57:I[6545,["/_next/static/chunks/18edpp3q7kide.js","/_next/static/chunks/2-lg_q3_g_vx-.js","/_next/static/chunks/1suervftjz1w3.js","/_next/static/chunks/1gwwpfbl6ew07.js","/_next/static/chunks/2xi47584r81jm.js","/_next/static/chunks/33zhqrivb54--.js","/_next/static/chunks/3p1i59i0zn5j1.js"],"default"] 53:["$","div","item-7",{"className":"w-[258px] flex flex-col flex-nowrap flex-none gap-y-2.5","children":[["$","div",null,{"className":"aspect-video w-full h-full bg-neutral-500 animate-pulse rounded-md"}],["$","div",null,{"className":"bg-neutral-500 animate-pulse rounded h-3 w-7/8"}],["$","div",null,{"className":"bg-neutral-500 animate-pulse rounded h-3 w-2/3"}]]}] 54:["$","div","item-8",{"className":"w-[258px] flex flex-col flex-nowrap flex-none gap-y-2.5","children":[["$","div",null,{"className":"aspect-video w-full h-full bg-neutral-500 animate-pulse rounded-md"}],["$","div",null,{"className":"bg-neutral-500 animate-pulse rounded h-3 w-7/8"}],["$","div",null,{"className":"bg-neutral-500 animate-pulse rounded h-3 w-2/3"}]]}] 55:["$","div","item-9",{"className":"w-[258px] flex flex-col flex-nowrap flex-none gap-y-2.5","children":[["$","div",null,{"className":"aspect-video w-full h-full bg-neutral-500 animate-pulse rounded-md"}],["$","div",null,{"className":"bg-neutral-500 animate-pulse rounded h-3 w-7/8"}],["$","div",null,{"className":"bg-neutral-500 animate-pulse rounded h-3 w-2/3"}]]}] 58:T1598a,00:00:02 - 00:00:05 The opinions expressed by the host and guests in this show are not 00:00:05 - 00:00:08 necessarily the views of children's health defense. 00:00:34 - 00:00:37 Good morning CHD. We're live here at CHD TV with 00:00:37 - 00:00:41 Financial Rebellion. It's Thursday the 23rd of 00:00:41 - 00:00:44 July 2026. We have an amazing guest for you today. 00:00:44 - 00:00:47 But first we're going to go to the news. So 00:00:47 - 00:00:51 first up is, well, this should disturb everyone. 00:00:51 - 00:00:54 Smart bass net for babies collects 00:00:54 - 00:00:57 and shares sensitive data. So these 00:00:57 - 00:01:01 are little bassinets that I mean I don't know why people would want 00:01:01 - 00:01:04 to put their baby in a bassinet. The baby should be on the mother in my 00:01:04 - 00:01:07 opinion, flesh to flesh type thing. As they say. 00:01:07 - 00:01:11 Skin to skin, I think is what they call it. Um, but everything about that 00:01:11 - 00:01:14 disturbs me greatly because I'm terribly old fashioned. What do you both 00:01:14 - 00:01:17 think? Well, this is this is about an 00:01:17 - 00:01:20 employer, which I think is a state government, um, 00:01:21 - 00:01:23 that's providing this $1,600 00:01:24 - 00:01:26 bassinet free for 00:01:26 - 00:01:30 employees. So I'm assuming the employer 00:01:30 - 00:01:32 has access to all the information that's being 00:01:32 - 00:01:35 collected. Now, apparently there's some sort of 00:01:36 - 00:01:38 shield that is supposed to protect 00:01:38 - 00:01:41 against, uh, the aspects of it that 00:01:42 - 00:01:45 involve EMF. So it cuts EMF 00:01:45 - 00:01:48 poisoning somewhat, but the data it's collecting is 00:01:49 - 00:01:51 date of birth name, you know, all this 00:01:51 - 00:01:54 information. And as always, the question is 00:01:55 - 00:01:58 where on earth is this data going? Who has 00:01:58 - 00:02:01 access to it? Is it going into the control grid? And if you have 00:02:01 - 00:02:05 a newborn baby, that's a great way to start the 00:02:05 - 00:02:08 file. The control file on the baby. 00:02:09 - 00:02:13 Catherine. Don't use it. 00:02:13 - 00:02:16 Just don't use it. I don't care if it's free. 00:02:16 - 00:02:19 It's not free. Nothing's free. Right. 00:02:20 - 00:02:23 So our next article up, so the 00:02:23 - 00:02:26 collaborative privacy advocacy community, Catherine, what 00:02:26 - 00:02:29 is this about? So the brouhaha over 00:02:29 - 00:02:33 license readers flock cameras and, and cameras in your 00:02:33 - 00:02:36 cars and other really invasive 00:02:36 - 00:02:40 methods of destroying privacy has got more and more people 00:02:40 - 00:02:43 worried about privacy. And this is a wonderful website. 00:02:43 - 00:02:46 It's a great resource to how to protect your 00:02:46 - 00:02:49 privacy. So lots of practical advice. And I wanted to make 00:02:49 - 00:02:52 sure you had this because I think it's a very useful tool 00:02:52 - 00:02:56 to help you think about and, and, and take 00:02:56 - 00:02:58 action to protect your privacy. But, and I'm sure Lou will 00:02:58 - 00:03:02 put that link in the show notes. Uh, so next 00:03:02 - 00:03:05 article please, Lou. Okay, so 00:03:05 - 00:03:08 salary report crime that pays is crime that 00:03:08 - 00:03:11 stays. DOJ just wants a, a percentage of the 00:03:11 - 00:03:15 profits. It's a kickback model that facilitates plunder. 00:03:15 - 00:03:18 Kathryn and Carolyn tell us what this is about. So this 00:03:18 - 00:03:22 is more bad news. Um, the Department of Justice has 00:03:22 - 00:03:25 had a long tradition of not prosecuting corporate crime, 00:03:25 - 00:03:28 including banking crime during the financial crisis. And of 00:03:28 - 00:03:31 course, the more individuals working for 00:03:31 - 00:03:34 corporations know they are free to break the law. 00:03:34 - 00:03:38 Um, and, and the only ramification could be a 00:03:38 - 00:03:41 percent of profit shared with the Department of Justice. Then the, 00:03:42 - 00:03:45 you know, sort of the more criminal behavior we've seen steadily for the 00:03:45 - 00:03:48 last two decades. And we get an announcement cover of 00:03:48 - 00:03:51 The Wall Street Journal this week saying the 00:03:51 - 00:03:55 Department of Justice is pulling back. Now, I didn't realize they could pull back 00:03:55 - 00:03:58 because I thought they were pulled back pretty far. So the fact that they're 00:03:58 - 00:04:01 pulling back further to me is very, very bad news. Because what it 00:04:01 - 00:04:04 means is as long as I kick back some of the profits of the 00:04:04 - 00:04:08 Department of Justice, it's a good crime, is a good business model for 00:04:08 - 00:04:11 corporations now. And and that's very, very 00:04:11 - 00:04:14 concerning. A lot of this has to do with 00:04:14 - 00:04:17 with pulling back on fines. Well, if they find the 00:04:17 - 00:04:21 actual costs of all the bad behavior of 00:04:21 - 00:04:24 the even just the banks, uh, they 00:04:24 - 00:04:27 could pay for a lot of the remedies, 00:04:27 - 00:04:31 uh, and force all sorts of things to happen, 00:04:31 - 00:04:34 like privacy in this context. But, you know, the 00:04:34 - 00:04:37 corporate crime is just this, all this does is, is 00:04:37 - 00:04:40 incentivize the companies to figure out a way 00:04:40 - 00:04:44 not to have anybody look at anything. The message to corporate 00:04:44 - 00:04:47 executives are your safe individually to break the 00:04:47 - 00:04:50 law. And then it's really a, um, 00:04:50 - 00:04:54 a, a financial equation as to how much of the profits from 00:04:54 - 00:04:57 the crime versus how much you're going to have to kick back to the Department of 00:04:57 - 00:05:00 Justice. And unfortunately, in many, many cases, 00:05:00 - 00:05:04 crime pays and you're creating yeah, you're 00:05:04 - 00:05:07 creating a business model that supports 00:05:07 - 00:05:11 and encourages crime. And just understand that 00:05:11 - 00:05:14 when in, when people, the regular populace 00:05:14 - 00:05:17 is harmed by something and the government finds a 00:05:17 - 00:05:20 company, they're not taking that fine and 00:05:20 - 00:05:24 paying it to the people who are injured. That's 00:05:24 - 00:05:25 going into the treasury. 00:05:27 - 00:05:30 All right, let's get the next article up. Hegseth 00:05:31 - 00:05:34 estimate Iran war has reached 00:05:34 - 00:05:37 $37.5 billion. So the Secretary 00:05:37 - 00:05:40 of War was testifying on the budget. They are asking 00:05:40 - 00:05:44 for 1.5 trillion, which is a $500 00:05:44 - 00:05:47 billion increase over last year. And and 00:05:47 - 00:05:50 last year had a pretty, pretty big increase over the year before. 00:05:50 - 00:05:53 This is an extraordinary amount of money. And the signal, 00:05:53 - 00:05:57 basically, if you look at the money that's 00:05:57 - 00:06:00 rolling into a lot of the infrastructure for the control 00:06:00 - 00:06:03 grid, it's a tsunami of money coming 00:06:03 - 00:06:07 out of the government in the Chips act 00:06:07 - 00:06:10 and and the big, beautiful bill that's basically creating corporate 00:06:10 - 00:06:13 monopolies. And war is now becoming the biggest 00:06:13 - 00:06:16 business. Now, when Hanks has said $38 00:06:16 - 00:06:20 trillion, uh, and I can't fathom how he 00:06:20 - 00:06:23 came up with that number, because if you look at the destruction of U.S. 00:06:23 - 00:06:26 bases in the Gulf, um, the number is much, 00:06:26 - 00:06:29 much bigger than that. So he has a very constricted 00:06:29 - 00:06:33 view of what the number is. The other reason it's 00:06:33 - 00:06:36 misleading. Um, we just on money and markets 00:06:36 - 00:06:39 this week, we, we showed a clip of a retired 00:06:39 - 00:06:43 admiral saying the Iraq war took 20 years. This will take 00:06:43 - 00:06:46 at least that. And instead of saying this war is going to last 4 to 5 00:06:46 - 00:06:49 weeks, which was what they said in February. Now we're 00:06:49 - 00:06:53 talking about a 20 year war. And if that's the 00:06:53 - 00:06:56 case, it's important to understand that a five week cash 00:06:56 - 00:06:59 flow produces a lot less value on a stock than 00:06:59 - 00:07:03 a 20 year cash flow. And, um, and so 00:07:03 - 00:07:06 between the appropriations going up 50%, 00:07:06 - 00:07:09 um, and, uh, you know, sort of what's 00:07:09 - 00:07:13 been happening yesterday, the House voted to pass 00:07:13 - 00:07:15 section 219 and the NDAA 00:07:16 - 00:07:19 integrating the Israeli military with the US government. So if you 00:07:19 - 00:07:23 if you look at the combination of things going and what the expense 00:07:23 - 00:07:26 is, we're talking about something, uh, that's going to 00:07:26 - 00:07:29 cost, you know, if Iraq costs 6 or $7 trillion, 00:07:29 - 00:07:33 this is, you know, if what if if what is being said 00:07:33 - 00:07:36 is true, this is going to cost, you know, double that 00:07:37 - 00:07:40 and, and it's going to swing an enormous sector of the economy 00:07:40 - 00:07:43 into defense out of, you know, much more 00:07:43 - 00:07:46 important, I would say technological innovation and, and 00:07:46 - 00:07:50 various innovations to make society better. So, 00:07:50 - 00:07:53 um, so I would say the Secretary of war dramatically 00:07:53 - 00:07:55 understated the cost of this war. 00:07:56 - 00:07:58 You can call it defense, but I call it war 00:08:00 - 00:08:03 and offense. Uh, just on another 00:08:03 - 00:08:06 note, I'm talking about war. I was talking to some, 00:08:07 - 00:08:10 let's say, military people the other day who were ordered to deploy to the 00:08:10 - 00:08:13 Middle East and are standing firm and refusing to 00:08:13 - 00:08:16 because they're making them take the anthrax, 00:08:17 - 00:08:20 which I don't know if you know about the anthrax, but it's it's the most 00:08:20 - 00:08:23 disabling. It's the yeah, the most 00:08:23 - 00:08:24 dangerous. 00:08:25 - 00:08:29 It's a death certificate with all kinds of problems going on in the military. 00:08:29 - 00:08:32 But I'm glad they're standing firm. I'm glad they're not just shooting themselves up with 00:08:32 - 00:08:35 anthrax. And, I mean, you couldn't get anything worse to put into someone. 00:08:35 - 00:08:38 But anyway, that's another subject. But just FYI, 00:08:39 - 00:08:42 um, I'm really proud of the military. Could the defender, could 00:08:42 - 00:08:45 the defender do an article on that? I'll speak to 00:08:46 - 00:08:49 Catherine. Okay. Yes. I mean, we we 00:08:49 - 00:08:52 need to but we need to be fighting for them. They should not be you know. 00:08:52 - 00:08:55 Absolutely. They have anthrax to fight for us. 00:08:56 - 00:09:00 Well here's the thing, you know, the what 00:09:00 - 00:09:03 you're you're, you know, if you're going to go to war, you don't want to lose a war. 00:09:03 - 00:09:07 And if you're going to poison your soldiers before they deploy, you've got a big 00:09:07 - 00:09:09 problem. That's right. That's right. 