposure, haven't you? 382 00:14:24,825 --> 00:14:25,804 Yeah. Exactly. 383 00:14:27,225 --> 00:14:27,725 So 384 00:14:28,660 --> 00:14:31,779 this component that we've got a DC close 385 00:14:31,779 --> 00:14:32,279 by, 386 00:14:32,740 --> 00:14:34,980 we've got all the promise of cost coming 387 00:14:34,980 --> 00:14:35,480 down, 388 00:14:36,580 --> 00:14:39,860 we've centralized on big providers with Five nine's 389 00:14:39,860 --> 00:14:42,845 SLAs, Although, I think probably from The US 390 00:14:42,845 --> 00:14:45,565 East outage, they just lost their five nines 391 00:14:45,565 --> 00:14:47,824 for a long period of time. 392 00:14:48,365 --> 00:14:50,044 Yep. I'm thinking of those pictures on the 393 00:14:50,044 --> 00:14:51,964 wall, you know, number of days since a 394 00:14:51,964 --> 00:14:55,004 major dinosaur attack on our infrastructure, and there's 395 00:14:55,004 --> 00:14:55,940 a tyrannosaurus 396 00:14:56,639 --> 00:14:58,259 claw marks down the wall. 397 00:14:59,279 --> 00:15:00,419 Did you have that? 398 00:15:02,240 --> 00:15:04,500 I should have. I should have. Yeah. 399 00:15:06,000 --> 00:15:06,980 So the downside 400 00:15:07,475 --> 00:15:10,355 of this it's fast because it's local, it's 401 00:15:10,355 --> 00:15:13,154 reliable because it's operated at scale is that 402 00:15:13,154 --> 00:15:15,235 when it fails, it fails hard. And when 403 00:15:15,235 --> 00:15:17,554 it fails, it fails bad. And the things 404 00:15:17,554 --> 00:15:19,875 that you actually come to rely on, you 405 00:15:19,875 --> 00:15:22,215 have to now think about duplicate and triplicate. 406 00:15:22,720 --> 00:15:25,120 Things get more expensive because you have to 407 00:15:25,120 --> 00:15:26,820 think about building the resilience 408 00:15:27,279 --> 00:15:29,679 you initially didn't think you wanted. You go 409 00:15:29,679 --> 00:15:30,419 from diversity, 410 00:15:30,879 --> 00:15:32,159 like you said. And 411 00:15:33,039 --> 00:15:35,865 Yeah. And and in fact and in fact, 412 00:15:36,105 --> 00:15:36,605 redundancies, 413 00:15:36,985 --> 00:15:38,845 this should go to show us that redundancy 414 00:15:39,065 --> 00:15:40,125 is not resilience. 415 00:15:41,144 --> 00:15:43,245 These are completely different things. 416 00:15:44,105 --> 00:15:46,845 You can redundancy is a component 417 00:15:47,784 --> 00:15:49,164 of a good resilient 418 00:15:49,544 --> 00:15:50,044 design. 419 00:15:50,500 --> 00:15:52,440 But just because something is redundant 420 00:15:53,139 --> 00:15:56,100 does not mean it is resilient. Don't if 421 00:15:56,100 --> 00:15:57,940 you've got those two things together in your 422 00:15:57,940 --> 00:15:58,440 head, 423 00:15:59,300 --> 00:16:02,340 separate them right now because because it's not 424 00:16:02,340 --> 00:16:02,840 true. 425 00:16:04,115 --> 00:16:06,214 So there's another part here, 426 00:16:06,834 --> 00:16:07,575 which is 427 00:16:07,954 --> 00:16:09,654 we've kind of walked into 428 00:16:10,995 --> 00:16:13,394 a political domain. And I'm not gonna get 429 00:16:13,394 --> 00:16:13,894 too 430 00:16:14,434 --> 00:16:16,595 heavy and specific about this because I don't 431 00:16:16,595 --> 00:16:17,495 think we need 432 00:16:17,875 --> 00:16:18,855 the atmosphere. 433 00:16:19,340 --> 00:16:20,080 But there's 434 00:16:20,460 --> 00:16:21,120 data extraterritoriality 435 00:16:24,779 --> 00:16:26,240 about this, that the company 436 00:16:26,860 --> 00:16:30,160 that becomes a major is almost certainly not 437 00:16:30,460 --> 00:16:33,654 a Vietnamese company or a Thai company or 438 00:16:33,654 --> 00:16:35,815 a Singaporean company. It's going to be an 439 00:16:35,815 --> 00:16:38,934 American company because the vast bulk of the 440 00:16:38,934 --> 00:16:40,475 providers that are hyperscaling 441 00:16:40,774 --> 00:16:41,835 to do these things 442 00:16:42,215 --> 00:16:45,195 have come out of the tech venture capital 443 00:16:45,254 --> 00:16:45,754 scene 444 00:16:46,070 --> 00:16:48,789 in America, and they've gone worldwide. And that's 445 00:16:48,789 --> 00:16:49,289 great, 446 00:16:49,990 --> 00:16:52,889 but the problem is their governance structures 447 00:16:53,350 --> 00:16:55,610 reflect the needs of being an American 448 00:16:55,990 --> 00:16:59,669 owned entity. And that doesn't necessarily align with 449 00:16:59,669 --> 00:17:02,264 what the Singaporean government needs in terms of 450 00:17:02,264 --> 00:17:05,544 its data states and integrity of control over 451 00:17:05,544 --> 00:17:08,105 its data. And we're starting to notice that 452 00:17:08,105 --> 00:17:10,105 a lot of economies, and I I do 453 00:17:10,105 --> 00:17:12,505 mean strategic partners of The US, so this 454 00:17:12,505 --> 00:17:13,164 is not 455 00:17:13,544 --> 00:17:16,849 adversarial in that sense, But people like Australia 456 00:17:16,910 --> 00:17:19,630 are saying, you know, there are some things 457 00:17:19,630 --> 00:17:21,869 that maybe we actually have to operate for 458 00:17:21,869 --> 00:17:24,829 ourselves here. And so although a hyperscaler like 459 00:17:24,829 --> 00:17:27,470 Amazon does an amazing service for us, we're 460 00:17:27,470 --> 00:17:29,070 gonna have to build out one that is 461 00:17:29,070 --> 00:17:31,009 just for us, where we buy the machines, 462 00:17:31,115 --> 00:17:33,375 we buy the racks, we buy the cables. 463 00:17:33,994 --> 00:17:35,055 Because otherwise, 464 00:17:35,674 --> 00:17:37,994 what's our level of control over things that 465 00:17:37,994 --> 00:17:39,855 are national strategic interest? 466 00:17:40,394 --> 00:17:40,894 And 467 00:17:41,275 --> 00:17:43,515 a number of people are starting to look 468 00:17:43,515 --> 00:17:44,575 at this and go, 469 00:17:45,200 --> 00:17:47,220 you know, I'm glad you're doing that. 470 00:17:47,839 --> 00:17:49,440 But I've got a bit of a moment 471 00:17:49,440 --> 00:17:52,480 here because I'm living in Google. I'm living 472 00:17:52,480 --> 00:17:55,680 in Amazon hosted services. I'm doing things where 473 00:17:55,680 --> 00:17:58,079 I think I'm subject to Australian laws and 474 00:17:58,079 --> 00:18:01,664 rules. So in Australia, it's totally legal to 475 00:18:01,664 --> 00:18:02,484 do circumvention. 476 00:18:02,785 --> 00:18:04,244 If there's a 477 00:18:04,545 --> 00:18:07,445 IPR lock that is part of the American 478 00:18:07,744 --> 00:18:11,184 intellectual property regime, it's legal in Australia to 479 00:18:11,184 --> 00:18:13,265 say, given the size of our market, our 480 00:18:13,265 --> 00:18:16,000 local laws permit us to circumvent those locks. 481 00:18:16,059 --> 00:18:17,359 They don't apply here. 482 00:18:17,819 --> 00:18:18,960 That's legal 483 00:18:19,259 --> 00:18:21,099 in my economy. If I have bought the 484 00:18:21,099 --> 00:18:23,599 goods, I'm allowed to break the lock. 485 00:18:24,059 --> 00:18:26,875 But if it's hosted on a machine that's 486 00:18:26,875 --> 00:18:29,214 being operated by Amazon, and Amazon 487 00:18:29,515 --> 00:18:32,554 operate Kindle, and Kindle operate the IPR, and 488 00:18:32,554 --> 00:18:34,174 the IPR terms and conditions 489 00:18:34,634 --> 00:18:36,414 are American terms and conditions, 490 00:18:37,275 --> 00:18:39,755 how does Amazon feel about me breaking the 491 00:18:39,755 --> 00:18:40,769 lock on the book? 492 00:18:41,170 --> 00:18:43,089 Yeah. Well, we had this problem with DNS 493 00:18:43,089 --> 00:18:44,710 in the early days, by the way. 494 00:18:45,009 --> 00:18:46,470 And and it's still 495 00:18:46,930 --> 00:18:49,049 it's still an issue in many ways. But 496 00:18:49,170 --> 00:18:49,670 yeah. 497 00:18:50,130 --> 00:18:52,049 But now this is only one component of 498 00:18:52,049 --> 00:18:54,049 what I would consider still to be the 499 00:18:54,049 --> 00:18:54,549 cost. 500 00:18:54,884 --> 00:18:56,805 Right? Oh, yeah. This is this this is 501 00:18:56,805 --> 00:18:59,045 a cost issue that people don't think about 502 00:18:59,045 --> 00:19:01,125 as a cost issue, but it actually is 503 00:19:01,125 --> 00:19:02,025 a cost issue. 504 00:19:02,724 --> 00:19:03,384 Oh, yeah. 505 00:19:03,845 --> 00:19:05,144 And this is cost 506 00:19:05,684 --> 00:19:08,759 in the sense of inside the framework, inside 507 00:19:08,900 --> 00:19:10,759 data, inside information, 508 00:19:11,700 --> 00:19:15,380 inside intellectual property. We're not thinking about political 509 00:19:15,380 --> 00:19:17,799 costs, social costs, legal costs, 510 00:19:18,180 --> 00:19:21,059 risk, consequence. No. There's there's Numbers. No. There's 511 00:19:21,059 --> 00:19:22,440 an economic cost 512 00:19:23,194 --> 00:19:25,994 when a government says, I want my data 513 00:19:25,994 --> 00:19:28,875 stored in my country, my people's data stored 514 00:19:28,875 --> 00:19:30,875 in my country. That's just So there just 515 00:19:30,875 --> 00:19:33,055 is. So so let's go there. 