00:09:10 - 00:09:13 Um, so yeah, that's another one. Maybe I'll get Meryl Nass on at some point to talk 00:09:13 - 00:09:17 about the dangers of anthrax, because, um, I don't know, there's a, there's a film I can't 00:09:17 - 00:09:20 remember, it's called Standing Order or something like that. And it shows how 00:09:20 - 00:09:24 the, the people who have, have children after the anthrax, they've come out with 00:09:24 - 00:09:27 no limbs. And like, it's, it's really toxic, bad 00:09:27 - 00:09:30 stuff. It should not be. So, uh, so Ty and 00:09:30 - 00:09:34 Charlene Bollinger made, uh, did a series on 00:09:34 - 00:09:37 vaccines and they had one piece of it on, 00:09:37 - 00:09:40 uh, on the anthrax, which is very, very good. 00:09:41 - 00:09:44 Yeah. So, uh, anyway, we need to stand by the military on 00:09:44 - 00:09:47 that. All right. Let's get the data center article up. It's our 00:09:47 - 00:09:50 last article before we bring our amazing guest on cheating 00:09:50 - 00:09:54 the system. How AI data centers bypass water 00:09:54 - 00:09:57 and air pollution laws and permits. Well, we've been talking 00:09:57 - 00:10:01 about data centers on Financial Rebellion for a long time 00:10:01 - 00:10:04 now, right? But the, the, you know, the 00:10:04 - 00:10:07 local debate is significant. I told 00:10:07 - 00:10:11 you the last time we discussed this, Tennessee passed a law saying 00:10:11 - 00:10:14 you can't do a data center if it increases water 00:10:14 - 00:10:18 or electricity charges or supply. Um, 00:10:18 - 00:10:21 and that's one of the big debates at the local area. And 00:10:21 - 00:10:24 of course, the question for many areas is, how do 00:10:24 - 00:10:28 you know for sure what, what the real 00:10:28 - 00:10:31 usage is going to be? Because there have been a lot of surprises 00:10:31 - 00:10:35 in, in different areas where they thought it was one thing and it's turned out to be 00:10:35 - 00:10:38 another. Yeah. Well, let's bring our 00:10:38 - 00:10:41 guests on. Catherine, can you? Okay, well, this is 00:10:41 - 00:10:45 a real treat. Travis Oliphant is a 00:10:45 - 00:10:47 CEO, head of open teams. He has a long 00:10:47 - 00:10:50 history as a software developer, a data 00:10:50 - 00:10:54 scientist, and an entrepreneur in the tech area. He 00:10:54 - 00:10:57 was part of the original Python ecosystem. He 00:10:57 - 00:11:00 helped create a great documentary that we had his movie of the week 00:11:00 - 00:11:03 about it. Um, but he created numpy package, 00:11:03 - 00:11:06 which was, um, helped with the 00:11:06 - 00:11:10 numerical, uh, computation in Python 00:11:10 - 00:11:13 and many other accomplishments. And, um, 00:11:13 - 00:11:16 co-founder or founder of several tech companies, but 00:11:16 - 00:11:19 now running open teams and is a great, 00:11:19 - 00:11:23 uh, I know of no one better to explain open 00:11:23 - 00:11:26 AI. Um, and, and what's happening in the debate 00:11:26 - 00:11:29 around artificial intelligence. He, as, uh, 00:11:29 - 00:11:33 Travis was originally, uh, introduced to me, he 00:11:33 - 00:11:36 was an OG as in, you know, original 00:11:36 - 00:11:39 gangster of, of the people who created AI. So we're talking 00:11:39 - 00:11:43 to one of the real founders of the AI, AI 00:11:43 - 00:11:46 phenomenon. Travis, thank you so much for getting up early. I know you're out 00:11:46 - 00:11:49 traveling on the West Coast, but it's a real treat. 00:11:49 - 00:11:53 And, uh, you join me not that long ago on the Solari report. It 00:11:53 - 00:11:56 was a very popular and wonderful conversation. So I 00:11:56 - 00:12:00 appreciate your taking the time out of your busy schedule to join us. 00:12:00 - 00:12:04 Oh, Catherine, thank you so much. It is a pleasure and honor to be here. 00:12:04 - 00:12:07 I really appreciate getting to know the both of the rest of the hosts here. 00:12:07 - 00:12:10 Uh, Carolyn and Polly Tommey. I'm thrilled to have the discussion. It's such an 00:12:10 - 00:12:14 important conversation. So we're going to start off with Carolyn. 00:12:14 - 00:12:17 We have a term for the day and it's open AI. So Carolyn, 00:12:17 - 00:12:21 take us away. But Travis, we're going to quickly turn it over to 00:12:21 - 00:12:22 you, open source AI. 00:12:26 - 00:12:27 Carolyn you're muted, I think. 00:12:30 - 00:12:33 Well, I, I use it without the microphone. Can you hear me 00:12:33 - 00:12:36 now? Yes we can. Okay. Said it was on Bumblebee, but 00:12:36 - 00:12:40 apparently it's not. Um, I originally came up 00:12:40 - 00:12:43 with open AI and then I realized, oh my 00:12:43 - 00:12:46 goodness, that is a that's the name 00:12:46 - 00:12:49 of a company. That's not actually the term. 00:12:49 - 00:12:52 The term is open source AI. So I did a little bit of 00:12:52 - 00:12:56 research. Oh, what's the difference? And I got a little bit of the 00:12:56 - 00:12:58 history, uh, of open AI, 00:12:59 - 00:13:02 which originally was a charitable 00:13:02 - 00:13:06 organization or at least, uh, had 00:13:06 - 00:13:09 charitable purposes under one of the corporate statutes. 00:13:09 - 00:13:12 And Elon Musk put a whole lot of money into it, 00:13:13 - 00:13:16 he says, on the assumption that this was going to be an open 00:13:16 - 00:13:19 source and that they published it the 00:13:19 - 00:13:22 way we would get, um, uh, Libre 00:13:22 - 00:13:25 office and, uh, ubuntu that we have 00:13:25 - 00:13:29 talked about. You can download the software, it's 00:13:29 - 00:13:32 free and you can do with it what you want. And 00:13:32 - 00:13:35 originally that was apparently the idea for 00:13:35 - 00:13:39 open AI. And then the board, I guess 00:13:39 - 00:13:41 against, uh, Elon Musk's, uh, 00:13:42 - 00:13:45 wishes went, went for profit. 00:13:45 - 00:13:48 And now it's private and they're not 00:13:48 - 00:13:52 as open, but that's, that's the background. And there actually, 00:13:52 - 00:13:55 there was a case, uh, Elon Musk sued 00:13:56 - 00:13:58 for if his money back or to force it to 00:13:58 - 00:14:02 go back to a nonprofit, uh, model and 00:14:02 - 00:14:05 the courts determined the case according to 00:14:05 - 00:14:09 procedural matters. So never ruled on that issue 00:14:09 - 00:14:11 because they said he missed the statute of limitations. 00:14:12 - 00:14:15 So that's the background on open AI. Open 00:14:15 - 00:14:18 source AI is somewhat like, 00:14:19 - 00:14:22 uh, the, uh, ubuntu, for example, 00:14:22 - 00:14:25 which is, uh, an operating system that you can 00:14:25 - 00:14:28 download as an alternative to 00:14:28 - 00:14:31 Microsoft. And that's something that I use on one of my 00:14:31 - 00:14:34 computers. Um, but it's not 00:14:34 - 00:14:37 completely open those. What we're used to 00:14:37 - 00:14:41 hearing about open source is the, all the code 00:14:41 - 00:14:44 is open as, and as I understand it, it's only 00:14:44 - 00:14:48 the weightings that are open, but 00:14:48 - 00:14:51 there are some open source AI 00:14:51 - 00:14:54 companies that will either license or 00:14:54 - 00:14:58 give you free, uh, a, uh, 00:14:58 - 00:15:02 software that you can take and run with. And 00:15:02 - 00:15:05 I'm interested to find out whether Travis started with one of 00:15:05 - 00:15:08 those and then embellished it and did what he wanted with it. 00:15:08 - 00:15:11 Because starting from scratch is like putting the millions of 00:15:11 - 00:15:14 dollars, uh, or billions of dollars that, that Elon 00:15:14 - 00:15:17 Musk put into open AI. I mean, starting from scratch is 00:15:17 - 00:15:21 really, really very, um, very expensive. But 00:15:21 - 00:15:24 can you tell us more about this? What's what's 00:15:24 - 00:15:27 open, like, what is this weighting and what 00:15:27 - 00:15:31 is involved in training the AI, 00:15:31 - 00:15:34 which is the source of the 00:15:34 - 00:15:37 information that the AI chatbot 00:15:37 - 00:15:40 is giving you? You've introduced a lot of 00:15:40 - 00:15:44 great concepts. Maybe we can unpack a little bit. Um, that 00:15:44 - 00:15:47 open source has a very long history. And you talked about 00:15:47 - 00:15:50 ubuntu. You talked about Linux. A lot of people, myself included, came 00:15:50 - 00:15:53 to open source via Linux, actually back in the 90s. 00:15:54 - 00:15:57 Uh, it started by a young man in Finland who just 00:15:57 - 00:16:00 basically saw the closed operating systems around. There was a 00:16:00 - 00:16:04 big movement, free software movement, open source movement emerging at the time 00:16:04 - 00:16:07 where people recognized, oh, once we have the ability to connect with each other, 00:16:07 - 00:16:10 humans are remarkably interested in, in cooperating 00:16:10 - 00:16:14 with each other. That's the one big thing learning that life has taught 00:16:14 - 00:16:17 me is humans cooperate all over the globe. Either 00:16:17 - 00:16:21 we want to cooperate, we're we're wired to cooperate. And the 00:16:21 - 00:16:24 challenge is how do we ensure that cooperation leads to 00:16:24 - 00:16:27 further cooperation, not to subjugation? It's very, uh, 00:16:27 - 00:16:30 unfortunately easy for, uh, kind of the, the biggest 00:16:30 - 00:16:34 bullies in the room to essentially use that cooperation against us 00:16:34 - 00:16:37 too quickly. It happens all throughout history and, 00:16:37 - 00:16:40 uh, open and open source. It becomes a method 00:16:40 - 00:16:44 for our cooperation to actually push back against that. So 00:16:44 - 00:16:47 it's been around for a while. I showed up in the AI 00:16:47 - 00:16:50 world 30 years ago, but essentially not, not because 00:16:50 - 00:16:54 AI has been around for a long time, but the core of AI is a 00:16:54 - 00:16:57 lot of math. You have to basically multiply, uh, 00:16:57 - 00:17:00 matrices together, multiply huge volumes of data together at 00:17:00 - 00:17:04 scale, in parallel. It takes and it's a lot of math and it's a lot of 00:17:04 - 00:17:07 computation. And the foundations of that computation have 00:17:07 - 00:17:11 existed in open source for at least three decades. So 00:17:11 - 00:17:14 numpy, the library I'm, I'm famous for, so to speak, for at 00:17:14 - 00:17:17 least the nerd world or the people that kind of know about that 00:17:17 - 00:17:20 space. Uh, this is something I, I, I wrote gave 00:17:20 - 00:17:24 away because it was really important to me that the 00:17:24 - 00:17:27 foundations of technology that could give rise to 00:17:27 - 00:17:30 millions of innovations be open to everybody. So to amplify 00:17:30 - 00:17:34 cooperation at the same time, how do you, how do you 00:17:34 - 00:17:37 create businesses that allow that cooperation to scale in, 00:17:37 - 00:17:40 in a way that goes beyond the tribal cooperation? This 00:17:40 - 00:17:44 is essentially the, the, the, the conflict or the 00:17:44 - 00:17:47 tension of society is how do you enable 00:17:47 - 00:17:50 cooperation at scale while continuing to to happen it 00:17:50 - 00:17:54 to occur cooperation, open AI. Just 00:17:54 - 00:17:57 another example of this tension between, uh, Elon who 00:17:57 - 00:18:00 wanted to, uh, was concerned that AI was 00:18:00 - 00:18:03 focusing under one company, Google in particular. 00:18:03 - 00:18:06 Google has been accumulating everybody's data. 00:18:07 - 00:18:10 Uh, you know, you have this free service and every time you have a free service, you're 00:18:10 - 00:18:13 really giving something away. And it's typically something pretty 00:18:13 - 00:18:17 important if you don't, if you're not buying something, you are the something. 00:18:17 - 00:18:20 If you're not buying a product, you are the product, as people say. And so it's 00:18:20 - 00:18:23 important to keep that in mind, understand that that's what was happening. Elon 00:18:23 - 00:18:27 said. He's like, wait, this isn't good. It's not good to have all the world's 00:18:27 - 00:18:30 data consolidated into a single organization, uh, because that 00:18:30 - 00:18:33 will reduce cooperation necessarily because there's 00:18:33 - 00:18:36 no organization that's, uh, that can 00:18:36 - 00:18:39 truly, uh, scale to the degree that humans can 00:18:39 - 00:18:42 scale. So OpenAI was constructed to keep AI 00:18:42 - 00:18:46 open. Well, it didn't do that. They saw the potential. 00:18:46 - 00:18:49 And, uh, once they released a version of AI that many people had, Google 00:18:49 - 00:18:52 had it. OpenAI did the thing of releasing it. These, this, 00:18:53 - 00:18:56 this chat bot that allowed you to have a conversation with a machine that 00:18:56 - 00:18:59 seemed very human. Uh, why? Because it's really an 00:18:59 - 00:19:03 interpolator. It orients your prompt, 00:19:03 - 00:19:06 it orients what you've said, goes into this machine, millions 00:19:06 - 00:19:09 of calculations, billions of calculations, results in a 00:19:09 - 00:19:13 continuation of that prompt. It's an interpolator and you're 00:19:13 - 00:19:16 orienting. The way I see it is you have a 00:19:16 - 00:19:19 massive machine that's taking all the internet's 00:19:19 - 00:19:22 data, and then when you give it a prompt, it's orienting it, 00:19:22 - 00:19:25 orienting those data to respond to you. And of course, 00:19:25 - 00:19:29 because it's trained on billions and trillions of data 00:19:29 - 00:19:32 points from humans, it, it reflects humans. It seems 00:19:32 - 00:19:36 like it's human. Uh, but it's not. It's a machine. Uh, I 00:19:36 - 00:19:39 like to call them cogs. Uh, and it's really, 00:19:39 - 00:19:42 really important that, that, that, that people can reproduce 00:19:42 - 00:19:46 that, that it's not just in the hands of a few folks. So this openness has 00:19:46 - 00:19:49 been around for a while. It gave rise to the technology 00:19:49 - 00:19:52 that led to AI. So Google, um, open 00:19:52 - 00:19:55 AI, everybody's using libraries that 00:19:55 - 00:19:58 themselves are open source. That's one thing I think it's important for 00:19:58 - 00:20:02 people to understand is, you know, we talk about the money Elon invested. 00:20:02 - 00:20:06 Actually, that money didn't go to build the infrastructure that was already 00:20:06 - 00:20:09 created by millions of people, essentially cooperating 00:20:09 - 00:20:13 over the past three decades to produce the foundations of technology 00:20:13 - 00:20:16 that would enable this AI revolution, this 00:20:16 - 00:20:19 AI ecosystem, to evolve. The money went to servers. 