516 00:19:33,515 --> 00:19:35,835 You know, I used to know some lovely 517 00:19:35,835 --> 00:19:37,454 people who worked as geologists. 518 00:19:37,799 --> 00:19:39,420 They were Australian geologists. 519 00:19:39,799 --> 00:19:41,720 They said they did shake and bake. They 520 00:19:41,720 --> 00:19:43,480 preferred to do the shake, but the problem 521 00:19:43,480 --> 00:19:45,320 is after a shake, you get the fire. 522 00:19:45,320 --> 00:19:47,559 So they did shake and bake, like shake 523 00:19:47,559 --> 00:19:48,779 and bake pancakes, 524 00:19:49,320 --> 00:19:50,619 except for its buildings. 525 00:19:51,065 --> 00:19:53,724 And they told me there's really good economic 526 00:19:53,785 --> 00:19:56,585 modeling that all that destructive energy from the 527 00:19:56,585 --> 00:19:58,184 earthquake that you guys are gonna get on 528 00:19:58,184 --> 00:19:59,484 the West Coast Of America 529 00:19:59,785 --> 00:20:01,804 is going to lead to a civil construction 530 00:20:02,025 --> 00:20:02,525 boom. 531 00:20:02,904 --> 00:20:04,700 There are gonna be so many jobs 532 00:20:05,079 --> 00:20:08,119 rebuilding the West Coast when it finally slides 533 00:20:08,119 --> 00:20:09,339 off into the ocean. 534 00:20:09,720 --> 00:20:12,059 It's gonna be great. There's an upside. 535 00:20:12,599 --> 00:20:15,480 And so you go The broken window theory. 536 00:20:15,480 --> 00:20:17,640 Let's all go break everybody's windows so they 537 00:20:17,640 --> 00:20:19,204 have to replace it. So Doesn't 538 00:20:19,664 --> 00:20:20,484 work right. 539 00:20:21,105 --> 00:20:22,565 Doesn't work. But 540 00:20:23,024 --> 00:20:23,924 if you ignore 541 00:20:24,865 --> 00:20:27,845 total loss rebuild and you think about constructing 542 00:20:27,984 --> 00:20:30,865 new investments to supply services in your data 543 00:20:30,865 --> 00:20:33,569 centers, that is economic investment. 544 00:20:33,869 --> 00:20:36,049 And so during the construction phase, 545 00:20:36,349 --> 00:20:39,230 you're employing jobs in civil engineering. There's people 546 00:20:39,230 --> 00:20:41,169 doing grading, then there's concreting, 547 00:20:41,549 --> 00:20:43,650 there's truck drivers doing delivery. 548 00:20:44,134 --> 00:20:46,615 That puts money into the local economy. It's 549 00:20:46,615 --> 00:20:49,095 really well known that during the construction phase, 550 00:20:49,095 --> 00:20:49,994 there's probably 551 00:20:50,694 --> 00:20:52,634 five times or more the jobs 552 00:20:52,934 --> 00:20:55,034 during that phase. And for a small town 553 00:20:55,174 --> 00:20:57,335 in a rural area that hasn't had much 554 00:20:57,335 --> 00:20:57,835 investment, 555 00:20:58,330 --> 00:21:01,130 that can be really massive upside. It's actually 556 00:21:01,130 --> 00:21:02,190 quite a big benefit. 557 00:21:02,730 --> 00:21:05,769 Alongside that, as you've already alluded to, there's 558 00:21:05,769 --> 00:21:09,549 power and cooling costs, which implies upgrading infrastructure. 559 00:21:10,090 --> 00:21:13,044 And again, that can be an amazing benefit 560 00:21:13,044 --> 00:21:14,345 for people in locations 561 00:21:14,804 --> 00:21:16,964 that are getting these services, because they don't 562 00:21:16,964 --> 00:21:17,784 have the population 563 00:21:18,404 --> 00:21:19,944 to pay for sewage, 564 00:21:20,404 --> 00:21:22,664 clean water, potable water, recycling, 565 00:21:23,690 --> 00:21:26,730 electricity supply at that scale. People suddenly go 566 00:21:26,730 --> 00:21:29,289 from one wire delivery of power to full 567 00:21:29,289 --> 00:21:32,569 three three phase power. Maybe they get local 568 00:21:32,569 --> 00:21:34,829 solar. So there are all of these components 569 00:21:35,049 --> 00:21:37,845 of spend that bring jobs into the area, 570 00:21:37,845 --> 00:21:39,865 that bring production into the area. 571 00:21:40,244 --> 00:21:41,785 Really great. It's lovely 572 00:21:42,244 --> 00:21:43,785 if that's what you want. 573 00:21:45,365 --> 00:21:46,105 But then 574 00:21:46,565 --> 00:21:48,025 the construction ends. 575 00:21:50,470 --> 00:21:52,089 Now what are you left with? 576 00:21:54,230 --> 00:21:56,970 How many jobs are there in a DC 577 00:21:57,029 --> 00:21:58,250 if it's running properly? 578 00:21:59,269 --> 00:22:00,710 Yeah. Right. That's what I was gonna say. 579 00:22:00,710 --> 00:22:01,769 This is all temporary. 580 00:22:02,434 --> 00:22:04,835 You and and and it doesn't necessarily mean 581 00:22:04,835 --> 00:22:06,994 you're employing local people because some of those 582 00:22:06,994 --> 00:22:07,974 trades and skills 583 00:22:08,275 --> 00:22:10,054 are very specialized. 584 00:22:10,994 --> 00:22:13,414 Oh, yeah. You have to write the contracts 585 00:22:13,634 --> 00:22:16,835 to require the people doing this to incur 586 00:22:16,835 --> 00:22:19,970 the cost of using your local labor, having 587 00:22:20,029 --> 00:22:20,529 apprenticeships, 588 00:22:20,910 --> 00:22:22,289 employing people locally, 589 00:22:22,590 --> 00:22:25,150 keeping people locally, not flying them in and 590 00:22:25,150 --> 00:22:26,450 out to do the construction. 591 00:22:26,910 --> 00:22:28,830 So we could talk about this as an 592 00:22:28,830 --> 00:22:31,410 effect in, you know, rust belt in America, 593 00:22:31,714 --> 00:22:33,974 but think about economies like Vietnam 594 00:22:34,674 --> 00:22:35,414 or Malaysia 595 00:22:35,954 --> 00:22:39,095 or Cambodia who are looking at potentially supplying 596 00:22:39,154 --> 00:22:40,535 service in their region. 597 00:22:41,075 --> 00:22:43,474 They're coming up from zero. So if people 598 00:22:43,474 --> 00:22:46,190 fly in and do all this construction, whoopee 599 00:22:46,190 --> 00:22:47,889 doo, there's all this local trade, 600 00:22:48,669 --> 00:22:50,669 it's a lot more like a US army 601 00:22:50,669 --> 00:22:52,929 base. You know? People fly in the food. 602 00:22:53,069 --> 00:22:55,549 They fly in the Wendy's from overseas. They 603 00:22:55,549 --> 00:22:56,369 fly in 604 00:22:56,750 --> 00:22:59,394 the product. They fly in the labor, and 605 00:22:59,394 --> 00:23:01,654 they go away again, and you're left with 606 00:23:01,875 --> 00:23:03,414 a cost burden. 607 00:23:03,795 --> 00:23:06,755 Because let's talk about that power and that 608 00:23:06,755 --> 00:23:07,255 water. 609 00:23:07,875 --> 00:23:11,700 How much power does a DC consume? Okay. 610 00:23:11,700 --> 00:23:13,619 I saw the other day now we talked 611 00:23:13,619 --> 00:23:15,859 to somebody before who does power on the 612 00:23:15,859 --> 00:23:17,539 hedge. I don't remember what her name was, 613 00:23:17,539 --> 00:23:19,460 but we did this whole thing. And she 614 00:23:19,460 --> 00:23:21,539 said they were starting to build 250 k 615 00:23:21,539 --> 00:23:21,859 w, 616 00:23:22,434 --> 00:23:25,234 racks. Now I'm hearing that's fairly common now. 617 00:23:25,234 --> 00:23:26,035 250 618 00:23:26,035 --> 00:23:28,355 k w. You have to, like, put this 619 00:23:28,355 --> 00:23:29,095 in context. 620 00:23:30,674 --> 00:23:32,934 Twenty years ago, when I built 621 00:23:33,315 --> 00:23:34,295 a new house 622 00:23:34,914 --> 00:23:35,414 Yep. 623 00:23:36,240 --> 00:23:37,119 We put in 624 00:23:37,759 --> 00:23:40,559 we requested an upgrade from a 100 k 625 00:23:40,559 --> 00:23:42,640 w to one seventy five k w so 626 00:23:42,640 --> 00:23:43,859 we could put up a shop. 627 00:23:45,200 --> 00:23:45,700 Today, 628 00:23:46,000 --> 00:23:47,220 the average house 629 00:23:47,519 --> 00:23:49,460 probably has a one seventy five. 630 00:23:49,804 --> 00:23:51,505 Some smaller ones have a 100. 631 00:23:51,964 --> 00:23:52,464 Townhouses 632 00:23:52,765 --> 00:23:54,525 and stuff like that have a have a 633 00:23:54,525 --> 00:23:55,904 100 k w feed. 634 00:23:56,525 --> 00:23:58,444 You know, like I said, larger houses may 635 00:23:58,444 --> 00:23:59,984 have one seventy five. Yeah. 636 00:24:00,525 --> 00:24:01,964 In the new house I'm trying to build 637 00:24:01,964 --> 00:24:03,565 right now, we're trying to get two fifty 638 00:24:03,565 --> 00:24:04,384 because, again, 639 00:24:04,845 --> 00:24:06,419 we have a shop, and we wanna be 640 00:24:06,419 --> 00:24:08,759 able to run two twenty volt equipment. 641 00:24:09,220 --> 00:24:11,220 Alright. But 250 642 00:24:11,220 --> 00:24:14,599 k w, we're talking about a single rack 643 00:24:15,059 --> 00:24:15,559 consuming 644 00:24:16,019 --> 00:24:17,159 what a house 645 00:24:17,859 --> 00:24:18,359 consumes. 