00:20:19 - 00:20:22 Primarily. The money went to, um, you know, large 00:20:22 - 00:20:26 amounts of compute that was necessary to train the early 00:20:26 - 00:20:29 models. Now, the cost of that training has been coming 00:20:29 - 00:20:32 down and the ability to then deploy it has also 00:20:32 - 00:20:35 coming down. And so I think it's definitely easy to get to, 00:20:35 - 00:20:39 to believe that somehow because something took a lot of money, it must necessarily 00:20:39 - 00:20:42 take a lot of money in the future. And that's just not true. Uh, 00:20:42 - 00:20:45 so because millions of people have 00:20:45 - 00:20:48 constructed a fabric of technology that is open, 00:20:49 - 00:20:52 so open AI used open source in order to produce their 00:20:52 - 00:20:55 solutions. And now going forward, it's important to 00:20:55 - 00:20:58 understand there are a lot of people out there promoting open weights 00:20:58 - 00:21:02 versus open source. The difference is, is, 00:21:03 - 00:21:06 is real. I mean, mostly open weights lets you use 00:21:06 - 00:21:09 AI for your purposes. You can take the weights and those weights 00:21:09 - 00:21:13 are numbers in a model. They're like parameters of a model that 00:21:13 - 00:21:16 you then deploy and do whatever you want with it on your hardware, on 00:21:16 - 00:21:20 your systems, and you can then use it for your purposes. That's open 00:21:20 - 00:21:23 weight. Open source gives you more of the 00:21:23 - 00:21:26 technology around the weights to adjust it to, uh, 00:21:26 - 00:21:29 change the algorithm to, uh, well, I see this 00:21:29 - 00:21:33 model use this particular detail of math or this particular concept, and 00:21:33 - 00:21:36 I want to adjust that. Well, if I have the source that that 00:21:36 - 00:21:40 of the training code, for example, or the source of the code that's used 00:21:40 - 00:21:43 to move around, they call it the harness around the 00:21:43 - 00:21:46 the model, but they're both useful. Both open 00:21:46 - 00:21:50 weights and open source are both useful. They have different implications in what they 00:21:50 - 00:21:53 enable. And so almost in order, you know, open source 00:21:53 - 00:21:56 is ideal. Open weight is is better and 00:21:56 - 00:21:59 then closed. Everything is, um, 00:22:00 - 00:22:03 very specific. I mean, it's not not saying everything has to be open. If 00:22:03 - 00:22:06 somebody does have, you know, if they take open and create closed, 00:22:07 - 00:22:10 at the very least they should be doing in a limited way and not in a way that's 00:22:10 - 00:22:13 trying to undermine the global cooperation that led to the 00:22:13 - 00:22:16 foundation of what they're of what they have. So, 00:22:16 - 00:22:19 Travis, we, we, we have 00:22:19 - 00:22:22 people in the tech industry who are 00:22:22 - 00:22:25 arguing vociferously for, uh, 00:22:25 - 00:22:28 closed models and for government support to keep models 00:22:28 - 00:22:32 closed. Yeah. Can you explain what, 00:22:32 - 00:22:34 what their argument is other than, 00:22:35 - 00:22:38 uh, you know, obviously that closed 00:22:38 - 00:22:41 model with them in control gives them enormous power. 00:22:41 - 00:22:45 Yeah, I'll, I'll try to steal their argument. I believe their 00:22:45 - 00:22:48 argument is, uh, that this is a technology that 00:22:48 - 00:22:51 enables, um, enables 00:22:51 - 00:22:55 information to flow and ideas to flow very quickly so that 00:22:55 - 00:22:58 you, that if in the hands of the wrong people, you know, bad people can do 00:22:58 - 00:23:01 bad things quickly, more quickly if they have access to these 00:23:01 - 00:23:05 models. And so they, the this is a common 00:23:05 - 00:23:08 refrain to use, you know, fear of, of fear 00:23:08 - 00:23:12 of harm, to reduce liberty, to reduce people's 00:23:12 - 00:23:15 access to things. Uh, I, I highly 00:23:15 - 00:23:18 suspicious of such arguments. Um, they're very, 00:23:18 - 00:23:22 uh, slippery slopes, but basically it's this, this is a critical 00:23:22 - 00:23:25 technology. I think one of the things they lean on is that it's, it's, 00:23:25 - 00:23:28 it's not trivial to take, uh, 3 00:23:28 - 00:23:32 trillion parameters, 3 trillion weights and then distribute 00:23:32 - 00:23:35 them and serve them. It actually does take some significant hardware, 00:23:35 - 00:23:39 not not as much as a huge data center, actually, but let's say it takes, 00:23:39 - 00:23:42 you know, not just one computer, but ten computers, right? Usually 00:23:42 - 00:23:46 serve a large model would take about ten high powered computers to, uh, 00:23:46 - 00:23:49 somewhere. So actually accessible to a lot of people. 00:23:49 - 00:23:52 But they take that and say, oh, this, you know, since this is expensive to 00:23:52 - 00:23:56 host, only a few people should be allowed to have it. And then, 00:23:56 - 00:23:59 you know, somebody needs to be in control of whether these models are, are 00:23:59 - 00:24:03 safe or not safe. The challenge with that is, what does that even 00:24:03 - 00:24:06 mean? What does it mean for a model to be safe or not safe? 00:24:06 - 00:24:10 I think that's a very, um, it's easy to say. And then to 00:24:10 - 00:24:13 point to a couple of instances, this is why the, the big model producers are 00:24:13 - 00:24:16 in the habit of essentially alarmism. They want 00:24:16 - 00:24:20 you to see bad things. They want you to see the harm AI can do. So they're 00:24:20 - 00:24:23 very it's a, it's a trade off because they want you to be excited about it. 00:24:23 - 00:24:27 The other time. They want you to be aware that, oh, this has got to be constrained in the hands 00:24:27 - 00:24:30 of. Only us can manage it. So. So 00:24:30 - 00:24:34 that's a tension happening right now. Let let me tell you what I hear 00:24:34 - 00:24:37 as a small business owner. What I hear 00:24:37 - 00:24:40 is I hear the Chinese marching around the world 00:24:40 - 00:24:44 saying, we're going to give open model AI. We're going to 00:24:44 - 00:24:47 spread it widely and use it on a very diversified basis so 00:24:47 - 00:24:51 that everybody can have it. And I hear 00:24:51 - 00:24:54 the Americans saying, we have to have a closed model. 00:24:54 - 00:24:57 And what that what I hear is the Chinese are saying, we're going to give 00:24:57 - 00:25:00 everybody a machine guns. And the Americans are saying, you, the 00:25:00 - 00:25:03 small business owner, have to keep using bow and arrows. 00:25:04 - 00:25:07 Yeah, that's exactly right. That's that's a that's a very good metaphor for 00:25:07 - 00:25:10 what they're saying. Of course, they wouldn't admit that's what they're saying because the 00:25:10 - 00:25:13 argument is ridiculous. Right. And if you put it like that, they're 00:25:13 - 00:25:17 like, oh yeah, I'm on the wrong side of this argument, which they 00:25:17 - 00:25:20 are. Unfortunately, however, there's a lot of money that's gone 00:25:20 - 00:25:23 into to prop up this model and a 00:25:23 - 00:25:27 lot of investors, a lot of people who want to get their money back. 00:25:27 - 00:25:31 Okay, there's so much money that's gone into this that of course, there's going to be a 00:25:31 - 00:25:34 lot of support for the idea that these companies need to thrive. Right. But 00:25:34 - 00:25:37 a surprisingly large amount of that money, if you peel back the 00:25:37 - 00:25:41 layers, is taxpayers money. Yes. 00:25:41 - 00:25:44 Sure. Right. Yeah, sure. Yeah. So, 00:25:44 - 00:25:45 um, 00:25:47 - 00:25:50 but that argument, regardless of the amount of money that's been put in already, 00:25:50 - 00:25:53 it dwarfs the amount of resources, 00:25:53 - 00:25:56 money, cooperation that's going to be limited if 00:25:56 - 00:25:59 we don't flip the script, if we don't, 00:26:00 - 00:26:03 uh, uh, go back to our roots in America 00:26:03 - 00:26:06 of independence, distributed prosperity, distributed 00:26:06 - 00:26:09 autonomy. I mean, this is the year we celebrate the independence 00:26:09 - 00:26:12 of the United States. Independence of America, 200, 00:26:12 - 00:26:16 250 years of independence. We have we have it is so 00:26:16 - 00:26:18 critical that we declare AI independence at the same time 00:26:19 - 00:26:22 that. So let me. Important moment, let me. Humanity. 00:26:23 - 00:26:26 Let me just play devil's advocate. I 00:26:26 - 00:26:29 don't think they can control it. They 00:26:29 - 00:26:33 can't. You're 100% right. That's why. And 00:26:33 - 00:26:36 in fact, I'm very passionate about people taking 00:26:36 - 00:26:39 control of their future by taking control of their data 00:26:40 - 00:26:43 and then control of their intelligence. But it's going to 00:26:43 - 00:26:46 happen. The seeds are already planted. The only question is 00:26:46 - 00:26:49 when, how much? How much suffering, how much? Uh, 00:26:49 - 00:26:53 poor decision making will be made in the United States before the 00:26:53 - 00:26:56 eventuality of open source AI becomes clear. 00:26:57 - 00:27:00 Uh, right. That because it's already 00:27:00 - 00:27:03 out, it's already out. Now, fortunately, many people in 00:27:03 - 00:27:07 the United States are waking up to this. I've actually been very excited over just the past 00:27:07 - 00:27:10 three months. I've been telling this story for about, you know, basically 00:27:10 - 00:27:13 30 years, but particularly with AI for the past three years, two and a half years, 00:27:14 - 00:27:17 um, the past two and a half months, I've been thrilled. 00:27:17 - 00:27:21 Satya Nadella is telling the story, uh, the all in podcast, telling 00:27:21 - 00:27:24 the story, uh, thinking machines, labs, people. There's, there's 00:27:24 - 00:27:27 investors in Silicon Valley who are putting money into companies that are, that are 00:27:27 - 00:27:31 driving open source. Nvidia is telling the story. Qualcomm is telling the 00:27:31 - 00:27:34 story. Apple is telling the story. So it's 00:27:34 - 00:27:37 actually fantastic news that it's just a 00:27:37 - 00:27:40 matter of which, you know, there unfortunately, there are still 00:27:40 - 00:27:44 a lot of big companies and and some large organizations, the federal 00:27:44 - 00:27:47 government that are still, uh, pushing. I mean, even Palantir, 00:27:47 - 00:27:50 Alex Karp got on the news and talked about the importance of openness, the importance of open 00:27:50 - 00:27:54 people controlling their own, their own models. So we've got 00:27:54 - 00:27:57 so many people are are explaining this, 00:27:57 - 00:28:01 but it's not over. And so we're, you know, that's one 00:28:01 - 00:28:04 message. We're going to have a retaliation for that retaliation. We have a we have 00:28:04 - 00:28:08 barbs thrown from the other side as well. I think that's what you're hearing. You see, 00:28:08 - 00:28:11 you know, you'll see news releases of, oh, we had a AI 00:28:11 - 00:28:14 escape and harm something. So see how bad it 00:28:14 - 00:28:18 can be. We better control it. Right. And well, but the 00:28:18 - 00:28:21 guys who are supposed to control are the ones who let it escape, I mean. Correct. 00:28:22 - 00:28:25 Exactly. It's like, I mean, this is 00:28:25 - 00:28:28 a an impressive technology. It's going to 00:28:28 - 00:28:32 require millions of people, billions of 00:28:32 - 00:28:35 people, preferably, but it's going to require millions of people putting on 00:28:35 - 00:28:38 their their accountability hats, their relationships that they 00:28:38 - 00:28:42 already have. The trust institutions that we've spent, uh, 00:28:42 - 00:28:45 centuries produce constructing and then hopefully 00:28:45 - 00:28:49 kind of deconstructing some of the, uh, antitrust 00:28:49 - 00:28:52 institutions that have already existed as well. And we all need 00:28:52 - 00:28:55 this to be supporting ourselves. So supporting the distributed 00:28:55 - 00:28:58 economy, right? So, so I'll say one thing, but 00:28:58 - 00:29:02 then I want to turn to Polly and, and Carolyn and get their questions. 00:29:02 - 00:29:06 Um, I was telling you, the marvelous young woman the other day, I heard 00:29:06 - 00:29:09 say somebody, somebody had said, uh, 00:29:09 - 00:29:13 you know, the young people hate AI. And she turned to them and said, we don't 00:29:13 - 00:29:15 hate the AI. We hate tech bros. 00:29:16 - 00:29:19 So, so part of it is we need a vision of, 00:29:20 - 00:29:24 you know, wise governance related to 00:29:24 - 00:29:27 all technology. And what we've seen in America 00:29:27 - 00:29:31 is not wise people in governance. 00:29:31 - 00:29:34 And, and I think there's a real hunger for wise people in 00:29:34 - 00:29:37 governance. And that's why I think open 00:29:37 - 00:29:40 source, you get millions of people involved. And from that can 00:29:40 - 00:29:44 emerge the wise governance we need. Mhm. I 00:29:44 - 00:29:47 agree, I agree, and that's where we can turn. I think that's part of my 00:29:47 - 00:29:50 message is we can turn to these millions of people 00:29:50 - 00:29:53 who have been at the heart of these communities. Uh, 00:29:53 - 00:29:56 yes, I started the NumPy project. NumPy was successful, 00:29:56 - 00:30:00 however, because of thousands of people, right? Just 00:30:00 - 00:30:03 like anything else. It wasn't because I did it alone. I sort of started a 00:30:03 - 00:30:07 movement and I constructed and I was very eager to get 00:30:07 - 00:30:10 collaborators. And then there's a lot of, you know, give and take and 00:30:10 - 00:30:14 compromise in order to maintain those collaborations. So 00:30:14 - 00:30:17 I would really encourage everyone to, to watch the Python 00:30:17 - 00:30:20 documentary that you had made because you meet those 00:30:20 - 00:30:24 people, you meet people who could govern a system like 00:30:24 - 00:30:27 this. Anyway, Polly and Carolyn, I've been, I've been hogging 00:30:27 - 00:30:31 well, where can we where can we watch the Python documentary? Is it, 00:30:31 - 00:30:34 uh, Lou, pull it up. I, I put, uh, we had it as movie of the 00:30:34 - 00:30:38 week, uh, on the 00:30:38 - 00:30:41 Solari report. Is there a link we can put in the. Yeah. Or or just 00:30:41 - 00:30:44 go, just go to YouTube and, and put in Python the 00:30:44 - 00:30:48 documentary. You'll pull it right up. It's it's on 00:30:48 - 00:30:51 YouTube. It's a freely available the producers, a 00:30:51 - 00:30:54 company called cult repo, uh, they've made some 00:30:54 - 00:30:57 impressive technology documentaries. Just came out with a C++ documentary. 