646 00:24:18,740 --> 00:24:21,399 And in some cases, two houses. One rack. 647 00:24:21,894 --> 00:24:22,954 So one rack. 648 00:24:23,494 --> 00:24:26,454 So end of Indiana Jones, they're pushing the 649 00:24:26,454 --> 00:24:28,694 Ark of the Covenant into the warehouse, and 650 00:24:28,694 --> 00:24:30,934 the wheels are going squeak, squeak, squeak, and 651 00:24:30,934 --> 00:24:33,335 you see the warehouse stretching off into the 652 00:24:33,335 --> 00:24:35,914 distance with all those piles of treasure. 653 00:24:36,380 --> 00:24:38,700 Think about a building that size Yep. But 654 00:24:38,700 --> 00:24:40,640 instead of treasure, it's a DC. 655 00:24:41,180 --> 00:24:44,299 That's how many racks there are. There's a 656 00:24:44,299 --> 00:24:46,320 thousand racks in a DC. 657 00:24:47,100 --> 00:24:49,980 They're not counting the total power supply for 658 00:24:49,980 --> 00:24:50,720 that location 659 00:24:51,100 --> 00:24:51,840 in kilowatts. 660 00:24:52,275 --> 00:24:55,734 They're counting it in hundreds of megawatts. Yep. 661 00:24:56,035 --> 00:24:57,654 Exactly. It's a small 662 00:24:57,955 --> 00:24:58,455 town 663 00:24:58,914 --> 00:24:59,974 in power supply. 664 00:25:00,355 --> 00:25:02,515 And and even if a small town. It's 665 00:25:02,515 --> 00:25:04,914 it's a mid sized city almost. And even 666 00:25:04,914 --> 00:25:07,690 if somebody like Amazon comes in and builds 667 00:25:07,690 --> 00:25:10,089 one data center and Akamai comes in and 668 00:25:10,089 --> 00:25:11,289 builds another little, 669 00:25:12,089 --> 00:25:15,369 another little data center for their whatever, and 670 00:25:15,369 --> 00:25:16,750 then, you know, 671 00:25:17,210 --> 00:25:19,210 Facebook comes in and build a little data 672 00:25:19,210 --> 00:25:21,845 center and Google. Well, you might think each 673 00:25:21,845 --> 00:25:24,164 of those are actually small. They're only, you 674 00:25:24,164 --> 00:25:26,164 know, oh, these people are only building 100 675 00:25:26,164 --> 00:25:28,404 rack data centers. You get 10 of them, 676 00:25:28,404 --> 00:25:30,504 you're still building a thousand racks. 677 00:25:30,964 --> 00:25:32,424 Yeah. Right? So 678 00:25:32,724 --> 00:25:33,464 the decision 679 00:25:34,164 --> 00:25:36,264 to say, we're going to direct 680 00:25:36,859 --> 00:25:38,880 400 megawatts of power 681 00:25:39,340 --> 00:25:42,700 into a location to supply data services twenty 682 00:25:42,700 --> 00:25:44,640 four hours a day because these 683 00:25:44,940 --> 00:25:46,799 places are not running 684 00:25:47,100 --> 00:25:50,000 diurnal cycle. They're running constant load. 685 00:25:50,380 --> 00:25:54,015 That's civil infrastructure. It's utility infrastructure. 686 00:25:54,794 --> 00:25:57,034 So that thing you said about getting right 687 00:25:57,034 --> 00:25:59,034 of way down the railway lines to put 688 00:25:59,034 --> 00:25:59,694 the fibres, 689 00:26:00,154 --> 00:26:02,414 you wanna string high voltage cables. 690 00:26:02,714 --> 00:26:05,194 You have to think extremely carefully about the 691 00:26:05,194 --> 00:26:07,549 ground you're going over. There's tree risk. 692 00:26:07,850 --> 00:26:10,809 There's frost risk. There's fire risk. Pacific Gas 693 00:26:10,809 --> 00:26:12,269 and Electricity in California 694 00:26:12,809 --> 00:26:14,970 can tell you a lot of stories about 695 00:26:14,970 --> 00:26:15,630 the problems 696 00:26:16,009 --> 00:26:16,670 with maintaining 697 00:26:16,970 --> 00:26:19,690 these power lines. Right? Yeah. Oh, yeah. Yeah. 698 00:26:19,690 --> 00:26:22,704 Definitely. So This is a decision that has 699 00:26:22,704 --> 00:26:23,204 consequence 700 00:26:23,825 --> 00:26:24,724 for decades. 701 00:26:25,424 --> 00:26:26,644 It's going to involve 702 00:26:27,105 --> 00:26:29,585 affecting the natural environment, and we're gonna come 703 00:26:29,585 --> 00:26:32,085 to that a bit more later. It's decisions 704 00:26:32,144 --> 00:26:34,724 about constructing an infrastructure investment 705 00:26:35,184 --> 00:26:37,950 that will affect the balance of power in 706 00:26:38,329 --> 00:26:40,569 a kind of electrical sense, not in a 707 00:26:40,569 --> 00:26:41,710 political sense, 708 00:26:42,009 --> 00:26:42,750 for decades. 709 00:26:44,009 --> 00:26:44,829 And so 710 00:26:45,769 --> 00:26:46,269 your 711 00:26:46,730 --> 00:26:50,345 domestic consumption pattern is a very well understood 712 00:26:50,345 --> 00:26:52,585 thing that the power companies have been used 713 00:26:52,585 --> 00:26:53,244 to balancing 714 00:26:53,944 --> 00:26:56,044 their generation cycles against. 715 00:26:56,585 --> 00:26:58,204 Constant load is different. 716 00:26:58,585 --> 00:27:00,904 You need a different source of power, and 717 00:27:00,904 --> 00:27:01,964 there are really 718 00:27:02,664 --> 00:27:04,605 kind of two or three choices. 719 00:27:05,400 --> 00:27:07,400 You can go all new agey, and you 720 00:27:07,400 --> 00:27:10,519 can install your own solar and wind and 721 00:27:10,519 --> 00:27:12,299 some kind of storage mechanism. 722 00:27:13,080 --> 00:27:15,880 Which isn't which actually isn't consistent enough. Yeah. 723 00:27:15,880 --> 00:27:17,815 You have to store it. Right? You do. 724 00:27:17,815 --> 00:27:19,894 It's not it's not just the generation. It's 725 00:27:19,894 --> 00:27:21,735 the store. No. And one of the things 726 00:27:21,735 --> 00:27:23,654 that I've heard about is people going out 727 00:27:23,654 --> 00:27:26,295 there and looking for old aluminium smelters that 728 00:27:26,295 --> 00:27:28,634 had at a dam or a power facility 729 00:27:28,695 --> 00:27:31,180 that's end of life. And they're saying, we'd 730 00:27:31,180 --> 00:27:32,940 really like to take that off your hands 731 00:27:32,940 --> 00:27:35,279 because it's absolutely perfect for them. 732 00:27:35,660 --> 00:27:38,220 Yeah. So the the second thing is that 733 00:27:38,220 --> 00:27:40,619 you look at doing something like that, but 734 00:27:40,619 --> 00:27:41,759 you also install 735 00:27:42,220 --> 00:27:43,119 gas turbines. 736 00:27:44,059 --> 00:27:45,815 So these are basically 737 00:27:46,115 --> 00:27:49,575 GE motors exactly the same as the turbines 738 00:27:49,634 --> 00:27:52,355 on a seven eighty seven, except for instead 739 00:27:52,355 --> 00:27:54,595 of pushing a plane, they're connected to a 740 00:27:54,595 --> 00:27:56,994 huge gearbox that connects to a dynamo and 741 00:27:56,994 --> 00:27:58,214 they generate electricity. 742 00:27:58,755 --> 00:28:00,710 But they are exceptionally 743 00:28:01,089 --> 00:28:01,990 fine machines. 744 00:28:02,690 --> 00:28:05,829 They're also very expensive machines. And they're loud. 745 00:28:06,450 --> 00:28:08,849 Oh, they're loud. So if you're the neighbor 746 00:28:08,849 --> 00:28:09,750 within about 747 00:28:10,049 --> 00:28:13,024 a four kilometer horizon of this place, you're 748 00:28:13,024 --> 00:28:15,125 putting up with a lot of noise now. 749 00:28:15,904 --> 00:28:18,784 But even better than this, they're in short 750 00:28:18,784 --> 00:28:19,284 supply, 751 00:28:19,984 --> 00:28:22,704 which means these companies are now bidding against 752 00:28:22,704 --> 00:28:25,105 each other, buying all the supply chain for 753 00:28:25,105 --> 00:28:25,924 these babies 754 00:28:26,625 --> 00:28:27,765 at the same time 755 00:28:28,170 --> 00:28:30,809 that the Texas power supply network is looking 756 00:28:30,809 --> 00:28:34,190 to buy some surplus gas generators to supplement 757 00:28:34,250 --> 00:28:36,910 its power burdens because of losses in system 758 00:28:37,049 --> 00:28:38,590 and to get system resiliency, 759 00:28:39,289 --> 00:28:40,750 those parts aren't available. 760 00:28:41,414 --> 00:28:43,275 You wanna get a thing called a syncom, 761 00:28:43,335 --> 00:28:45,595 a synchronous condenser that regenerates 762 00:28:45,975 --> 00:28:48,695 inertia back into the system just by being 763 00:28:48,695 --> 00:28:49,674 a huge flywheel. 764 00:28:50,134 --> 00:28:52,215 They're on, like, a five year waiting list 765 00:28:52,215 --> 00:28:55,674 now. So the effect of that is 766 00:28:56,049 --> 00:28:57,029 a lot because 767 00:28:57,410 --> 00:28:58,150 the hyperscaler 768 00:28:58,529 --> 00:29:00,869 data centers are buying up the technology 769 00:29:01,250 --> 00:29:03,329 to give you and me access to better 770 00:29:03,329 --> 00:29:03,829 Netflix, 771 00:29:04,289 --> 00:29:05,430 but it's starting 772 00:29:05,890 --> 00:29:06,630 to interfere 773 00:29:07,170 --> 00:29:10,394 with the utility function of general electricity supply. 774 00:29:10,394 --> 00:29:12,734 So we're talking about reliable electricity 775 00:29:13,195 --> 00:29:13,934 for hospitals, 776 00:29:14,634 --> 00:29:18,394 for public transport systems For people's houses. Yeah. 777 00:29:18,394 --> 00:29:19,695 Everything. People's houses. 778 00:29:20,075 --> 00:29:21,775 So these decisions, 779 00:29:22,315 --> 00:29:23,855 I'm gonna have a hyperscaler 780 00:29:24,234 --> 00:29:25,855 build a center near me. 781 00:29:26,200 --> 00:29:29,000 The power side of it is a whole 782 00:29:29,000 --> 00:29:30,299 bag of concern. 783 00:29:31,000 --> 00:29:33,960 So out in left field is something quite 784 00:29:33,960 --> 00:29:34,460 interesting. 785 00:29:34,840 --> 00:29:37,960 There are actually amazing machines that will generate 786 00:29:37,960 --> 00:29:40,840 power 24 by seven, and they don't need 787 00:29:40,840 --> 00:29:41,340 coal. 788 00:29:42,134 --> 00:29:44,234 They actually make less radioactive 789 00:29:44,535 --> 00:29:46,075 waste than coal does, 790 00:29:46,535 --> 00:29:47,515 and they're called 791 00:29:47,974 --> 00:29:49,115 nuclear reactors. 792 00:29:49,894 --> 00:29:50,394 And 793 00:29:50,855 --> 00:29:52,855 I just saw a story I just saw 794 00:29:52,855 --> 00:29:54,555 a story this morning or yesterday. 795 00:29:55,210 --> 00:29:58,410 Microsoft is asking has asked for permitting to 796 00:29:58,410 --> 00:29:59,630 open a closed 797 00:30:00,250 --> 00:30:01,710 nuclear power plant 798 00:30:02,090 --> 00:30:03,850 Yeah. Because they want the whole thing. And 799 00:30:03,850 --> 00:30:04,910 3 Mile Island 800 00:30:05,450 --> 00:30:07,130 has been bought up by some data. I 801 00:30:07,130 --> 00:30:08,650 don't know if it was Microsoft who it 802 00:30:08,650 --> 00:30:11,055 was. And there, they were like, they were 803 00:30:11,055 --> 00:30:13,055 supposed to shut that down. And nope. They're 804 00:30:13,055 --> 00:30:15,474 not shutting it down. They're gonna run it. 805 00:30:16,335 --> 00:30:16,835 So 806 00:30:17,375 --> 00:30:20,434 we are on the cusp of a technology 807 00:30:20,654 --> 00:30:22,734 decision that is going to have a long 808 00:30:22,734 --> 00:30:23,954 term tail here. 809 00:30:24,255 --> 00:30:24,755 Because 810 00:30:25,559 --> 00:30:28,599 to date, nuclear power investment has been somewhat 811 00:30:28,599 --> 00:30:31,259 stalled in the West and proceeding at breakneck 812 00:30:31,399 --> 00:30:32,460 speed in Asia, 813 00:30:32,839 --> 00:30:35,659 but it's centered around traditional large 814 00:30:35,960 --> 00:30:38,359 reactor investment. You make a hole in the 815 00:30:38,359 --> 00:30:41,724 ground and you spend about $15,000,000,000, 816 00:30:41,865 --> 00:30:43,964 and you make a gigawatt of power, 817 00:30:44,345 --> 00:30:46,424 and then you're really happy for the next 818 00:30:46,424 --> 00:30:50,045 fifty years modulo some operational issues and fuel. 819 00:30:50,664 --> 00:30:53,065 Nobody wants to do that anymore. It's very 820 00:30:53,065 --> 00:30:53,565 risky. 821 00:30:53,960 --> 00:30:56,220 If they run over cost, then they're late. 822 00:30:56,279 --> 00:30:58,680 So Rolls Royce and a bunch of other 823 00:30:58,680 --> 00:31:00,140 companies like GE 824 00:31:00,519 --> 00:31:03,640 have proposed building something they call SMRs, small 825 00:31:03,640 --> 00:31:06,599 modular reactors. These are much more like things 826 00:31:06,599 --> 00:31:08,144 that admiral Rickover 827 00:31:08,684 --> 00:31:10,924 put inside tin cans and sent around the 828 00:31:10,924 --> 00:31:13,404 world on a goodwill mission for the US 829 00:31:13,404 --> 00:31:17,005 Navy. They're amazing machines. They're tiny, and they're 830 00:31:17,005 --> 00:31:20,285 even fuel once, operate for fifty years. They're 831 00:31:20,285 --> 00:31:21,585 not gonna make a gigawatt. 832 00:31:22,045 --> 00:31:22,980 They're gonna make 833 00:31:23,779 --> 00:31:26,500 couple 100 megawatts. They're pretty good size for 834 00:31:26,500 --> 00:31:28,819 a data center. They could be really nice. 835 00:31:28,819 --> 00:31:29,640 But, Tom, 836 00:31:30,179 --> 00:31:32,500 is anyone actually sending you a letter through 837 00:31:32,500 --> 00:31:34,419 the mail saying how do you feel about 838 00:31:34,419 --> 00:31:36,944 having a small nuclear reactor up the road? 839 00:31:37,345 --> 00:31:37,845 Yep. 840 00:31:38,384 --> 00:31:40,724 Yep. So they're not asking me 841 00:31:41,184 --> 00:31:43,585 if this is something I feel comfortable with. 842 00:31:43,585 --> 00:31:44,804 And I have to say, 843 00:31:45,105 --> 00:31:46,804 I'm not actually oppositional, 844 00:31:47,105 --> 00:31:48,884 but they're not even talking 845 00:31:49,424 --> 00:31:51,764 about asking. Right? Yeah. 846 00:31:52,359 --> 00:31:55,720 So the agenda of social control that comes 847 00:31:55,720 --> 00:31:58,220 along with the deployment of nuclear technology 848 00:31:59,000 --> 00:32:01,480 does not get discussed very much. You can't 849 00:32:01,480 --> 00:32:03,819 just put one of these babies out there 850 00:32:03,960 --> 00:32:06,679 without building a risk model around what does 851 00:32:06,679 --> 00:32:08,595 that do to us to have one of 852 00:32:08,595 --> 00:32:11,795 these puppies running? Well, they're still regulated by 853 00:32:11,795 --> 00:32:13,234 the NRC, aren't they, at least in The 854 00:32:13,234 --> 00:32:15,555 United States? It's not like they state regulation 855 00:32:15,555 --> 00:32:16,535 because they're private. 856 00:32:17,154 --> 00:32:20,694 No. Indeed. But that regulation includes that for 857 00:32:20,755 --> 00:32:22,934 a circle around them, normal 858 00:32:23,900 --> 00:32:27,500 social constructs around engagement with law enforcement are 859 00:32:27,500 --> 00:32:29,579 radically different. It's the same as the effect 860 00:32:29,579 --> 00:32:30,319 of the borders. 861 00:32:30,619 --> 00:32:32,460 If you live within five miles of the 862 00:32:32,460 --> 00:32:35,179 border in America, your normal civil rights have 863 00:32:35,179 --> 00:32:35,679 been 864 00:32:36,295 --> 00:32:39,195 interrupted in the interests of guardians 865 00:32:39,894 --> 00:32:42,535 of the border being able to do what 866 00:32:42,535 --> 00:32:45,174 they have to do to protect mainstream America. 867 00:32:45,174 --> 00:32:47,755 It's exactly the same with nuclear power plant. 868 00:32:48,660 --> 00:32:51,299 Rights of entry, rights of evidence, rights of 869 00:32:51,299 --> 00:32:51,799 action 870 00:32:52,100 --> 00:32:54,660 completely change in a circle around these things. 871 00:32:54,660 --> 00:32:56,180 And the people who do it are good 872 00:32:56,180 --> 00:32:58,259 people. I'm not saying that we're dealing with 873 00:32:58,259 --> 00:33:00,759 bad people, but the social construct 874 00:33:01,174 --> 00:33:03,575 around how do we conduct our lives is 875 00:33:03,575 --> 00:33:07,095 subtly different when you have nuclear power plants 876 00:33:07,095 --> 00:33:09,994 next door. But we're not talking about it. 877 00:33:10,615 --> 00:33:12,454 Yeah. And that's just one of the costs. 878 00:33:12,454 --> 00:33:13,894 And now, you know, you talk about the 879 00:33:13,894 --> 00:33:15,035 cost of the day center, 880 00:33:15,390 --> 00:33:17,710 That also comes back to economic cost. Like, 881 00:33:17,710 --> 00:33:20,269 it's not cheap to build those small those 882 00:33:20,269 --> 00:33:20,769 SMRs. 883 00:33:21,309 --> 00:33:23,470 And I know. And so, you know, we're 884 00:33:23,470 --> 00:33:24,210 out here 885 00:33:24,589 --> 00:33:26,830 not only and those SMRs, by the way, 886 00:33:26,830 --> 00:33:29,325 also require water. And that's something else we 887 00:33:29,325 --> 00:33:31,805 haven't talked about is that the data center 888 00:33:31,805 --> 00:33:32,205 itself 889 00:33:33,005 --> 00:33:35,744 list. Yeah. The data center itself consumes 890 00:33:36,285 --> 00:33:38,765 huge amounts of water even if it's not 891 00:33:38,765 --> 00:33:39,265 liquid. 892 00:33:39,805 --> 00:33:43,529 Amounts. And and SMRs require water as well 893 00:33:43,529 --> 00:33:45,470 because they're reliant on cooling 894 00:33:45,769 --> 00:33:47,150 from these things. So 895 00:33:47,529 --> 00:33:49,150 let's talk a bit about water. 896 00:33:50,170 --> 00:33:50,670 Water 897 00:33:51,210 --> 00:33:52,670 is really interesting 898 00:33:53,444 --> 00:33:56,404 because we just take it for granted that 899 00:33:56,404 --> 00:33:58,505 it's available in a lot of places. 900 00:33:58,884 --> 00:34:00,825 But when you get into an economy 901 00:34:01,765 --> 00:34:04,724 like, let's say, The Middle East where pure 902 00:34:04,724 --> 00:34:07,065 water is extremely hard to come by, 903 00:34:07,444 --> 00:34:10,079 you have to invest energy to take seawater 904 00:34:10,219 --> 00:34:12,619 that you have access to and convert it 905 00:34:12,619 --> 00:34:14,699 to pure water to use for things like 906 00:34:14,699 --> 00:34:18,319 drinking and flushing toilets and running data centers. 