00:30:58 - 00:31:01 Oh, really illustrates. Yes. And they have a 00:31:01 - 00:31:04 JavaScript documentary. I think this is one of the things 00:31:04 - 00:31:07 that many people don't understand broadly, 00:31:07 - 00:31:10 is just how incredible the global 00:31:10 - 00:31:14 cooperation of technologists has been. And 00:31:14 - 00:31:17 it's important in people, especially people's minds, as they think about 00:31:17 - 00:31:20 the technology future. They recognize it isn't 00:31:20 - 00:31:24 all coming out of one part of the world. You know, the tech 00:31:24 - 00:31:27 bros, so-called, are essentially an 00:31:27 - 00:31:30 anomaly. They're basically a group of 00:31:30 - 00:31:33 consolidators and, you know, and trying to, you know, provide 00:31:33 - 00:31:36 return to investors. I'm not they're not, you know, all bad. It's just 00:31:36 - 00:31:39 simply the incentive structures are limited. 00:31:39 - 00:31:42 Meanwhile, there's been millions of people cooperating, you know, 00:31:42 - 00:31:45 starting from the roots of the scientific endeavor and 00:31:45 - 00:31:48 constructing amazing collaborative activities, 00:31:49 - 00:31:52 uh, despite being poorly funded, despite 00:31:52 - 00:31:56 having that's the amazing thing is just with a little bit of funding, a little bit 00:31:56 - 00:31:58 of, of support, it's on the 00:31:58 - 00:32:02 unlocking of cooperation among millions 00:32:02 - 00:32:05 has done amazing things. And what enabled that 00:32:05 - 00:32:08 communication, just the internet enabling people to communicate with each 00:32:08 - 00:32:12 other that once you, you know, in humans, that's the way we are. Once we 00:32:12 - 00:32:15 enable people to communicate and kind of get out of their way, 00:32:15 - 00:32:18 provide some basic, um, support for that 00:32:18 - 00:32:21 communication. It's, it's incredible the 00:32:21 - 00:32:24 prosperity and the support we generate together. So 00:32:24 - 00:32:27 Travis, one of the reasons I think this is so important is if 00:32:27 - 00:32:31 you look at the, the economy of 00:32:31 - 00:32:34 any place, so you're probably headed to 00:32:34 - 00:32:37 Utah. So you live in Utah, but you know, so we'll just take the 00:32:37 - 00:32:41 if you look at the economy, you know, whether it's a county or a state, 00:32:41 - 00:32:44 we've centralized so much wealth in a way that the 00:32:44 - 00:32:48 government has been financing, pouring a lot of money into 00:32:48 - 00:32:51 creating corporate monopolies. And it's dramatically lowered the 00:32:51 - 00:32:54 possibility of the wealth creation that can come from this 00:32:54 - 00:32:58 technology. And that's why, um, freeing the 00:32:58 - 00:33:01 technology and freeing the local 00:33:01 - 00:33:03 markets, if you will, and the local capital 00:33:03 - 00:33:06 circulation is, you know, they need each other, 00:33:07 - 00:33:10 you need one to do the other. Totally agree. That's 00:33:10 - 00:33:13 why I'm completely aligned with your, some of the conversations you've had and how 00:33:13 - 00:33:16 critical it is. And they can be synergistic. They can support each 00:33:16 - 00:33:20 other. If, for example, every jurisdiction had 00:33:20 - 00:33:23 an AI that was supportive of them. I can almost imagine a chamber of commerce where 00:33:23 - 00:33:27 there was a the Chamber of Commerce, if its purpose was essentially 00:33:27 - 00:33:30 ensure there was an open weight model that every business could use, 00:33:30 - 00:33:33 you know, just by fabulously by support, by conversations. I go to 00:33:33 - 00:33:37 countries and talk to ministers, economic ministers, and try to help 00:33:37 - 00:33:40 them understand that if they're going to support their economy and grow it 00:33:40 - 00:33:44 in the in the next century, they have to start 00:33:44 - 00:33:47 with AI that's under their control, AI that they can then post 00:33:47 - 00:33:50 train. And the phenomena, the amazing phenomena is that that's that's 00:33:50 - 00:33:53 happening. I think the United States has to get on top of 00:33:53 - 00:33:57 this. You know, China, uh, is doing some interesting things or 00:33:57 - 00:34:00 promoting these open way models. Uh, they're more free 00:34:00 - 00:34:03 essentially in that regard than we are in the United States in terms of their 00:34:03 - 00:34:06 ability to do that. Now, I, I think it's really important we have a 00:34:06 - 00:34:10 diversity of open models, because ultimately what goes 00:34:10 - 00:34:13 into those models is the data that the, the, the 00:34:13 - 00:34:16 curators decide. And that data ultimately defines the 00:34:16 - 00:34:20 interpolation surface that your prompt will be oriented 00:34:20 - 00:34:23 to. And then now the good news about open weight 00:34:23 - 00:34:26 models. You might start with an open weight model from China. And then many 00:34:26 - 00:34:29 people are concerned that, oh, that means it's going to give me wrong answers, or it's going to send 00:34:29 - 00:34:33 me code that that dials home. There's a lot of FUD about that, a lot 00:34:33 - 00:34:36 of fear, uncertainty, doubt, a lot of fear of, oh, I, I, I, 00:34:36 - 00:34:40 I can't touch these models. Well, you, you 00:34:40 - 00:34:43 can always retrain them. I mean, you start with the open weight models and 00:34:43 - 00:34:46 then start training it with new data. It's called post training 00:34:46 - 00:34:50 and it's instruction following. It's actually what led to, you know, 00:34:50 - 00:34:53 any raw trained model it we're not going 00:34:53 - 00:34:56 to like it very much as humans. And so there's a post training that every 00:34:56 - 00:35:00 model creator, that model distributor does. OpenAI does it differently 00:35:00 - 00:35:02 than anthropic. They all pretty much do the same thing on the 00:35:03 - 00:35:06 pre-training, but on the post trainings where their differentiation 00:35:06 - 00:35:09 arises. And that's okay. What instructions, what what questions? 00:35:09 - 00:35:12 Answer pairs? Do I give it the final in the final 00:35:12 - 00:35:16 answer so that post training can be done. So 00:35:16 - 00:35:19 in every organization could do it. And it's actually not as expensive as training as 00:35:19 - 00:35:23 the as initial training. This is why thinking machines, there's a 00:35:23 - 00:35:26 company that just showed up, was able to put out a 2.8 trillion 00:35:26 - 00:35:29 parameter model, a very large model. Uh, not sure if 00:35:29 - 00:35:33 it was Kimmy. Actually, I think there's a billion parameter 983 million parameter model 00:35:33 - 00:35:36 from thinking machines lab. They made an open source model where 00:35:36 - 00:35:39 they started with kind of some open weight models, then just did 00:35:39 - 00:35:43 instruction following, just trained it to, to in a way 00:35:43 - 00:35:46 they want. Uh, and so this is happening right now 00:35:46 - 00:35:49 every, I mean, it's, it's critical, I think that people get in, get in 00:35:49 - 00:35:52 front of that. That is going to require 00:35:52 - 00:35:56 compute, distributed compute that is going to require power and 00:35:56 - 00:35:58 energy. And how that comes to, I mean, I think the 00:35:58 - 00:36:02 battle of the data centers is interesting one, but it's also kind of 00:36:02 - 00:36:06 been framed according to the large, uh, AI 00:36:06 - 00:36:09 providers. They've framed the conversation, which is, you know, 00:36:09 - 00:36:13 my, in my view, a distributed AI economy has data centers 00:36:13 - 00:36:16 that are distributed. Right? Exactly. You know, let's 00:36:16 - 00:36:19 say you have eight machines, you have a rack. There could be, you know, 00:36:20 - 00:36:23 thousands of data centers in a city. There could be 00:36:23 - 00:36:26 millions across a state, there could be billions 00:36:26 - 00:36:29 across the world, essentially micro data centers, you know, micro 00:36:29 - 00:36:33 data centers instead of and that are that nobody cares. So it 00:36:33 - 00:36:36 seems to be much stronger and much more resilient to do it that way. 00:36:36 - 00:36:40 Hugely. I mean, this is one of the challenges. You just make a target 00:36:40 - 00:36:43 for your. So the whole point of the internet was to 00:36:43 - 00:36:46 distribute, uh, connectivity and communication. So it 00:36:46 - 00:36:49 was less easy to be targeted. And why are we 00:36:49 - 00:36:53 centralizing it yet again? It's a, oh, I, I, I have 00:36:53 - 00:36:56 a, I, as you know, I have an opinion on why we're centralizing, but let me 00:36:56 - 00:37:00 jump here. One of my favorite books several, many years, about five 00:37:00 - 00:37:03 years ago was a Chinese venture capitalist who wrote a 00:37:03 - 00:37:06 book called AI superpowers and talked about the danger of 00:37:07 - 00:37:10 the United States and China just taking such an extraordinary 00:37:10 - 00:37:13 lead in this technology that it would leave the rest of the world behind. 00:37:14 - 00:37:16 Now we see China, uh, 00:37:17 - 00:37:20 saying at least that they're that they don't want that, that they're 00:37:20 - 00:37:23 going to try and, and push, you know, open source 00:37:23 - 00:37:27 around the world, decentralized around the world and get other people 00:37:27 - 00:37:31 up to speed on this technology. So do you 00:37:31 - 00:37:34 think we're at risk of sort of two AI superpowers 00:37:34 - 00:37:37 and, and the rest of the world left behind? What how do you think it's 00:37:37 - 00:37:41 going to go? There is some risk of that, basically, depending 00:37:41 - 00:37:44 on how the power, energy and financial 00:37:44 - 00:37:48 worlds evolve, or frankly, because there's not 00:37:48 - 00:37:51 a technology risk because, um, as long as people 00:37:51 - 00:37:54 have access to some amounts of distributed power and energy and some 00:37:54 - 00:37:58 amounts of capital, but there are, you know, some, some tension 00:37:58 - 00:38:01 points, some choke points in terms of, uh, now on the 00:38:01 - 00:38:04 compute and power side, um, there's enough you 00:38:04 - 00:38:08 can do even with the current models that exist 00:38:08 - 00:38:11 in a distributed way, which doesn't require a lot of power, doesn't require a lot of compute. 00:38:12 - 00:38:15 So I feel like that, that's just a matter of timing. Is that going to be, 00:38:15 - 00:38:19 you know, in a, is it going to take just a year or is it going to take, you know, five 00:38:19 - 00:38:22 years? Uh, that's going to happen. But on the financial 00:38:22 - 00:38:25 side, you know, we, we still live in a very brittle economy 00:38:25 - 00:38:28 financially, unfortunately. And that could absolutely 00:38:28 - 00:38:31 disrupt, you know, uh, the distribution of 00:38:31 - 00:38:34 prosperity and wealth. The thing I don't understand, um, 00:38:35 - 00:38:38 we all, we're all somewhat familiar. Most of us haven't studied in 00:38:38 - 00:38:41 depth, but we are aware of the concept of the dark ages and medieval period. 00:38:41 - 00:38:44 And, and that was a period where wealth was concentrated. 00:38:44 - 00:38:48 You know, people didn't own their land. They didn't own their own 00:38:48 - 00:38:51 ability to create. And that led to, you know, even for 00:38:51 - 00:38:54 the most wealthy people in that society, they didn't live as well as some 00:38:54 - 00:38:57 people live today. And so it's, 00:38:58 - 00:39:01 it's, it's, it's antithetical to the prosperity of even the people in 00:39:01 - 00:39:04 charge for them to want this. It's going to 00:39:04 - 00:39:08 result in a world that they don't want to be a part of. Uh, so it's, 00:39:08 - 00:39:11 it's, it's a much better idea to expand 00:39:11 - 00:39:14 humanity, to expand humanity's ability to, to, to 00:39:14 - 00:39:17 partner, to share, to invent, to create. Uh, 00:39:17 - 00:39:20 the possibilities are amazing. I'm actually a huge believer in 00:39:20 - 00:39:24 optimist in the possibility of AI supported, because AI is 00:39:24 - 00:39:27 just tools, intelligent tools supported 00:39:27 - 00:39:30 prosperity, but that has to be done in a distributed way. That has to 00:39:30 - 00:39:34 be done by focusing again on, on the distribution that made 00:39:34 - 00:39:37 it possible. And that's mainly one thing I want to make people understand 00:39:37 - 00:39:40 is AI is only possible because the work of 00:39:40 - 00:39:43 millions of people that came before, that provided the 00:39:43 - 00:39:46 foundation and open source foundation for these latest 00:39:46 - 00:39:50 iterations to sit on top of. So the other thing that's happening, 00:39:50 - 00:39:53 Travis, is that many people do not 00:39:53 - 00:39:56 understand the tool or how to use it effectively. Yeah. 00:39:56 - 00:39:59 So where, where do you go to get that education? If, if I 00:39:59 - 00:40:03 have, if I have, you know, college kids or 00:40:03 - 00:40:06 kids or grandkids that I want to make sure get an, 00:40:07 - 00:40:10 a solid grounding on how to use and not 00:40:10 - 00:40:13 misuse the tool. How do I do that? It's a good question. 00:40:13 - 00:40:16 I think, I think fundamentally one is, is start to use 00:40:16 - 00:40:20 it, recognize that, uh, where is your data 00:40:20 - 00:40:23 going? I think part of it is just asking that question. You know, you can even 00:40:23 - 00:40:27 ask AI that question. You know, where is my data going when I, when I talk to you now, 00:40:27 - 00:40:30 it might lie to you or hallucinate as they call it. It's not really lying. It's 00:40:30 - 00:40:33 sort of implies intent, but it might, you know, not tell you 00:40:33 - 00:40:37 exactly truth, but, uh, just you need to be aware of it's going there's 00:40:37 - 00:40:40 actually lots of places to go today. Uh, there's, there's a, there's something 00:40:40 - 00:40:44 called open router. We're trying to make that available at open teams 00:40:44 - 00:40:47 via free, a free download called, you know, Colab. So look for a 00:40:47 - 00:40:50 free download. We're trying to help folks, but there are other free downloads you can get 00:40:50 - 00:40:53 already to get access to, uh, models that are 00:40:53 - 00:40:56 out there. You know, there's Hermes agent, there's Jan, 00:40:56 - 00:40:59 there's just go to open router. Uh, so, 00:40:59 - 00:41:03 um, there are, especially over the next several months, you're 00:41:03 - 00:41:06 going to have more and more of this availability thinking machines lab 00:41:06 - 00:41:10 is going to be, they have open source. They have a lot of money. They have open source 00:41:10 - 00:41:13 models. They're going to be promoting how they make it. Yeah. Open 00:41:13 - 00:41:15 router. There's there's morpheus.ai 00:41:16 - 00:41:19 that morpheus.ai is a open source 00:41:19 - 00:41:23 ecosystem of model providers that lets you find, you 00:41:23 - 00:41:26 know, where are the machines out there? I can run an open model 00:41:26 - 00:41:28 on. Um, and so, 00:41:30 - 00:41:33 and then just start experimenting, start using, just being aware, I 00:41:33 - 00:41:36 think is one of the big things. Um, right now that 00:41:36 - 00:41:40 is the, that is the trade off that's happening is the big, uh, 00:41:40 - 00:41:43 the, the big monopolists essentially are trading 00:41:43 - 00:41:46 convenience for autonomy. And that's always been the trade off. And so 00:41:46 - 00:41:50 they're making it the easiest, the easiest thing to do is go to their website type 00:41:50 - 00:41:53 things in. You get an answer. And they've, they've also provided desktop apps you 00:41:53 - 00:41:56 can download. And it takes a little bit of work to produce that 00:41:56 - 00:41:59 infrastructure. And so they're spending the investment money to 00:41:59 - 00:42:03 produce that infrastructure so that you, as the consumer become dependent 00:42:03 - 00:42:06 on their delivery of that. Right. And so 00:42:07 - 00:42:10 being aware that's happening and then looking for alternatives, 00:42:11 - 00:42:14 and then having conversations with other providers in the open 00:42:14 - 00:42:17 source ecosystem, which there are many that are growing. So I'm excited about that. Obviously 00:42:17 - 00:42:21 an open teams, we're happy to help. But I also understand the problem is so big that 00:42:21 - 00:42:24 we're not going to be the only ones we actually need to promote an ecosystem of, 00:42:24 - 00:42:28 of providers, not just one. Uh, in fact, I say this all the 00:42:28 - 00:42:31 time, open teams, we are doing this so we can provide an example. 00:42:31 - 00:42:34 Uh, but we are not going to be the only ones. In fact, that would be an 00:42:34 - 00:42:38 anti-pattern if we do become the only ones providing this support. 00:42:38 - 00:42:41 Uh, we need to be supporting an example, a pattern that others can follow. 00:42:41 - 00:42:45 And then, uh, you know, we'll have our, uh, we're eager 00:42:45 - 00:42:48 to build trust relationships with folks so that we can get the help they need, because 00:42:48 - 00:42:51 we're not saying, you know, go off and do it on your own. We're saying get 00:42:51 - 00:42:55 the help you need, but get it from trusted sources that are 00:42:55 - 00:42:58 not at the same time trying to also, you know, become your AI 00:42:58 - 00:43:01 overlord that are trying to basically become your AI supporter. So that AI 00:43:01 - 00:43:05 becomes a tool that you use that continues to help you 00:43:05 - 00:43:08 and that you stay in charge of so poorly. 00:43:08 - 00:43:12 Questions. Well, I have lots of well, I mean, 00:43:12 - 00:43:15 really, I, I think my main concern is Children's Health Defense is, 00:43:15 - 00:43:18 is, you know, as I said before we went live, AI is brilliant. And I don't 00:43:18 - 00:43:22 think anyone can say that it's not, but it's also dangerous. 00:43:22 - 00:43:25 So, so how do we protect our, 00:43:26 - 00:43:29 our I know Kathryn's kind of asked that already, but our children who are 00:43:29 - 00:43:32 just have access to everything in this day and age. 00:43:32 - 00:43:36 Yeah, it's a great point. I think this is another example where already our children have 00:43:36 - 00:43:39 been exposed to eyes that have been training them 00:43:39 - 00:43:43 to think a particular way or effectively. It's a dumb AI. Dumb AI has 00:43:43 - 00:43:46 already hurt our children with the social media, effectively training 00:43:46 - 00:43:49 them just to be, you know, give them information they already want. You 00:43:49 - 00:43:52 know, one of the things that I've, I've six children. One of the things 00:43:52 - 00:43:56 children, uh, when they're, when they're, they're very eager learners, 00:43:56 - 00:43:59 very eager participants in the world, but they'll just pursue things, 00:43:59 - 00:44:03 you know, very linearly, very, you know, low hanging fruit. And there are definitely things 00:44:03 - 00:44:06 that are bad patterns, you know, oh, this tastes good. I should eat it. Well, 00:44:06 - 00:44:09 not if it's going to be addictive and cause you a lifetime of 00:44:09 - 00:44:13 regret, right? There's sort of certain choices that are really bad. 00:44:13 - 00:44:16 Oh, I like this, this, this, this, this feed that makes me 00:44:16 - 00:44:20 angry, that shows me all the bad in the world. I really like it. Well, yes, but not if it affects 00:44:20 - 00:44:23 your mental health or, you know, I like seeing everybody's successes that 00:44:23 - 00:44:26 social media is sending me. Well, not if it's going to cause you to feel bad about yourself 00:44:26 - 00:44:30 and and not recognize your own gifts. Like, this is the thing that social 00:44:30 - 00:44:33 media has been doing to our children already. And so it 00:44:33 - 00:44:36 requires some discipline, you know, some intentional discipline. I 00:44:36 - 00:44:40 think that's right. Now, kind of the message it's there's a little bit of 00:44:40 - 00:44:43 intentional discipline required right now. So you don't become 00:44:43 - 00:44:46 addicted to the, uh, the drug that's being 00:44:46 - 00:44:49 fed to you by this really cool tool that's just seems free and 00:44:49 - 00:44:53 open, but has these consequences buried in it, which 00:44:53 - 00:44:56 are giving away your autonomy, giving away your context, giving away your 00:44:56 - 00:44:58 information to some other, to the 00:44:58 - 00:45:02 provider. The way we've done with Google, the way we've done 00:45:02 - 00:45:06 with, uh, so many of these applications that have, uh, 00:45:07 - 00:45:09 come into our lives and have these unintended consequences. 00:45:09 - 00:45:13 Doomscrolling. Right. We've all seen it. We've all participated in 00:45:13 - 00:45:16 this, in this, um, negative outcome of 00:45:16 - 00:45:19 dumb AI. Like this is AI that is just simple models, right? 00:45:19 - 00:45:23 Not even in very intentional, they have these negative outcomes. So I 00:45:23 - 00:45:26 think it's important just to, um, you know, as a 00:45:26 - 00:45:29 parent, as anybody who cares about children, you have to 00:45:29 - 00:45:32 recognize that children need guidance, children need guidelines, any guidelines from 00:45:32 - 00:45:36 loving people, not guidelines from a nanny state, not guidelines from super 00:45:36 - 00:45:39 organization that knows everything, not some platonic, 00:45:39 - 00:45:43 you know, child rearing ecosystem. You need guidelines 00:45:43 - 00:45:46 from loving trusted connections. 00:45:47 - 00:45:50 And, uh, and we need the same for AI. You know, think of AI 00:45:50 - 00:45:53 like a child because it basically, in many ways there's 00:45:53 - 00:45:56 a metaphor there. It's not the same thing as a child, but it is this 00:45:56 - 00:45:59 intelligent system that is that requires 00:45:59 - 00:46:02 some oversight, requires some accountability, requires some 00:46:02 - 00:46:06 guidance. And it's really important that we don't have this Lord of the 00:46:06 - 00:46:09 flies situation where, you know, the children are put in charge of 00:46:09 - 00:46:13 everybody else, and the children being both the some of the 00:46:13 - 00:46:16 misaligned incentives of the investment community that have 00:46:16 - 00:46:20 concentrated in certain parts of the world, and then also 00:46:20 - 00:46:23 the intelligence systems they've produced that are also like children. 00:46:24 - 00:46:27 And the. And so when you hear safety, you'll hear 00:46:27 - 00:46:30 that, you know, Orwellian trust and safety, you'll hear it, oh, we need this 00:46:30 - 00:46:34 to for safety. What are the guardrails? Yeah, I agree, we need guardrails. 00:46:34 - 00:46:37 Just like children need guardrails. How do we do that? We don't do it by 00:46:37 - 00:46:40 creating the concentrated, you know, mother of 00:46:40 - 00:46:43 all. You don't do that with children. So with 00:46:43 - 00:46:47 AI it's the same. How do we put guardrails on AI? We put it in the 00:46:47 - 00:46:50 hands of accountable intelligence, accountable people. 00:46:50 - 00:46:54 And so we distribute the power into, uh, millions of 00:46:54 - 00:46:57 people and use the same trust accountability network. So we've already 00:46:57 - 00:47:00 established over ten 000 years in order to constrain it. 00:47:00 - 00:47:04 So guardrails have to come from people, and it needs to 00:47:04 - 00:47:07 come from millions of people. We're not going to get 00:47:07 - 00:47:10 there by creating a new organization, a nanny state that helps guard 00:47:10 - 00:47:14 it. And that's the that's currently right now. But 00:47:14 - 00:47:17 it's so tempting to do that. It's so tempting to think that there's 00:47:17 - 00:47:21 going to be some super genius that'll help us. And they have all the 00:47:21 - 00:47:24 powers. They'll protect us that, that, um, 00:47:24 - 00:47:27 that, that whisper, that, that temptation 00:47:27 - 00:47:31 hasn't panned out for thousands of years. It's not 00:47:31 - 00:47:33 going to pan out here either. But I point out is to 00:47:33 - 00:47:37 distribute the capability of this tool in the hands of millions of people. 00:47:37 - 00:47:40 I don't think I ever told you this, Travis, but when, uh, 00:47:41 - 00:47:44 when Carolyn and I were doing Hamilton Securities 00:47:44 - 00:47:47 Group in Washington, we had hired AT&T Bell 00:47:47 - 00:47:50 Laboratories to help us, uh, 00:47:50 - 00:47:54 process mortgage auctions so that we 00:47:54 - 00:47:57 could let different industries sell, stratify the 00:47:57 - 00:48:00 packages and bid against Each other, you know? It just took 00:48:00 - 00:48:03 incredible compute power to look at all the possible 00:48:03 - 00:48:06 permutations and figure out which made the most for 00:48:06 - 00:48:10 taxpayers anyway. So I spent a lot of time at AT&T Bell 00:48:10 - 00:48:13 Laboratories trying to sort of translate 00:48:13 - 00:48:16 their optimization software for 00:48:16 - 00:48:20 routing telephone calls and IT to mortgage auctions. And it 00:48:20 - 00:48:23 turned out that that they had applied it to airline scheduling for many 00:48:23 - 00:48:26 decades. And after many decades of doing highly complex 00:48:27 - 00:48:30 airline scheduling, where all the personnel are giving you their 00:48:30 - 00:48:33 preferences and what they want and, and you 00:48:33 - 00:48:36 optimizing and doing the schedules, they were 00:48:36 - 00:48:39 able to, they had the data to analyze the cost of 00:48:39 - 00:48:42 a rule. What did it cost the company for the 00:48:42 - 00:48:45 unions or the government to have rules that affected the 00:48:45 - 00:48:48 scheduling? And one of the things they concluded was 00:48:48 - 00:48:52 that the privileges afforded to senior 00:48:52 - 00:48:55 pilots so constrained the model that 00:48:55 - 00:48:58 if the pilots gave up that rule, the pilots would get 00:48:58 - 00:49:02 better scheduling for themselves. And so 00:49:02 - 00:49:05 their requirements of privilege was so 00:49:06 - 00:49:09 killing the optimization that even the pilots were worse 00:49:09 - 00:49:12 off. Wow. I love that story. 00:49:13 - 00:49:16 Yeah. So that's what it feels like every day as I'm 00:49:16 - 00:49:19 watching the the people who are arguing for 00:49:19 - 00:49:22 control, the control model, you know, 00:49:22 - 00:49:25 I'm saying they're so shrinking the optimization, they're going to shrink 00:49:26 - 00:49:28 their own. But that, you know, tier I 00:49:29 - 00:49:32 words cannot express to you how mathematically expensive tyranny is. 00:49:33 - 00:49:36 It's just it's, it's even more than you would intuitively guess. 00:49:37 - 00:49:40 So I think Kevin has a question. Carolyn. 00:49:40 - 00:49:44 Yeah. I have a couple of questions. One is let let's 00:49:44 - 00:49:46 go down to the regular people level. 00:49:48 - 00:49:51 As I understand it, a. And you're in this 00:49:51 - 00:49:54 business, you can build your own AI within a 00:49:54 - 00:49:57 company and you start with some kind of open 00:49:57 - 00:50:00 thing. But, um, you could 00:50:00 - 00:50:03 train. And as I understand it, part of 00:50:03 - 00:50:07 training has to do with what 00:50:07 - 00:50:10 are your sources? And I've done some 00:50:10 - 00:50:13 work with Claude and, you know, we're 00:50:13 - 00:50:16 interested in certain things that are not in Wikipedia. 