907 00:34:18,380 --> 00:34:20,000 That water is expensive. 908 00:34:20,859 --> 00:34:23,905 So we now have people who are talking 909 00:34:23,905 --> 00:34:24,804 quite seriously 910 00:34:25,184 --> 00:34:28,085 about the booming market for information 911 00:34:28,385 --> 00:34:30,864 services in The Middle East, in Oman, and 912 00:34:30,864 --> 00:34:33,360 Dubai, and Saudi Arabia, and Yemen. 913 00:34:33,680 --> 00:34:34,500 These countries 914 00:34:34,960 --> 00:34:37,780 are totally up for a global Internet experience. 915 00:34:38,239 --> 00:34:40,480 It's part of their social structure now that 916 00:34:40,480 --> 00:34:41,380 people expect 917 00:34:41,760 --> 00:34:44,000 for having put up with a very odd 918 00:34:44,000 --> 00:34:45,059 political economy 919 00:34:45,440 --> 00:34:48,180 that they're getting access to these amazing benefits 920 00:34:48,320 --> 00:34:48,980 of buying 921 00:34:49,494 --> 00:34:49,994 EA. 922 00:34:50,534 --> 00:34:53,894 They own esports. Right? They're not gonna watch 923 00:34:53,894 --> 00:34:56,214 esports that are being played on servers over 924 00:34:56,214 --> 00:34:56,954 in America. 925 00:34:57,414 --> 00:34:59,574 They want to be what an Australian would 926 00:34:59,574 --> 00:35:01,994 colloquially call a loping bastard. 927 00:35:02,489 --> 00:35:05,389 They want those esports servers in racks 928 00:35:05,690 --> 00:35:08,989 where they are in Downtown Dubai, which means 929 00:35:09,130 --> 00:35:12,329 the water supply for a massive city now 930 00:35:12,329 --> 00:35:13,789 has to be put up against 931 00:35:14,090 --> 00:35:16,269 the needs to supply data services 932 00:35:16,570 --> 00:35:18,030 to that massive city. 933 00:35:18,385 --> 00:35:20,545 It's the same gig as the power. This 934 00:35:20,545 --> 00:35:23,204 has got fifty to a hundred year consequence. 935 00:35:24,144 --> 00:35:25,925 Yeah. If it weren't for latency, 936 00:35:26,305 --> 00:35:28,625 and latency is the real killer, but if 937 00:35:28,625 --> 00:35:29,845 it weren't for latency, 938 00:35:30,305 --> 00:35:32,325 I actually worked on a project in Greenland 939 00:35:32,630 --> 00:35:34,150 where we were talking about going up the 940 00:35:34,150 --> 00:35:35,289 West Coast Of Greenland, 941 00:35:35,590 --> 00:35:38,410 and there are these huge, huge mountain, 942 00:35:39,590 --> 00:35:42,230 valleys that go that that fall right into 943 00:35:42,230 --> 00:35:42,969 the ocean. 944 00:35:43,349 --> 00:35:45,429 And there's a huge amount, I mean, millions 945 00:35:45,429 --> 00:35:48,010 of gallons of water melt off every spring. 946 00:35:48,924 --> 00:35:50,924 And they just it just goes right into 947 00:35:50,924 --> 00:35:52,684 the ocean. And they were thinking, well, you 948 00:35:52,684 --> 00:35:55,105 know, we could damn one of those, 949 00:35:56,045 --> 00:35:58,605 and we could not only have millions of 950 00:35:58,605 --> 00:36:00,864 gallons of water to cool data centers, 951 00:36:01,460 --> 00:36:03,159 but we could have millions of gallons 952 00:36:03,699 --> 00:36:04,359 of water 953 00:36:04,659 --> 00:36:05,719 to push through, 954 00:36:06,980 --> 00:36:07,480 hydroelectric 955 00:36:07,859 --> 00:36:10,659 dam, and we could build data centers, and 956 00:36:10,659 --> 00:36:12,980 it wouldn't really cost us a lot other 957 00:36:12,980 --> 00:36:14,359 than building the dam 958 00:36:14,744 --> 00:36:17,384 and and blasting out a flat spot. Because 959 00:36:17,384 --> 00:36:19,384 there's no flat spots. I mean, there's nothing 960 00:36:19,625 --> 00:36:21,545 There's a lot of people in Iceland who 961 00:36:21,545 --> 00:36:23,704 are thinking very much the same. We have 962 00:36:23,704 --> 00:36:26,264 glaciers and we have geothermal power. Why don't 963 00:36:26,264 --> 00:36:28,424 we operate as a data center for the 964 00:36:28,424 --> 00:36:31,940 Northern sphere? And people are going, yeah. But 965 00:36:32,320 --> 00:36:34,239 No. The problem the problem with, like seconds 966 00:36:34,239 --> 00:36:34,900 of delay. 967 00:36:35,280 --> 00:36:36,960 Yeah. But it's not only that, but Iceland 968 00:36:36,960 --> 00:36:37,599 has a lot of, 969 00:36:39,519 --> 00:36:40,019 geoactivity. 970 00:36:40,400 --> 00:36:42,400 They hit a lot of little bitty earthquakes, 971 00:36:42,400 --> 00:36:44,355 and people don't realize that. Because you don't 972 00:36:44,355 --> 00:36:46,434 feel so cool. Hey. Yeah. That's not so 973 00:36:46,434 --> 00:36:48,434 good for a data center. It's much much 974 00:36:48,434 --> 00:36:49,335 harder on 975 00:36:50,114 --> 00:36:52,594 a hard drive, a spinning hard drive, or 976 00:36:52,594 --> 00:36:54,135 even electronics connectivity 977 00:36:54,755 --> 00:36:56,355 than it is. You can't even feel it 978 00:36:56,355 --> 00:36:58,880 walking across the ground, but there are hundreds 979 00:36:58,880 --> 00:37:00,579 of these little tremors all the time. 980 00:37:00,880 --> 00:37:03,140 So yeah. So this but but the delay 981 00:37:03,200 --> 00:37:05,440 and all of those kinds of environmental factors 982 00:37:05,440 --> 00:37:06,900 trying to build in those areas, 983 00:37:07,599 --> 00:37:10,464 it's just it's just very hard. What we 984 00:37:10,464 --> 00:37:13,264 basically need is we need to convert all 985 00:37:13,264 --> 00:37:14,484 networking to quantum. 986 00:37:15,425 --> 00:37:15,905 So the 987 00:37:18,385 --> 00:37:19,284 No more latency. 988 00:37:19,664 --> 00:37:20,565 No more latency. 989 00:37:21,505 --> 00:37:23,585 I already told you that. And the reverse 990 00:37:23,585 --> 00:37:25,920 of it. You got both. Pick one. What 991 00:37:25,920 --> 00:37:26,739 do I care? 992 00:37:27,119 --> 00:37:27,840 But I wonder 993 00:37:28,639 --> 00:37:29,380 No. Yes. 994 00:37:30,159 --> 00:37:31,859 So it's so it's a little bit ridiculous, 995 00:37:31,920 --> 00:37:33,760 but what would happen if there was no 996 00:37:33,760 --> 00:37:34,260 latency? 997 00:37:35,119 --> 00:37:38,135 What centralization would happen? And, you know, people 998 00:37:38,135 --> 00:37:41,195 wouldn't be considering Iceland anymore. Right? Like No. 999 00:37:42,135 --> 00:37:42,635 So 1000 00:37:43,014 --> 00:37:43,514 this 1001 00:37:44,534 --> 00:37:47,494 is something that an economist might call a 1002 00:37:47,494 --> 00:37:48,635 negative externality 1003 00:37:49,014 --> 00:37:50,809 if I have the jargon right, that you 1004 00:37:50,809 --> 00:37:52,650 make all these plans, then it would make 1005 00:37:52,650 --> 00:37:54,570 a lot of sense to do this build 1006 00:37:54,570 --> 00:37:55,949 out and do something. 1007 00:37:56,489 --> 00:37:58,409 Yeah. I'm gonna build a data center, and 1008 00:37:58,409 --> 00:38:00,650 I'm gonna build it to take GPUs from 1009 00:38:00,650 --> 00:38:02,909 NVIDIA, and I'm going to run a massive 1010 00:38:03,130 --> 00:38:05,610 AI farm in this data center, and I'm 1011 00:38:05,610 --> 00:38:07,445 gonna build it in Malaysia 1012 00:38:08,065 --> 00:38:08,565 because 1013 00:38:08,945 --> 00:38:11,585 they have an amazing tax offset right now. 1014 00:38:11,585 --> 00:38:13,505 I've done a deal with a landholder. I'm 1015 00:38:13,505 --> 00:38:15,025 gonna make a lot of money all the 1016 00:38:15,025 --> 00:38:17,184 way down the line. You get six months 1017 00:38:17,184 --> 00:38:19,765 of the way in, and NVIDIA say, 1018 00:38:21,549 --> 00:38:22,989 I'm really sorry to tell you this, but 1019 00:38:22,989 --> 00:38:25,069 the Chinese did this thing called DeepSeek, and 1020 00:38:25,069 --> 00:38:26,909 they can run the same model we sold 1021 00:38:26,909 --> 00:38:28,750 you but on a tenth of the capital 1022 00:38:28,750 --> 00:38:30,750 cost. And a guy up the road has 1023 00:38:30,750 --> 00:38:33,710 actually done that. He's already built out because 1024 00:38:33,710 --> 00:38:36,175 it's a tenth size, tenth the power budget. 1025 00:38:36,554 --> 00:38:38,175 He stole your lunch, dude. 1026 00:38:38,795 --> 00:38:40,474 Yeah. Well, this is the other yeah. This 1027 00:38:40,474 --> 00:38:41,534 is another economic 1028 00:38:41,914 --> 00:38:43,714 and social cost we don't think about Yeah. 1029 00:38:43,835 --> 00:38:46,335 Because we are building permanent infrastructure 1030 00:38:47,150 --> 00:38:48,369 around specific 1031 00:38:48,829 --> 00:38:50,210 task and specific 1032 00:38:50,750 --> 00:38:52,130 ways of building networks. 1033 00:38:52,989 --> 00:38:55,090 Those that infrastructure is not going away. 1034 00:38:55,390 --> 00:38:57,230 I mean, you can't convert that into a 1035 00:38:57,230 --> 00:38:59,010 mall and go shopping in it. 1036 00:38:59,309 --> 00:39:01,010 That is the externality. 1037 00:39:01,635 --> 00:39:03,494 Because when somebody decides 1038 00:39:03,875 --> 00:39:05,795 that the deal isn't going to work, but 1039 00:39:05,795 --> 00:39:08,275 they've already clear felled the forest, laid the 1040 00:39:08,275 --> 00:39:11,175 concrete floor, built the trunking, built the utility, 1041 00:39:11,715 --> 00:39:13,894 and then they walk away from the deal, 1042 00:39:14,355 --> 00:39:15,655 what they leave behind 1043 00:39:15,954 --> 00:39:19,130 isn't their problem. It's somebody else's problem. 1044 00:39:19,429 --> 00:39:21,609 And we are actually already 1045 00:39:21,909 --> 00:39:24,409 beginning to see signs of this in Southeast 1046 00:39:24,469 --> 00:39:27,530 Asia. It's a growth market. It's the logical 1047 00:39:27,589 --> 00:39:29,750 place for a lot of data centers to 1048 00:39:29,750 --> 00:39:32,389 build out because it's got very cheap cost 1049 00:39:32,389 --> 00:39:33,050 of construction, 1050 00:39:33,724 --> 00:39:35,885 and it's got very strong potential for a 1051 00:39:35,885 --> 00:39:37,905 future market. But the problem is 1052 00:39:38,285 --> 00:39:39,905 some locations like Singapore 1053 00:39:40,364 --> 00:39:43,644 have already exceeded their foreseeable power budget, and 1054 00:39:43,644 --> 00:39:44,465 they passed 1055 00:39:44,925 --> 00:39:45,664 new legislation 1056 00:39:46,525 --> 00:39:49,190 basically saying no more DCs. 1057 00:39:49,569 --> 00:39:50,069 And 1058 00:39:50,449 --> 00:39:53,250 then people who bought land on the prospect 1059 00:39:53,250 --> 00:39:55,650 of a future DC sale with some joint 1060 00:39:55,650 --> 00:39:56,550 venture partners 1061 00:39:56,929 --> 00:39:59,329 are sitting there wondering, what exactly am I 1062 00:39:59,329 --> 00:40:00,949 going to do with this asset? 1063 00:40:01,344 --> 00:40:04,724 So then there's hungry developers looking at Malaysia, 1064 00:40:04,864 --> 00:40:05,925 say, or Vietnam, 1065 00:40:06,864 --> 00:40:09,445 buying into the idea of a DC build, 1066 00:40:09,585 --> 00:40:12,305 and they make an investment there. Okay. Let's 1067 00:40:12,305 --> 00:40:14,565 talk a little bit about power in Vietnam. 1068 00:40:14,625 --> 00:40:17,480 What is Vietnam's primary source of electricity 1069 00:40:17,780 --> 00:40:20,599 power? It's not hydro. It's coal. 1070 00:40:21,699 --> 00:40:24,280 Yeah. So Yeah. You are now moving 1071 00:40:24,900 --> 00:40:28,360 a decision about megawatts of power generation 1072 00:40:28,660 --> 00:40:30,920 into an economy that is going to perpetuate 1073 00:40:31,884 --> 00:40:34,224 coal market, which is great for my economy 1074 00:40:34,444 --> 00:40:36,625 because I get to sell them the coal. 1075 00:40:37,085 --> 00:40:38,605 But we were kind of on a road 1076 00:40:38,605 --> 00:40:41,164 map to try and get out of coal 1077 00:40:41,164 --> 00:40:41,664 production. 1078 00:40:43,164 --> 00:40:45,244 Oh, that's no problem, you say. Let's do 1079 00:40:45,244 --> 00:40:45,904 an SMR. 1080 00:40:46,280 --> 00:40:48,699 I'm sorry. You cannot strategically 1081 00:40:49,079 --> 00:40:49,579 sell 1082 00:40:49,960 --> 00:40:50,460 SMRs 1083 00:40:50,920 --> 00:40:51,420 into 1084 00:40:51,719 --> 00:40:52,619 that economy. 1085 00:40:53,000 --> 00:40:55,960 You can't. It it also impacts connectivity in 1086 00:40:55,960 --> 00:40:58,454 ways we don't think about this costing model 1087 00:40:58,535 --> 00:41:00,695 and where we build things and where the 1088 00:41:00,695 --> 00:41:02,315 power is. Right? Because 1089 00:41:02,855 --> 00:41:04,635 when I've worked for multiple providers, 1090 00:41:06,215 --> 00:41:08,155 operators, whatever you wanna say, 1091 00:41:08,454 --> 00:41:11,675 they have always chosen Singapore or Hong Kong 1092 00:41:12,110 --> 00:41:14,530 as a way a gateway into China 1093 00:41:14,910 --> 00:41:16,530 because you don't want to necessarily 1094 00:41:16,910 --> 00:41:19,550 have a drop point. It's like it's like 1095 00:41:19,550 --> 00:41:20,929 India has its weirdness 1096 00:41:21,309 --> 00:41:23,630 that you can go through India cheaply, but 1097 00:41:23,630 --> 00:41:25,730 you cannot terminate circuits in India. 1098 00:41:26,114 --> 00:41:27,715 You can, but they cost a lot of 1099 00:41:27,715 --> 00:41:28,215 money. 1100 00:41:28,755 --> 00:41:31,155 You can't build a fiber that terminates in 1101 00:41:31,155 --> 00:41:33,635 US territories like Guam or on the West 1102 00:41:33,635 --> 00:41:36,434 Coast. Yeah. Huawei is a venture capital, and 1103 00:41:36,434 --> 00:41:39,130 you cannot terminate it in Hong Kong. You 1104 00:41:39,130 --> 00:41:42,090 can't. Right. Exactly. And so you have these 1105 00:41:42,090 --> 00:41:44,409 you have these weird things like this. And 1106 00:41:44,409 --> 00:41:46,489 Singapore was always this place where you could 1107 00:41:46,489 --> 00:41:47,949 go. You could get to China. 1108 00:41:48,409 --> 00:41:50,989 You could get to the entire Asian Pacific. 1109 00:41:51,449 --> 00:41:54,164 It's an intrapot in every sense. But unfortunately, 1110 00:41:55,025 --> 00:41:57,984 Singapore can't afford to host the DCs that 1111 00:41:57,984 --> 00:41:59,764 people would need to have there 1112 00:42:00,065 --> 00:42:02,704 to use it to provide this service. The 1113 00:42:02,704 --> 00:42:06,065 markets are having to move, which means fifty 1114 00:42:06,065 --> 00:42:08,829 year plans about cable routing and investment in 1115 00:42:08,829 --> 00:42:10,530 power and investment in water 1116 00:42:10,909 --> 00:42:14,510 are moving around. They're sloshing around like giant 1117 00:42:14,510 --> 00:42:16,289 buckets of money and consequence, 1118 00:42:17,630 --> 00:42:19,489 and you or I as consumers 1119 00:42:20,445 --> 00:42:22,844 don't actually have an awful lot to say 1120 00:42:22,844 --> 00:42:23,585 in this. 1121 00:42:24,285 --> 00:42:27,724 Yeah. Yeah. We don't, honestly. Even many operators 1122 00:42:27,724 --> 00:42:29,565 don't have anything to say in this because 1123 00:42:29,565 --> 00:42:31,105 it's all on the real estate 1124 00:42:31,565 --> 00:42:34,204 and the and the power side, power, water, 1125 00:42:34,204 --> 00:42:36,640 and real estate side. So if you're somebody 1126 00:42:36,640 --> 00:42:39,460 like Akamai or Facebook or LinkedIn or whoever, 1127 00:42:39,760 --> 00:42:42,159 Microsoft, whatever it is, you're just trying to 1128 00:42:42,159 --> 00:42:44,019 play the market. You don't have any. 1129 00:42:44,320 --> 00:42:46,079 Like, you can't say, oh, I'm gonna open 1130 00:42:46,079 --> 00:42:47,539 in Singapore this year. 1131 00:42:47,920 --> 00:42:49,380 Okay. Good luck with that. 1132 00:42:50,094 --> 00:42:50,594 Because, 1133 00:42:50,894 --> 00:42:51,555 you know, 1134 00:42:52,255 --> 00:42:53,235 no. You're not. 1135 00:42:55,215 --> 00:42:56,275 So in the margins, 1136 00:42:57,135 --> 00:43:00,035 because I don't like only bad news stories, 1137 00:43:00,175 --> 00:43:02,639 there are actually some good news stories that 1138 00:43:02,639 --> 00:43:04,400 come out of things like that. You being 1139 00:43:04,400 --> 00:43:05,139 a positive 1140 00:43:05,440 --> 00:43:05,940 Australian. 1141 00:43:06,480 --> 00:43:08,719 I know. I know. I just I can't 1142 00:43:08,719 --> 00:43:10,719 do doom and gloom, Russ. It's just not 1143 00:43:10,719 --> 00:43:11,619 in my making. 1144 00:43:12,239 --> 00:43:12,739 So 1145 00:43:13,359 --> 00:43:16,260 there aren't many jobs, but there are jobs. 1146 00:43:16,875 --> 00:43:19,594 And those jobs during the construction phase leave 1147 00:43:19,594 --> 00:43:23,434 behind people who've got amazing skills in rapid 1148 00:43:23,434 --> 00:43:24,974 construction, tilt slab, 1149 00:43:25,355 --> 00:43:29,054 power engineering, driving. It actually does leave behind 1150 00:43:29,550 --> 00:43:32,430 capacity and capability for the local economy to 1151 00:43:32,430 --> 00:43:35,489 do other things. So people start building piggeries. 1152 00:43:35,630 --> 00:43:36,769 People start building 1153 00:43:37,309 --> 00:43:39,789 other things that consume the margins of this 1154 00:43:39,789 --> 00:43:42,670 power that wouldn't have been economically feasible for 1155 00:43:42,670 --> 00:43:45,170 them before because the power wasn't being delivered. 1156 00:43:45,525 --> 00:43:46,984 There are side benefits. 