00:50:16 - 00:50:19 Now, I understand Wikipedia is a major source 00:50:20 - 00:50:23 for Claude or the search engine or the the 00:50:23 - 00:50:26 AI's that we have access to. If I want something that would 00:50:26 - 00:50:30 be on, on children's health defense, it does 00:50:30 - 00:50:33 not pick that up. But if I say go to Children's 00:50:33 - 00:50:36 Health defense, which I have done, and here's an 00:50:36 - 00:50:40 article and then I ask it, does your training 00:50:40 - 00:50:43 pick up that you just admitted this from an 00:50:43 - 00:50:46 article that you don't usually look at, and it says, oh, no, 00:50:46 - 00:50:49 that that doesn't go into my training. It 00:50:49 - 00:50:52 remembers, if you will excuse that word, 00:50:52 - 00:50:56 it remembers what I've been working on and 00:50:56 - 00:50:59 have pointed it toward to look at. And 00:50:59 - 00:51:02 it will continue to use that information 00:51:03 - 00:51:06 in further work that I'm doing, but it 00:51:06 - 00:51:08 doesn't learn that information. So as I 00:51:08 - 00:51:12 understand it, you could create an in-house 00:51:12 - 00:51:15 AI and put in children's health, defense and 00:51:15 - 00:51:19 all the alt media things and have 00:51:19 - 00:51:22 completely different results from what you get from 00:51:22 - 00:51:25 Claude or ChatGPT. Is that right? Yeah. That's right, that's 00:51:25 - 00:51:28 correct. There's actually three levels you described in that in that 00:51:28 - 00:51:31 description, three levels of training. You've got what's called 00:51:31 - 00:51:35 pre-training, which is these raw data that was used to train 00:51:35 - 00:51:38 the original weights. And you're correct that, you know, it doesn't have 00:51:38 - 00:51:41 proprietary information like, say, from children's health, health, health, uh, 00:51:41 - 00:51:44 um, it doesn't have that information. You've also got the 00:51:44 - 00:51:48 instruction following data. This is the list of questions, answers 00:51:48 - 00:51:51 that are particularly oriented to help orient the 00:51:51 - 00:51:54 model, to be more human, to be more responsive to people. That's 00:51:54 - 00:51:58 another level of training. It's called fine tuning or just fine tuning of the 00:51:58 - 00:52:01 model. So you've got the pre-training that gives it the basic substrate, 00:52:02 - 00:52:05 you've got the fine tuning, and then you have the outer 00:52:05 - 00:52:08 context. This is what you're describing that Claude has. When 00:52:08 - 00:52:12 you give it information, it remembers. It's the it's the memory that it 00:52:12 - 00:52:15 has about you. And that's actually one of the critical things right now 00:52:15 - 00:52:18 that the big models Claude and OpenAI are doing to 00:52:18 - 00:52:22 addict you to their model to make it hard to switch to something else is 00:52:22 - 00:52:25 they're essentially creating a model of you that they that they 00:52:25 - 00:52:28 only hold that you can't get out, you don't get access to 00:52:28 - 00:52:32 that helps them, helps that model respond better to 00:52:32 - 00:52:35 you. Right? It's an important thing, actually. And so it's one of the critical 00:52:35 - 00:52:39 things happening right now and why the sooner we can do this, the easier 00:52:39 - 00:52:42 it'll be to move people away from that addictive. 00:52:42 - 00:52:46 Um, because basically like a friend, I mean, it's the same thing that friends 00:52:46 - 00:52:49 do, friends learn about you and you like to be around them because they know you. 00:52:49 - 00:52:53 And it's not as hard to kind of pre-train them. Well, that's what's 00:52:53 - 00:52:56 Claude's doing. Claude's learning about you, and it's responding to you 00:52:56 - 00:52:59 in a similar way that a friend would. Now that's fine, as 00:52:59 - 00:53:03 long as it. It doesn't become this addictive relationship you can't get away 00:53:03 - 00:53:06 from, uh, because it isn't trained on the 00:53:06 - 00:53:09 foundational model you need it to. That's the opportunity right now. There's these three 00:53:09 - 00:53:12 ways you can take and take open weight models. 00:53:12 - 00:53:16 Great. Trained on the internet. Fine. You don't need to. You can you 00:53:16 - 00:53:19 can pre-train on your data if you'd like. Often just the 00:53:19 - 00:53:23 fine tuning step is enough to orient that massive data 00:53:23 - 00:53:26 set to your specific concerns. And you don't want that data set 00:53:26 - 00:53:30 uploaded to the internet. You want that data set to be specific to you, and now you have a 00:53:30 - 00:53:33 model that's private, a model that you can host, the model you don't have to give away. You keep it 00:53:33 - 00:53:36 on your servers, and it becomes the foundation for your 00:53:36 - 00:53:40 chats, for your conversations, for what you give to internally, what you give 00:53:40 - 00:53:43 externally. And you can even, you know, change those to, 00:53:43 - 00:53:46 in fact, you could have a foundation model and then little 00:53:46 - 00:53:49 adjustments to that model that you serve up to different categories of 00:53:49 - 00:53:53 people, different personas, different tranches of people you're trying to 00:53:53 - 00:53:56 help. This is the world that we should have. We should have billions of 00:53:56 - 00:54:00 models. We couldn't there be models? 00:54:00 - 00:54:02 Couldn't there be an alt, uh, an 00:54:03 - 00:54:06 alternative medicine AI that anybody 00:54:06 - 00:54:09 could purchase? Yes, absolutely. This could mean 00:54:10 - 00:54:14 legal. It could help every single space of 00:54:14 - 00:54:17 human, you know, exploration and innovation, but 00:54:17 - 00:54:20 only if it's served in a distributed way 00:54:20 - 00:54:23 where. And that's in fact, how we can compete. We can serve, we can help each other. We can 00:54:23 - 00:54:27 cooperate together. Each person, each organization, person, 00:54:27 - 00:54:30 organization, country level has what I call an intelligence 00:54:30 - 00:54:33 hub, a hub of information that integrates to them. 00:54:33 - 00:54:36 And then they share part of that intelligence hub with 00:54:36 - 00:54:40 others and then enriched by others. And that keep in a 00:54:40 - 00:54:43 in a way that essentially mirrors an economy of people trading and serving each 00:54:43 - 00:54:46 other. So there's a massive opportunity here to 00:54:46 - 00:54:49 create a beautiful distributed AI economy. And it 00:54:49 - 00:54:52 starts with recognizing open source, the foundation 00:54:52 - 00:54:56 building, um, intelligence hubs, everybody understand 00:54:56 - 00:54:59 their need to create one for their organization, for their 00:54:59 - 00:55:02 family, for themselves, for their country. And then 00:55:02 - 00:55:06 it's, it's really critical right now. It's not enough just to lean on one 00:55:06 - 00:55:09 person's API or one person's model. The idea that you would 00:55:09 - 00:55:12 rent intelligence from another company is, is ludicrous. 00:55:12 - 00:55:16 It's absolutely ludicrous. Like, it makes no sense. It's like renting your 00:55:16 - 00:55:19 autonomy. No, you need to take this capability and bring it 00:55:19 - 00:55:23 in. And, uh, we're here to help that happen and here to train 00:55:23 - 00:55:26 many people to help that happen. I see a world where, you know, 00:55:26 - 00:55:29 thousands and thousands of folks, it's like, it's like the PC 00:55:29 - 00:55:32 revolution. Remember when PCs, you know, the computers are all in the 00:55:33 - 00:55:36 mainframes and we didn't get access to them. And then PCs became 00:55:36 - 00:55:40 out there and everybody had a PC on every desktop. We 00:55:40 - 00:55:42 need an AI in every organization, 00:55:43 - 00:55:46 right? I you also, you need, you need sovereignty, 00:55:46 - 00:55:50 you need AI sovereignty, and you need data sovereignty. Yeah. 00:55:50 - 00:55:53 Internal to the person or the organization. It starts with 00:55:53 - 00:55:56 protecting your data. So that's how do you do that? If you're 00:55:56 - 00:56:00 using what, how do you use that? Do that. If you're using 00:56:00 - 00:56:03 Claude or chat, you're not, you're not if you're using Claude or chat, 00:56:04 - 00:56:07 you're not protecting your data because essentially you're, and you know, 00:56:07 - 00:56:10 a little bit, that's okay. You know, maybe you don't care about some of your data going out, but you 00:56:10 - 00:56:14 just have to recognize that's what you're doing. You have to recognize that if you're sharing 00:56:14 - 00:56:17 information with Claude or ChatGPT, it is not yours 00:56:17 - 00:56:20 anymore. In fact, you know, in the United States, the law has recognized that 00:56:20 - 00:56:24 if, if if a lawyer uses ChatGPT, they violated attorney 00:56:24 - 00:56:27 client privilege effectively. They, they no longer have it 00:56:27 - 00:56:30 because they've, they've sort of opened the door and had someone listen in. 00:56:31 - 00:56:34 And so that's why a lot of law firms are recognizing this and saying, oh, we gotta have our own. 00:56:34 - 00:56:37 And they do. Right. So. So is it 00:56:37 - 00:56:41 possible? So the, the number one way that I would use 00:56:41 - 00:56:44 an AI is for search. Yeah. It's a good search 00:56:44 - 00:56:47 engine. Yeah. And we have a great researcher who checks 00:56:47 - 00:56:50 everything. So we use it for search. Then we check 00:56:50 - 00:56:54 everything. And that's pretty good. You're usually not giving away a lot when you're 00:56:54 - 00:56:57 just searching, but you're giving away a little bit. You're giving away a little bit. Is it 00:56:57 - 00:57:00 possible to build your own AI that can be 00:57:00 - 00:57:04 sufficiently trained using the open source AI 00:57:04 - 00:57:07 so that you it'll function as a search engine for 00:57:07 - 00:57:11 you? Yeah, that's the good news right now. That is very doable. 00:57:11 - 00:57:14 Now and then you just have to we're preparing what the cost of that would be 00:57:14 - 00:57:17 because you have to, you know, for it to be truly, truly yours, 00:57:17 - 00:57:21 you have to host it on, on hardware you control to that hardware can be 00:57:21 - 00:57:24 rented hardware. I'm not saying you have to, you know, be or have your own buildings. 00:57:24 - 00:57:28 Actually, the cloud did a lot of work to get enterprise comfortable with the fact that they're just 00:57:28 - 00:57:31 renting the compute, not giving away their knowledge. So there are lots of 00:57:31 - 00:57:34 ways to have hosted compute that you can then 00:57:34 - 00:57:37 have. That's your data, but you put the models up there. It takes a little bit of 00:57:37 - 00:57:41 elbow grease, but that's pretty straightforward. And so the cost of that has come down 00:57:41 - 00:57:44 precipitously. It is, um, it's still a, it's a little 00:57:44 - 00:57:48 bit more than 200 a month, right. Per person, you know, for a single 00:57:48 - 00:57:51 person. So if you're a group of 4 or 5 people, you're used to spending 00:57:51 - 00:57:54 essentially a thousand a month to get access to cloud, that's a pretty good deal 00:57:55 - 00:57:58 because it might cost you 4000 a month, 5000 a month, for 00:57:58 - 00:58:01 example, to host it yourself. So that's currently, I think, 00:58:01 - 00:58:04 what the the opportunity and challenge is, is to understand that 00:58:04 - 00:58:08 trade off of renting versus owning, you know, there's some 00:58:08 - 00:58:11 financial engineering that can take place because for many people, it's 00:58:11 - 00:58:14 worth it maybe to, you know, because essentially you can pay that 00:58:14 - 00:58:17 cost, maybe get a loan to pay some of that cost. 00:58:18 - 00:58:21 So just before we close, because we only have 1.5 00:58:21 - 00:58:24 minutes, one of the most important things you've said, 00:58:24 - 00:58:27 I think, and critical, I want to bring this point home is 00:58:28 - 00:58:31 don't fall for the fear tactics. Yeah. Could you 00:58:31 - 00:58:34 could you repeat that again? Do not fall for the fear tactics. 00:58:35 - 00:58:39 It's a very common refrain. And so I think you 00:58:39 - 00:58:42 have to fall. You have to lean into hope, lean into people 00:58:42 - 00:58:45 around you, lean into love, basically, you know, 00:58:46 - 00:58:49 lean into love. That's what drives human flourishing, uh, 00:58:49 - 00:58:52 leading to love. Lean into faith. Um, the fear 00:58:53 - 00:58:56 is what's going to try to get you controlled. And it's 00:58:56 - 00:58:59 easy because we all do it. I mean, fear is 00:59:00 - 00:59:03 pretty easy to hit on, and it's an important thing to help you avoid things you 00:59:03 - 00:59:06 shouldn't be doing. But, you know, fear, fear, 00:59:07 - 00:59:10 fear the fear. That's like, thank you so 00:59:10 - 00:59:14 much for your education. It's been utterly amazing today. And on that 00:59:14 - 00:59:17 note, to thank our viewers and your comments, we really 00:59:17 - 00:59:21 appreciate you and to be kind to each other. It costs nothing to be kind to 00:59:21 - 00:59:24 each other, and there's so much passive aggressiveness out there at the 00:59:24 - 00:59:27 moment and competition and nastiness 00:59:27 - 00:59:31 and just that's exhausting. Let it go. Just 00:59:31 - 00:59:34 let it go and be kind to each other. My brothers and sisters, at the end of the day, 00:59:34 - 00:59:36 Travis, you're amazing and thank you so much for being on this morning. 