1157 00:43:48,005 --> 00:43:48,744 There's also 1158 00:43:49,284 --> 00:43:53,045 the data component that as more places build 1159 00:43:53,045 --> 00:43:53,944 out DCs, 1160 00:43:54,324 --> 00:43:56,744 it's going to get more common for governments, 1161 00:43:56,885 --> 00:43:57,864 local governments, 1162 00:43:58,324 --> 00:43:58,824 communities 1163 00:43:59,204 --> 00:43:59,864 to say, 1164 00:44:00,219 --> 00:44:02,559 I really want this within my own territory. 1165 00:44:03,099 --> 00:44:04,000 And the hyperscalers 1166 00:44:04,780 --> 00:44:07,180 may not be the ones, but somebody's gonna 1167 00:44:07,180 --> 00:44:08,800 stick their hand up and say, 1168 00:44:09,340 --> 00:44:11,500 okay. I have a price point to hone 1169 00:44:11,500 --> 00:44:13,840 your data on machines you can understand 1170 00:44:14,285 --> 00:44:16,625 that are within your geopolitical control. 1171 00:44:16,925 --> 00:44:19,644 If that's what you want, this technology has 1172 00:44:19,644 --> 00:44:20,465 been commoditized. 1173 00:44:21,244 --> 00:44:23,585 Tell me the supply chains that you're comfortable 1174 00:44:23,644 --> 00:44:26,364 with. For Americans, it's gonna be Dell or 1175 00:44:26,364 --> 00:44:28,070 it's going to be whoever. 1176 00:44:28,369 --> 00:44:30,930 Maybe for Chinese backed entities, it's gonna be 1177 00:44:30,930 --> 00:44:32,930 a different source of product going in those 1178 00:44:32,930 --> 00:44:35,170 racks. That's okay. It's the nature of the 1179 00:44:35,170 --> 00:44:37,829 beast. But the localized control story, 1180 00:44:38,130 --> 00:44:40,505 I think, is going to get better. And 1181 00:44:40,505 --> 00:44:44,025 I think the RTT story about everybody wants 1182 00:44:44,025 --> 00:44:45,005 data local, 1183 00:44:45,385 --> 00:44:47,224 I don't think that's wrong. I think it's 1184 00:44:47,224 --> 00:44:50,445 true. I I didn't need my finance transactions 1185 00:44:50,664 --> 00:44:52,744 going all the way to Switzerland and coming 1186 00:44:52,744 --> 00:44:53,485 back again. 1187 00:44:54,179 --> 00:44:56,260 So there is a component of this where 1188 00:44:56,260 --> 00:44:56,920 the resilience 1189 00:44:57,619 --> 00:44:59,539 sort of comes back into the system if 1190 00:44:59,539 --> 00:45:02,260 you build it right. Because if power goes 1191 00:45:02,260 --> 00:45:02,760 out 1192 00:45:03,059 --> 00:45:04,920 at the wrong end of a long line, 1193 00:45:05,219 --> 00:45:06,440 you have no recourse. 1194 00:45:07,255 --> 00:45:09,494 When power goes out close to you and 1195 00:45:09,494 --> 00:45:11,655 you can't do things, if it's taken out 1196 00:45:11,655 --> 00:45:12,875 your local DC, 1197 00:45:13,175 --> 00:45:14,934 it's kind of the island of you and 1198 00:45:14,934 --> 00:45:17,655 the DC that are offline. The other people 1199 00:45:17,655 --> 00:45:20,155 who have other DCs to operate from 1200 00:45:20,530 --> 00:45:23,730 probably get continuing service. So we talked about 1201 00:45:23,730 --> 00:45:25,269 US East going offline. 1202 00:45:25,970 --> 00:45:27,750 It didn't affect everything. Right? 1203 00:45:28,289 --> 00:45:31,010 Same with this Azure routing. It was outage. 1204 00:45:31,010 --> 00:45:33,329 It was it was local effect. It was 1205 00:45:33,329 --> 00:45:35,240 a little bit distributed because these guys don't 1206 00:45:35,240 --> 00:45:36,309 necessarily follow boundaries, and sometimes things have 1207 00:45:48,155 --> 00:45:50,440 political side of this, one thing I would 1208 00:45:50,440 --> 00:45:52,539 want to be doing or saying is 1209 00:45:53,160 --> 00:45:55,000 if you're gonna build a data center in 1210 00:45:55,000 --> 00:45:55,739 my area, 1211 00:45:56,119 --> 00:45:58,599 I want the physical facilities to be at 1212 00:45:58,599 --> 00:45:59,900 least somewhat fungible. 1213 00:46:00,599 --> 00:46:03,180 I want them to be usable for something. 1214 00:46:03,565 --> 00:46:06,125 If you decide to drop out, I don't 1215 00:46:06,125 --> 00:46:09,025 want you building a building where the doorways 1216 00:46:09,164 --> 00:46:09,905 are specifically 1217 00:46:10,765 --> 00:46:14,605 designed to fit this particular NVIDIA processor. As 1218 00:46:14,605 --> 00:46:15,819 stupid as that might sound, you know 1219 00:46:16,699 --> 00:46:19,179 Oh, no. That's real. These doorways have to 1220 00:46:19,179 --> 00:46:21,659 have enough system strength that if I convert 1221 00:46:21,659 --> 00:46:23,039 it to repairing buses, 1222 00:46:23,420 --> 00:46:25,019 I can fit them in the doorway and 1223 00:46:25,019 --> 00:46:27,819 drive them on the concrete floor. Exactly. Yes. 1224 00:46:27,819 --> 00:46:29,684 There have to be there has to be 1225 00:46:29,684 --> 00:46:30,585 some fungibility, 1226 00:46:30,965 --> 00:46:31,785 some amount 1227 00:46:32,565 --> 00:46:33,305 of flexibility 1228 00:46:34,244 --> 00:46:36,244 built into these things. And, yes, that's gonna 1229 00:46:36,244 --> 00:46:38,905 cost more upfront. I know that. That's okay. 1230 00:46:39,285 --> 00:46:41,445 Yeah. But that's that's the that's that's the 1231 00:46:41,445 --> 00:46:43,980 social trade off that this So these up 1232 00:46:44,059 --> 00:46:44,960 front costs 1233 00:46:45,420 --> 00:46:47,900 aren't that dissimilar to the kinds of costs 1234 00:46:47,900 --> 00:46:50,639 that people doing green power build outs routinely 1235 00:46:50,780 --> 00:46:52,780 have to talk about. You want to put 1236 00:46:52,780 --> 00:46:55,119 wind towers up, you pay the farmer 1237 00:46:55,500 --> 00:46:57,039 rent for use of the land. 1238 00:46:57,340 --> 00:46:58,320 You pay the neighbors 1239 00:46:58,859 --> 00:47:01,655 lower rent for the effect on their environment 1240 00:47:01,715 --> 00:47:03,635 having that power there. You wanna drag a 1241 00:47:03,635 --> 00:47:06,454 power line? Rights of way come with rent. 1242 00:47:06,594 --> 00:47:07,414 So maybe 1243 00:47:07,875 --> 00:47:10,535 data centers have a social capital rent. 1244 00:47:10,994 --> 00:47:12,695 They have to employ locals, 1245 00:47:13,000 --> 00:47:15,500 and they have to build fungible facilities. 1246 00:47:15,880 --> 00:47:18,039 And that power budget that's bought in for 1247 00:47:18,039 --> 00:47:20,780 them has to have a percentage cap component 1248 00:47:21,320 --> 00:47:23,719 that's power for the local schools, power for 1249 00:47:23,719 --> 00:47:26,219 the local churches Yeah. Power for the community. 1250 00:47:26,964 --> 00:47:28,964 Yeah. Because, like, you would never buy a 1251 00:47:28,964 --> 00:47:29,464 house 1252 00:47:29,844 --> 00:47:32,164 that if you repainted the walls, the house 1253 00:47:32,164 --> 00:47:34,105 could could no longer be used. 1254 00:47:34,724 --> 00:47:36,964 Or if you wanted to knock out a, 1255 00:47:36,964 --> 00:47:38,405 you know, if you wanted to knock out 1256 00:47:38,405 --> 00:47:40,484 a wall and make the make the kitchen 1257 00:47:40,484 --> 00:47:42,769 bigger, well, no. Every wall in the house 1258 00:47:42,769 --> 00:47:43,670 is load bearing. 1259 00:47:43,969 --> 00:47:45,989 Nope. You can't touch any of the infrastructure. 1260 00:47:46,610 --> 00:47:49,010 Oh, if you if you if you hang 1261 00:47:49,010 --> 00:47:50,769 up, it's your I live in I live 1262 00:47:50,769 --> 00:47:53,510 in Australia. We build houses out of asbestos 1263 00:47:53,889 --> 00:47:54,949 fibro cement. 1264 00:47:55,489 --> 00:47:57,885 Yeah. But I mean, even like even, like, 1265 00:47:57,945 --> 00:48:00,105 oh, you can't you can't drill it. Like, 1266 00:48:00,105 --> 00:48:01,945 I I once lived in a house where 1267 00:48:01,945 --> 00:48:03,545 I wanted to be able to hang stuff 1268 00:48:03,545 --> 00:48:05,005 from the ceiling of the garage. 1269 00:48:05,305 --> 00:48:07,164 And the builder said, absolutely 1270 00:48:07,625 --> 00:48:10,139 not. If you drill if you drill any 1271 00:48:10,139 --> 00:48:12,000 sort of lag bolt into those, 1272 00:48:12,780 --> 00:48:16,059 joists or whatever, those rafters or whatever, you 1273 00:48:16,059 --> 00:48:18,619 are destroying the structural integrity of the whole 1274 00:48:18,619 --> 00:48:19,119 house. 1275 00:48:19,420 --> 00:48:21,785 And you're like, if that's not cool, I'm 1276 00:48:21,785 --> 00:48:23,785 sorry. No. There's something that's not cool about 1277 00:48:23,785 --> 00:48:24,285 this. 1278 00:48:24,744 --> 00:48:26,425 That's not how it was meant to be, 1279 00:48:26,425 --> 00:48:29,385 was it? Just not like that. Kinda different. 1280 00:48:29,385 --> 00:48:30,525 I wanna democratize 1281 00:48:30,825 --> 00:48:33,065 net. I kind of wanna get to power 1282 00:48:33,065 --> 00:48:35,065 to the people on this. You know? And 1283 00:48:35,065 --> 00:48:35,639 I think 1284 00:48:35,960 --> 00:48:38,920 democratizing the DC would be a really, really 1285 00:48:38,920 --> 00:48:40,139 good first step. 1286 00:48:42,199 --> 00:48:43,980 Yeah. I still know how you do it. 1287 00:48:46,440 --> 00:48:48,619 It's gonna look different for every region. 