00:59:37 - 00:59:40 Thank you Travis. Thank you. Catherine Polly Tommey 00:59:40 - 00:59:44 Carolyn, great to see you all. Yeah. And to our viewers, 00:59:44 - 00:59:47 pretty big. Stand firm, never settle. And we'll be back with Financial 00:59:47 - 00:59:51 Rebellion. Will we be back next week or is it Meryl? Meryl. 00:59:51 - 00:59:54 Meryl. We'll be back the week after. Meryl. Okay, we'll see you then. 01:00:10 - 01:00:13 All right. So we're live. Um, can you tell us your name, please? 01:00:13 - 01:00:16 My name is Theodora Cohen. And this is a story about 01:00:16 - 01:00:20 your tenant, my tenant, a young couple, and they lost their 01:00:20 - 01:00:23 baby in 2016. The baby's name was 01:00:23 - 01:00:26 Brian Morales. He wasn't even a year 01:00:26 - 01:00:30 old. And they lived in the back of my 01:00:30 - 01:00:33 house in a cottage. And, uh, one day I saw all these 01:00:33 - 01:00:36 officers and police cars going up my driveway, 01:00:36 - 01:00:40 and I ran back there with them. And little Brian, 01:00:40 - 01:00:43 who had already passed away in his sleep while taking a nap in the 01:00:43 - 01:00:47 back cottage. Uh, they had him out on the grass trying to 01:00:47 - 01:00:50 do CPR on him, but he was white as a ghost, and I knew he 01:00:50 - 01:00:54 was gone. And, uh, with that, they 01:00:54 - 01:00:57 took him away. And, um, I had to call the 01:00:57 - 01:01:01 dad who was working a young couple. And, uh, 01:01:01 - 01:01:04 it was really heartbreaking. They had a little daughter. 01:01:04 - 01:01:07 This was her younger brother. Okay. With that? I 01:01:07 - 01:01:10 was involved with the opioid crisis, with the, uh, 01:01:11 - 01:01:14 recovery center out here on Long Island. And I got 01:01:14 - 01:01:17 to know the coroner who did the ugly truth for the opioid 01:01:17 - 01:01:20 crisis and all the schools. And I knew him. 01:01:21 - 01:01:24 And I asked the dad, was this 01:01:24 - 01:01:27 baby at the doctors before he 01:01:27 - 01:01:30 died? And he was and he got me the doctor's report of all 01:01:30 - 01:01:33 the vaccines that this little Brian had 01:01:34 - 01:01:37 before a year old, he had 01:01:37 - 01:01:40 four vaccines that week, including MMR. 01:01:41 - 01:01:45 Uh, you can see the circled ones of the four vaccines 01:01:45 - 01:01:48 he just had that week. Can you read those out? And 01:01:48 - 01:01:51 they were the ones that he had that week. He had hepatitis A, 01:01:52 - 01:01:55 he had P influenza, he had p f 01:01:55 - 01:01:58 quad. Uh, he had the MMR. 01:01:58 - 01:02:02 And he also had, uh, the, the varivax, 01:02:02 - 01:02:05 the varicella, the varicella vaccine. 01:02:06 - 01:02:09 And when I looked at the whole list, remember, he wasn't even a year old. He 01:02:09 - 01:02:11 had 16 vaccines 01:02:13 - 01:02:16 before he was a year old. The mom was saying how 01:02:16 - 01:02:19 tired he would always be. And this time he fell asleep 01:02:19 - 01:02:22 in the nap and he never woke up. They 01:02:22 - 01:02:25 considering it. I called, I called the coroner 01:02:26 - 01:02:29 in Suffolk County, Michael Kaplan, and we discussed the 01:02:29 - 01:02:32 case, and I told him that I felt this was a 01:02:32 - 01:02:33 vaccine death. 01:02:35 - 01:02:38 And he says, well, they gave the vaccines according to the 01:02:38 - 01:02:41 schedule. And and I said, no, 01:02:41 - 01:02:45 no, we have to take this further. And it brought the, uh, 01:02:45 - 01:02:48 centers to the New York University Hospital for the 01:02:48 - 01:02:52 parents to have a discussion with them. And they got nowhere 01:02:52 - 01:02:54 with it. They didn't even want to hear the word vaccines. 01:02:55 - 01:02:58 And so, um, to my knowledge, 01:02:58 - 01:03:02 this little baby died as a result. And do you have permission from them 01:03:02 - 01:03:05 to tell this story? Pardon me. Do you have permission from the parents to tell 01:03:05 - 01:03:09 this story? Not yet. Okay. 01:03:09 - 01:03:12 We are live right now, so that might be an issue, but it is 01:03:12 - 01:03:15 tragic. And you see the babies here? I'm. I'm 01:03:15 - 01:03:18 here. I am here because not long 01:03:18 - 01:03:21 after that, uh, there was another baby up not 01:03:21 - 01:03:24 far from us that was in the newspapers 01:03:24 - 01:03:28 that they considered another SIDs case. Uh, 01:03:28 - 01:03:31 toddlers. Sudden death leaves family searching for answers. 01:03:32 - 01:03:35 Uh, that was, uh, in 2018 in the same 01:03:35 - 01:03:38 town out East Riverhead. And, 01:03:38 - 01:03:41 um, yeah, another one. And so, um. 01:03:44 - 01:03:47 There's an organization to, I don't know, you might want to take a look at this 01:03:47 - 01:03:50 SUDCCC it's the, uh, sudden 01:03:50 - 01:03:54 infant deaths and they, they 01:03:54 - 01:03:57 don't know why. And the sudden infant deaths, they just that's what they title it. 01:03:57 - 01:04:01 Now, nobody's really looking into the vaccine issue. Haven't 01:04:01 - 01:04:04 been for a while. Maybe they are now, but 01:04:04 - 01:04:07 um, I will tell the parents that, that I was 01:04:07 - 01:04:11 here and um, they would like 01:04:11 - 01:04:14 some closure, I'm sure. Yeah. Young couple. How 01:04:14 - 01:04:17 long ago did that happen? The 2016, 2016, 01:04:18 - 01:04:21 mid-October 2016. And um, 01:04:21 - 01:04:25 those, um, he got the shots, uh, 01:04:25 - 01:04:28 on here. October 4th, 2016. Is there 01:04:28 - 01:04:30 anything else you want to tell the people? 01:04:32 - 01:04:34 Uh, this is not new. 01:04:36 - 01:04:39 I was in Hippocrates Health Institute in the 01:04:39 - 01:04:43 90s with Brian Clemens, and someone came down with 01:04:43 - 01:04:46 Lenny. Horace was speaking. A woman came down. Her 01:04:46 - 01:04:49 baby died the next day as a result of vaccines. This 01:04:49 - 01:04:53 goes all the way back to the 90s that I know of. That's 01:04:53 - 01:04:56 why I was able to identify this. And I've been 01:04:56 - 01:05:00 keeping up with, um, uh, children's 01:05:00 - 01:05:03 health, defense, high wire. And I watch you, I watch 01:05:03 - 01:05:06 you speak and, um, it has to get out 01:05:06 - 01:05:10 there. It's just covering this up with 01:05:10 - 01:05:13 sudden infant deaths syndromes and, 01:05:13 - 01:05:16 um, but it's even bigger than that. It's bigger. The 01:05:16 - 01:05:20 pharmaceutical companies are, um, 01:05:21 - 01:05:24 have some kind of plan going on to, uh, 01:05:24 - 01:05:28 well, thank you for telling us about your 01:05:28 - 01:05:30 tenants, baby. Mhm. Um, 01:05:31 - 01:05:35 and I hope you know that they understand. Now, are they 01:05:35 - 01:05:38 still your tenants or have they moved? Have they moved 01:05:38 - 01:05:42 or are they still your tenants? They moved. Not only did they move, 01:05:42 - 01:05:45 but they ended up splitting up, which I hear happens a 01:05:45 - 01:05:48 lot from families that lose their children. They ended up. So 01:05:48 - 01:05:52 I'm in touch with the dad still, you know, and I'm going to let him 01:05:52 - 01:05:55 know. Okay. Um, all right. Well, we we 01:05:55 - 01:05:58 we'll wait for that permission. But thank you so much for doing this. 01:05:58 - 01:06:01 Thank you. All right. Thank you. 01:06:03 - 01:06:07 This and all CHD TV programing is made 01:06:07 - 01:06:09 possible by generous support from viewers like you. 01:06:10 - 01:06:14 Please consider making a contribution today to keep our programing going 01:06:14 - 01:06:15 strong. Thank you.59:T788,[‘Smart’ Bassinet for Babies Collects — and Shares — Sensitive Data](https://childrenshealthdefense.org/defender/smart-bassinet-for-babies-collects-and-shares-sensitive-data-snoo/) [Privacy Guides ](https://www.privacyguides.org/en/) [Cheating the System: How AI Data Centers Bypass Water and Air Pollution Laws and Permits](https://childrenshealthdefense.org/defender/ai-data-centers-bypass-water-air-pollution-laws-permits/) [Crime that pays is crime that stays. ](https://x.com/solari_the/status/2079167662316212737) [ Hegseth $37.5 Billion spent estimate](https://x.com/solari_the/status/2079840577202376871) [House adopts Biggs’s amendment to slightly cut defense budget if Pentagon fails to get audit](https://thehill.com/homenews/house/5982387-biggs-pentagon-audit-ndaa/) [Travis Oliphant on X ](https://x.com/teoliphant?lang=en) [Travis Oliphant on Open Source, AI & Building for Everyone ](https://www.youtube.com/watch?v=fhkEIEVcWjc) [Download Linux](https://www.linux.org/pages/download/) [OpenTeams](https://openteams.com/) [Travis Oliphant from Wikipedia](https://en.wikipedia.org/wiki/Travis_Oliphant) [A Discussion of Open-Source AI with Travis Oliphant](https://solari.com/a-discussion-of-open-source-ai-with-travis-oliphant/) [Python: The Documentary](https://solari.com/movie-of-the-week-june-27-2026-python-the-documentary/) [The Story of Python and how it took over the world](https://www.youtube.com/watch?v=GfH4QL4VqJ0) [Why Python Became the Language of AI](https://inferencebysequoia.substack.com/p/why-python-became-the-language-of) [LibreOffice](https://www.libreoffice.org/) [Ubuntu](https://ubuntu.com/) [Applied AI Society](https://appliedaisociety.org/) [Morpheus AI](https://www.morpheus.ai) [OpenRouter](https://openrouter.ai) [AI Superpowers: China, Silicon Valley, and the New World Order](https://www.amazon.com/AI-Superpowers-China-Silicon-Valley/dp/132854639X) 56:["$","$L57",null,{"isEpisode":true,"videoData":{"show":{"__typename":"Show","sys":{"id":"62CSug7PQruJSFf61WQfgd"},"title":"Financial Rebellion with Catherine Austin Fitts","slug":"financial-rebellion-with-catherine-austin-fitts","broadcastChannelId":"a1xbld7jizbgusezhp5g","description":"The revolution will not be centralized. Financial Rebellion is a weekly show featuring Catherine Austin Fitts and attorney Carolyn Betts, general counsel of Solari, Inc. who call upon their years of experience in the financial and legal sectors to provide us with the tools we need to powershift our money and reclaim financial independence from the monopolizing grip of the central banks and digital currency titans.","hosts":{"items":[{"sys":{"id":"5GBlIFHlUY15arDIUbUhDp"},"image":null,"name":"Catherine Austin Fitts"},{"sys":{"id":"4unAPJ4gwBjhtfTE6kIzGp"},"name":"Polly Tommey","image":{"url":"https://images.childrenshealthdefense.org/fpx75ohbqjq3/2PVUrHyGCU4updSj8uK2F7/76ecd58e0666451a59f7be8d7b4423fe/Polly_Tommey_720x940.jpg","width":720,"height":940,"description":"Headshot photo of Polly Tommey"}},{"sys":{"id":"3lJPKvlZ8HMcFwNQesb8bF"},"image":null,"name":"Carolyn Betts"}]},"heroImage":null,"linkedFrom":null},"episode":{"__typename":"Episode","sys":{"id":"2p6zN4DYBnvERIJmVRKv9x"},"title":"Open Source AI: What It Is, Why It Matters With Travis Oliphant","description":"Travis Oliphant makes the following eye-opening argument about the future of AI: Without sovereignty over data, organizations and people lose control over their own decision-making; concentrated power creates a weapon that has the ability to influence an entire population. So what’s the real solution to this threat? Watch this edition of \"Financial Rebellion” to find out!","videoTags":"Catherine Austin Fitts, Travis Oliphant, Carolyn Betts, data collection, privacy, surveillance, tracking, personal information, AI - artificial intelligence, data centers, open AI, OpenTeams, NumPy, Python, SciPy, data scientist, software developer, ","slug":"open-source-ai-what-it-is-why-it-matters-with-travis-oliphant-show-page","publishDate":"2026-07-23T13:50:00.000Z","live":false,"videoPlaybackId":"open-source-ai-what-it-is-why-it-matters-with-travis-oliphant-iajmdgdscft4yslxgy9j","broadcastId":"fxblkplua9w5uiekgsrh","podcastListenUrl":null,"transcript":{"json":{"data":{},"content":[{"data":{},"content":[{"data":{},"marks":[],"value":"$58","nodeType":"text"}],"nodeType":"paragraph"}],"nodeType":"document"}},"resourceLinks":"$59","thumbnailMediaStandard":{"url":"https://images.childrenshealthdefense.org/fpx75ohbqjq3/2MYWcGKxCQG22hWRp6fP9K/fd660dbb4947d500fefb68655a97fadf/072326_FR_Cover.jpg","width":1920,"height":1080},"showsCollection":{"items":[{"__typename":"Show","sys":{"id":"62CSug7PQruJSFf61WQfgd"},"title":"Financial Rebellion with Catherine Austin Fitts","slug":"financial-rebellion-with-catherine-austin-fitts","broadcastChannelId":"a1xbld7jizbgusezhp5g","description":"The revolution will not be centralized. Financial Rebellion is a weekly show featuring Catherine Austin Fitts and attorney Carolyn Betts, general counsel of Solari, Inc. who call upon their years of experience in the financial and legal sectors to provide us with the tools we need to powershift our money and reclaim financial independence from the monopolizing grip of the central banks and digital currency titans.","hosts":"$56:props:videoData:show:hosts"}]}}},"donateLink":{"__typename":"ExternalLink","_id":"MTpjcm46Y29udGVudGZ1bDo6OmNvbnRlbnQ6c3BhY2VzL2ZweDc1b2hicWpxMy9lbnZpcm9ubWVudHMvbWFzdGVyL2VudHJpZXMvNGNLMFUxcnJNamFmWldPY2czbUEzQQ==","title":"Million Match","url":"https://childrenshealthdefense.org/fallmatch","queryId":"chdtv-donate-link"},"insiderSignUpLink":{"__typename":"ExternalLink","_id":"MTpjcm46Y29udGVudGZ1bDo6OmNvbnRlbnQ6c3BhY2VzL2ZweDc1b2hicWpxMy9lbnZpcm9ubWVudHMvbWFzdGVyL2VudHJpZXMvMWczdFpuTG9ubmQ1V21Gd3RQM3ZxMQ==","title":"Become an Insider","url":"https://childrenshealthdefense.org/become-a-chd-insider/","queryId":"become-insider"},"signUpLink":{"__typename":"ExternalLink","_id":"MTpjcm46Y29udGVudGZ1bDo6OmNvbnRlbnQ6c3BhY2VzL2ZweDc1b2hicWpxMy9lbnZpcm9ubWVudHMvbWFzdGVyL2VudHJpZXMvN3ZyWFhKbHZRQ3FBcEloY0JwV3F1dw==","title":"Join Our List","url":"https://childrenshealthdefense.org/about-us/sign-up/?itm_term=chd-tv","queryId":"chdtv-join-us-link"},"relatedVideos":[{"id":"5gEQnhsbs3zmUXfk8LeTch","contentful_id":"5gEQnhsbs3zmUXfk8LeTch","videoTitle":"Beware Smart Cities and 