1288 00:48:49,025 --> 00:48:50,805 Some places are gonna be, 1289 00:48:51,505 --> 00:48:53,765 you know, they're gonna be much more about, 1290 00:48:54,704 --> 00:48:57,765 you know, like freedom or not freedom, but 1291 00:48:58,224 --> 00:49:00,305 tilting towards the business because they want the 1292 00:49:00,305 --> 00:49:00,805 business 1293 00:49:01,289 --> 00:49:03,130 there and other places are gonna be more 1294 00:49:03,130 --> 00:49:05,210 tilting towards the community because they want the 1295 00:49:05,210 --> 00:49:06,110 community there. 1296 00:49:06,410 --> 00:49:07,849 I I don't know how you you can't 1297 00:49:07,849 --> 00:49:09,309 one size fits it all. 1298 00:49:09,610 --> 00:49:11,289 It's all There's gonna be a range of 1299 00:49:11,289 --> 00:49:13,849 solutions. After the microphone goes off, the three 1300 00:49:13,849 --> 00:49:15,449 of us can have a good old sing 1301 00:49:15,449 --> 00:49:16,429 song of Kumbaya 1302 00:49:16,730 --> 00:49:18,724 and fix the problems of the world. 1303 00:49:19,025 --> 00:49:19,525 No. 1304 00:49:21,025 --> 00:49:22,864 The four of us because the plant has 1305 00:49:22,864 --> 00:49:24,644 the drum kit. Yeah. That's right. 1306 00:49:27,824 --> 00:49:30,085 So we've talked about the water cost. 1307 00:49:30,619 --> 00:49:32,140 A lot of people don't think about the 1308 00:49:32,140 --> 00:49:34,079 water cost, the power cost, 1309 00:49:34,539 --> 00:49:36,960 the social cost. Is there anything else? 1310 00:49:38,539 --> 00:49:40,539 Like, we we we've tried to we've tried 1311 00:49:40,539 --> 00:49:42,059 to push some of the good stuff back 1312 00:49:42,059 --> 00:49:44,160 in, but is there anything else 1313 00:49:44,780 --> 00:49:45,280 that 1314 00:49:46,795 --> 00:49:47,855 we can think of? 1315 00:49:48,235 --> 00:49:49,675 And I'm gonna say this. This is kind 1316 00:49:49,675 --> 00:49:51,675 of tangential. A lot of people say, like 1317 00:49:51,675 --> 00:49:54,974 you said earlier, oh, it's leaving behind electricians 1318 00:49:55,114 --> 00:49:56,394 and people who know how to who know 1319 00:49:56,394 --> 00:49:58,235 how to do concrete and plumbing and stuff 1320 00:49:58,235 --> 00:50:00,519 like that. I am utterly convinced 1321 00:50:01,139 --> 00:50:02,119 that the trades 1322 00:50:02,820 --> 00:50:03,559 will eventually 1323 00:50:04,500 --> 00:50:05,880 be the same thing 1324 00:50:06,500 --> 00:50:08,039 as taxi drivers. 1325 00:50:09,460 --> 00:50:12,179 You think they'll skill down? I think they'll 1326 00:50:12,179 --> 00:50:14,594 skill down. I think there will be a 1327 00:50:14,675 --> 00:50:14,994 ski 1328 00:50:16,914 --> 00:50:19,074 a a cream on the top of elite 1329 00:50:19,074 --> 00:50:20,534 drivers who drive, 1330 00:50:21,315 --> 00:50:24,275 not taxis, but they're, like, well trained in 1331 00:50:24,275 --> 00:50:25,094 how to do 1332 00:50:25,554 --> 00:50:26,375 evasion and 1333 00:50:26,675 --> 00:50:29,170 all this other stuff who are driving around 1334 00:50:29,170 --> 00:50:31,489 very important people and doing limos because they 1335 00:50:31,489 --> 00:50:33,410 have the big cars, and then there'll be 1336 00:50:33,410 --> 00:50:34,150 Uber drivers. 1337 00:50:34,530 --> 00:50:36,050 And I think this is the same thing 1338 00:50:36,050 --> 00:50:37,429 is gonna happen to electrical 1339 00:50:37,969 --> 00:50:38,469 and 1340 00:50:38,769 --> 00:50:41,090 plumbing at least. I don't know about concrete 1341 00:50:41,090 --> 00:50:43,030 because concrete's actually to me harder 1342 00:50:43,815 --> 00:50:44,315 and 1343 00:50:44,775 --> 00:50:46,315 and stuff like that. But 1344 00:50:46,855 --> 00:50:48,635 I think that's a thing that's coming. 1345 00:50:49,175 --> 00:50:50,474 So I think resilience 1346 00:50:50,934 --> 00:50:52,795 in a local community comes 1347 00:50:53,655 --> 00:50:55,835 in many shapes and sizes. 1348 00:50:56,809 --> 00:50:57,309 And 1349 00:50:57,690 --> 00:50:59,610 I used to believe very strongly that the 1350 00:50:59,610 --> 00:51:01,849 technology was more good than bad and that 1351 00:51:01,849 --> 00:51:03,070 exposure to technology 1352 00:51:03,449 --> 00:51:05,389 tended to the more good than bad. 1353 00:51:05,849 --> 00:51:06,349 And 1354 00:51:07,050 --> 00:51:09,690 I have become slightly more conservative as I 1355 00:51:09,690 --> 00:51:11,769 got older. I swore I wouldn't do this, 1356 00:51:11,769 --> 00:51:14,105 but I have. And I'm not convinced 1357 00:51:14,804 --> 00:51:17,204 it's a good idea for local communities to 1358 00:51:17,204 --> 00:51:19,704 expose kids to an online life early. 1359 00:51:20,244 --> 00:51:21,224 And I worry 1360 00:51:21,525 --> 00:51:22,505 that the increasing 1361 00:51:23,204 --> 00:51:25,625 investment in technology we're making 1362 00:51:26,119 --> 00:51:28,119 means that people are going to look for 1363 00:51:28,119 --> 00:51:30,539 things that they can invest in to leverage 1364 00:51:30,679 --> 00:51:33,660 that technology investment. And I'm pretty sure education 1365 00:51:34,280 --> 00:51:36,440 and social structures for kids are gonna get 1366 00:51:36,440 --> 00:51:37,660 there. Now we're scared 1367 00:51:38,164 --> 00:51:40,644 of the aspect of kids getting into big 1368 00:51:40,644 --> 00:51:43,284 online environments where we don't have a boundary 1369 00:51:43,284 --> 00:51:45,684 of control. There are social issues there that 1370 00:51:45,684 --> 00:51:47,764 are just horrible. I don't want to even 1371 00:51:47,764 --> 00:51:50,085 think about them. Yeah. But I would actually 1372 00:51:50,085 --> 00:51:51,844 say show. We should do that. I would 1373 00:51:51,844 --> 00:51:52,679 actually say 1374 00:51:53,239 --> 00:51:56,440 ignoring all that, just the mere existence of 1375 00:51:56,440 --> 00:51:58,460 a kid's life bounded online, 1376 00:51:58,839 --> 00:52:01,319 to me, is not a resilient life. It's 1377 00:52:01,319 --> 00:52:03,719 not a good idea, folks. And if we're 1378 00:52:03,719 --> 00:52:06,780 building these DCs and we're building social structure 1379 00:52:07,000 --> 00:52:08,219 into these DCs, 1380 00:52:08,744 --> 00:52:10,505 we have to be very careful to build 1381 00:52:10,505 --> 00:52:12,364 a component of our lived experience, 1382 00:52:13,224 --> 00:52:14,364 not in the DC. 1383 00:52:14,984 --> 00:52:15,884 Yeah. Yeah. 1384 00:52:16,184 --> 00:52:17,944 I know Tom needs to drop, so I 1385 00:52:17,944 --> 00:52:19,784 know he has some other meeting. I don't 1386 00:52:19,784 --> 00:52:21,464 have any other meetings ever, so, you know, 1387 00:52:21,464 --> 00:52:23,510 I could just go on for hours. We 1388 00:52:23,510 --> 00:52:24,889 can talk about your Cherokee. 1389 00:52:26,389 --> 00:52:28,170 So alright. Cool. 1390 00:52:28,550 --> 00:52:31,609 Alright. So, Tom, where can people find you? 1391 00:52:33,589 --> 00:52:34,089 LinkedIn. 1392 00:52:35,190 --> 00:52:35,690 And 1393 00:52:35,989 --> 00:52:38,230 the after the dramatic pause, where can people 1394 00:52:38,230 --> 00:52:39,130 find the plant? 1395 00:52:40,364 --> 00:52:42,765 No. No. Nowhere yet, but I'll say this. 1396 00:52:42,765 --> 00:52:44,465 If, if the people want, 1397 00:52:45,244 --> 00:52:47,244 a social media account for the plant, then 1398 00:52:47,244 --> 00:52:49,164 I who am I to deny the people? 1399 00:52:49,164 --> 00:52:49,664 Right? 1400 00:52:52,780 --> 00:52:55,579 To the people. Alright. George, where can people 1401 00:52:55,579 --> 00:52:56,320 find you? 1402 00:52:56,700 --> 00:52:58,380 You can find me at a p Nick's 1403 00:52:58,380 --> 00:53:00,720 blog, blog.apnick.net. 1404 00:53:01,180 --> 00:53:04,140 LinkedIn, Twitter, anything else? X? I'm sorry. Whatever 1405 00:53:04,140 --> 00:53:06,320 else it is. No. Don't do that stuff. 1406 00:53:06,675 --> 00:53:08,515 Alright. Cool. Alright. I'm Russ White. You can 1407 00:53:08,515 --> 00:53:09,555 always find me here at the hedge. You 1408 00:53:09,555 --> 00:53:10,914 can find me at rule eleven dot tech. 1409 00:53:10,914 --> 00:53:12,434 You can find me on x every now 1410 00:53:12,434 --> 00:53:13,094 and again, 1411 00:53:13,394 --> 00:53:14,135 and LinkedIn. 1412 00:53:14,755 --> 00:53:16,515 We know that we live in an attention 1413 00:53:16,515 --> 00:53:17,335 driven economy 1414 00:53:17,635 --> 00:53:18,934 and that you've paid attention 1415 00:53:19,235 --> 00:53:21,269 all the way to the bitter end of 1416 00:53:21,269 --> 00:53:23,269 this episode of The Hedge. So thank you 1417 00:53:23,269 --> 00:53:25,590 for listening, and we will catch you next 1418 00:53:25,590 --> 00:53:26,090 time.