15-Minute Cities","videoDescription":"Smart cities are no conspiracy theory. Under the guise of “environmentalism” and “equity,” a surveillance initiative has been long underway in urban centers across the globe. Using digital technology to collect data, these communities become hubs for implementing and enforcing “eco-friendly” infrastructure and lifestyles. What privacy concerns arise from these programs? Are citizens responding positively to these changes? Can we expect to see more of these projects popping up in the coming years? Watch this episode of “Financial Rebellion” to find out!","publishedDate":"2/13/2025","publishedDateTimestamp":1739458740,"videoYear":"2025","videoSlug":"/chd-tv/shows/financial-rebellion-with-catherine-austin-fitts/beware-smart-cities-and-15-minutes-cities-show-page/","videoThumbnailUrl":"https://images.childrenshealthdefense.org/fpx75ohbqjq3/1pkfrkXQuG6YjSH1nZCj5b/a57be38aaa31c8175db1ddad189723b1/021325_FR_Cover_rev2.jpg","contentType":"Episode","tags":["Catherine Austin Fitts","Carolyn Betts","smart cities","15-minute cities","Sandi Adams","Glastonbury Town Council","Global Covenant of Mayors","Coalition for High Ambition Multilevel Partnerships - CHAMP","COP28","Global Covenant of Mayors for Climate & Energy - GCoM","cities","infrastructure","surveillance state","privacy","facial recognition","tracking","SDG - sustainable development goals","sustainability","UN - United Nations","climate change","C40 cities","clean energy","greenhouse gas","urban living","eco-friendly","mayors","local government","Paris Agreement","equity","resilience","Pact For the Future","Summit of the Future","ESG - Environmental Social Governance","stakeholders","net zero","carbon neutral","human capital management - HCM","greenwashing","environmentalism","corporate social responsibility - CSR","electric energy","fossil fuels","emissions","traffic","transportation","utility","water supply","waste management","communities","data collection","digital surveillance","digital ID","personal data","algocracy","AI - artificial intelligence","crime","police state","information and communication technology - ICT","technology","Garden Cities","utopia","Smarter Planet","Cisco Systems","urban development","philanthropy","connectivity","analytics","energy-efficient","bike-sharing","autonomous vehicles","public transit","e-governance","telecommunications - telecom","smart meters","DER - distributed energy resources","electric grid","smart grid","Energy Data Management Systems - EDMS","Big Data","Big Tech","Closed-circuit television - CCTV","City Brain","Safe City","ASEAN Smart Cities Network - ASCN","New Urban Agenda","carfree city","eco-city","mass surveillance","plannet community","smart highway","technocracy"],"showTitle":"Financial Rebellion with Catherine Austin Fitts","objectID":"5gEQnhsbs3zmUXfk8LeTch_financial-rebellion-with-catherine-austin-fitts","_highlightResult":{"videoTitle":{"value":"Beware Smart Cities and 15-Minute Cities","matchLevel":"none","matchedWords":[]},"videoDescription":{"value":"Smart cities are no conspiracy theory. Under the guise of “environmentalism” and “equity,” a surveillance initiative has been long underway in urban centers across the globe. Using digital technology to collect data, these communities become hubs for implementing and enforcing “eco-friendly” infrastructure and lifestyles. What privacy concerns arise from these programs? Are citizens responding positively to these changes? Can we expect to see more of these projects popping up in the coming years? Watch this episode of “Financial Rebellion” to find out!","matchLevel":"partial","fullyHighlighted":false,"matchedWords":["data","privacy","surveillance","data","centers","data"]},"tags":[{"value":"Catherine Austin Fitts","matchLevel":"partial","fullyHighlighted":true,"matchedWords":["catherine","austin","fitts"]},{"value":"Carolyn Betts","matchLevel":"partial","fullyHighlighted":true,"matchedWords":["carolyn","betts"]},{"value":"smart cities","matchLevel":"none","matchedWords":[]},{"value":"15-minute cities","matchLevel":"none","matchedWords":[]},{"value":"Sandi Adams","matchLevel":"none","matchedWords":[]},{"value":"Glastonbury Town Council","matchLevel":"none","matchedWords":[]},{"value":"Global Covenant of Mayors","matchLevel":"none","matchedWords":[]},{"value":"Coalition for High Ambition Multilevel Partnerships - CHAMP","matchLevel":"none","matchedWords":[]},{"value":"COP28","matchLevel":"none","matchedWords":[]},{"value":"Global Covenant of Mayors for Climate & Energy - GCoM","matchLevel":"none","matchedWords":[]},{"value":"cities","matchLevel":"none","matchedWords":[]},{"value":"infrastructure","matchLevel":"none","matchedWords":[]},{"value":"surveillance state","matchLevel":"partial","fullyHighlighted":false,"matchedWords":["surveillance"]},{"value":"privacy","matchLevel":"partial","fullyHighlighted":true,"matchedWords":["privacy"]},{"value":"facial recognition","matchLevel":"none","matchedWords":[]},{"value":"tracking","matchLevel":"partial","fullyHighlighted":true,"matchedWords":["tracking"]},{"value":"SDG - sustainable development goals","matchLevel":"none","matchedWords":[]},{"value":"sustainability","matchLevel":"none","matchedWords":[]},{"value":"UN - United Nations","matchLevel":"none","matchedWords":[]},{"value":"climate change","matchLevel":"none","matchedWords":[]},{"value":"C40 cities","matchLevel":"none","matchedWords":[]},{"value":"clean energy","matchLevel":"none","matchedWords":[]},{"value":"greenhouse gas","matchLevel":"none","matchedWords":[]},{"value":"urban living","matchLevel":"none","matchedWords":[]},{"value":"eco-friendly","matchLevel":"none","matchedWords":[]},{"value":"mayors","matchLevel":"none","matchedWords":[]},{"value":"local government","matchLevel":"none","matchedWords":[]},{"value":"Paris Agreement","matchLevel":"none","matchedWords":[]},{"value":"equity","matchLevel":"none","matchedWords":[]},{"value":"resilience","matchLevel":"none","matchedWords":[]},{"value":"Pact For the Future","matchLevel":"none","matchedWords":[]},{"value":"Summit of the Future","matchLevel":"none","matchedWords":[]},{"value":"ESG - Environmental Social Governance","matchLevel":"none","matchedWords":[]},{"value":"stakeholders","matchLevel":"none","matchedWords":[]},{"value":"net zero","matchLevel":"none","matchedWords":[]},{"value":"carbon neutral","matchLevel":"none","matchedWords":[]},{"value":"human capital management - HCM","matchLevel":"none","matchedWords":[]},{"value":"greenwashing","matchLevel":"none","matchedWords":[]},{"value":"environmentalism","matchLevel":"none","matchedWords":[]},{"value":"corporate social responsibility - CSR","matchLevel":"none","matchedWords":[]},{"value":"electric energy","matchLevel":"none","matchedWords":[]},{"value":"fossil fuels","matchLevel":"none","matchedWords":[]},{"value":"emissions","matchLevel":"none","matchedWords":[]},{"value":"traffic","matchLevel":"none","matchedWords":[]},{"value":"transportation","matchLevel":"none","matchedWords":[]},{"value":"utility","matchLevel":"none","matchedWords":[]},{"value":"water supply","matchLevel":"none","matchedWords":[]},{"value":"waste management","matchLevel":"none","matchedWords":[]},{"value":"communities","matchLevel":"none","matchedWords":[]},{"value":"data collection","matchLevel":"partial","fullyHighlighted":true,"matchedWords":["data","collection","data","data"]},{"value":"digital surveillance","matchLevel":"partial","fullyHighlighted":false,"matchedWords":["surveillance"]},{"value":"digital ID","matchLevel":"none","matchedWords":[]},{"value":"personal data","matchLevel":"partial","fullyHighlighted":true,"matchedWords":["data","personal","data","data"]},{"value":"algocracy","matchLevel":"none","matchedWords":[]},{"value":"AI - artificial intelligence","matchLevel":"partial","fullyHighlighted":false,"matchedWords":["artificial","intelligence"]},{"value":"crime","matchLevel":"none","matchedWords":[]},{"value":"police state","matchLevel":"none","matchedWords":[]},{"value":"information and communication technology - ICT","matchLevel":"partial","fullyHighlighted":false,"matchedWords":["information"]},{"value":"technology","matchLevel":"none","matchedWords":[]},{"value":"Garden Cities","matchLevel":"none","matchedWords":[]},{"value":"utopia","matchLevel":"none","matchedWords":[]},{"value":"Smarter Planet","matchLevel":"none","matchedWords":[]},{"value":"Cisco Systems","matchLevel":"none","matchedWords":[]},{"value":"urban development","matchLevel":"none","matchedWords":[]},{"value":"philanthropy","matchLevel":"none","matchedWords":[]},{"value":"connectivity","matchLevel":"none","matchedWords":[]},{"value":"analytics","matchLevel":"none","matchedWords":[]},{"value":"energy-efficient","matchLevel":"none","matchedWords":[]},{"value":"bike-sharing","matchLevel":"none","matchedWords":[]},{"value":"autonomous vehicles","matchLevel":"none","matchedWords":[]},{"value":"public transit","matchLevel":"none","matchedWords":[]},{"value":"e-governance","matchLevel":"none","matchedWords":[]},{"value":"telecommunications - telecom","matchLevel":"none","matchedWords":[]},{"value":"smart meters","matchLevel":"none","matchedWords":[]},{"value":"DER - distributed energy resources","matchLevel":"none","matchedWords":[]},{"value":"electric grid","matchLevel":"none","matchedWords":[]},{"value":"smart grid","matchLevel":"none","matchedWords":[]},{"value":"Energy Data Management Systems - EDMS","matchLevel":"partial","fullyHighlighted":false,"matchedWords":["data","data","data"]},{"value":"Big Data","matchLevel":"partial","fullyHighlighted":false,"matchedWords":["data","data","data"]},{"value":"Big Tech","matchLevel":"none","matchedWords":[]},{"value":"Closed-circuit television - CCTV","matchLevel":"none","matchedWords":[]},{"value":"City Brain","matchLevel":"none","matchedWords":[]},{"value":"Safe City","matchLevel":"none","matchedWords":[]},{"value":"ASEAN Smart Cities Network - ASCN","matchLevel":"none","matchedWords":[]},{"value":"New Urban Agenda","matchLevel":"none","matchedWords":[]},{"value":"carfree city","matchLevel":"none","matchedWords":[]},{"value":"eco-city","matchLevel":"none","matchedWords":[]},{"value":"mass surveillance","matchLevel":"partial","fullyHighlighted":false,"matchedWords":["surveillance"]},{"value":"plannet community","matchLevel":"none","matchedWords":[]},{"value":"smart highway","matchLevel":"none","matchedWords":[]},{"value":"technocracy","matchLevel":"none","matchedWords":[]}],"showTitle":{"value":"Financial Rebellion with Catherine Austin Fitts","matchLevel":"partial","fullyHighlighted":false,"matchedWords":["catherine","austin","fitts"]}}},{"id":"5vlcIG5I7zwF8KxxU2XF5O","contentful_id":"5vlcIG5I7zwF8KxxU2XF5O","videoTitle":"What Does Surveillance Have To Do With Sovereignty?","videoDescription":"You might be surprised as to how many forms of personal data tracking have been, will be and are currently rolled out on the public. It’s almost as if the dystopian future books and movies warned us about is already here. From doorbells to drones, location tracking, internet privacy and beyond — there really is no escaping the surveillance state. Tune in to “Financial Rebellion” to learn all about it! \n","publishedDate":"2/19/2026","publishedDateTimestamp":1771512600,"videoYear":"2026","videoSlug":"/chd-tv/shows/financial-rebellion-with-catherine-austin-fitts/what-does-surveillance-have-to-do-with-sovereignty-show-page/","videoThumbnailUrl":"https://images.childrenshealthdefense.org/fpx75ohbqjq3/3Z4WTZV3hM4DSQR0rDLYeY/1f6d5f1504afa0758d0afaab334d8a94/021926_FR_Cover.jpg","contentType":"Episode","tags":["Reinette Senum","Catherine Austin Fitts","Twila Brase","Carolyn Betts","open-air prison","imprisonment","sovereignty","surveillance","privacy","control","tracking","Turkey","travel prices","inflation","slaughterbots","AI - artificial intelligence","drones","Pete Hegseth","Secretary of War","Department of War - DOW","DOD - department of defense","DDP - drone dominance program","military","airforce","Gauntlet","Fort Benning","combat drones","Palantir","data sovereignty","personal data","genocide","tech companies","Apple","pre-speech","first amendment - 1st amendment","freedom of speech","Ring","Flock","Super Bowl ads","data centers","doorbells","cameras","XTEND","security threats","border","CBP","license plate readers","censorship","ICE","immigration","biometrics","location data","LPR","DEA","fourth amendment - 4th amendment"],"showTitle":"Financial Rebellion with Catherine Austin Fitts","objectID":"5vlcIG5I7zwF8KxxU2XF5O_financial-rebellion-with-catherine-austin-fitts","_highlightResult":{"videoTitle":{"value":"What Does Surveillance Have To Do With Sovereignty?","matchLevel":"partial","fullyHighlighted":false,"matchedWords":["surveillance"]},"videoDescription":{"value":"You might be surprised as to how many forms of personal data tracking have been, will be and are currently rolled out on the public. It’s almost as if the dystopian future books and movies warned us about is already here. From doorbells to drones, location tracking, internet privacy and beyond — there really is no escaping the surveillance state. Tune in to “Financial Rebellion” to learn all about it! \n","matchLevel":"partial","fullyHighlighted":false,"matchedWords":["data","privacy","surveillance","tracking","personal","data","data"]},"tags":[{"value":"Reinette Senum","matchLevel":"none","matchedWords":[]},{"value":"Catherine Austin Fitts","matchLevel":"partial","fullyHighlighted":true,"matchedWords":["catherine","austin","fitts"]},{"value":"Twila Brase","matchLevel":"none","matchedWords":[]},{"value":"Carolyn Betts","matchLevel":"partial","fullyHighlighted":true,"matchedWords":["carolyn","betts"]},{"value":"open-air prison","matchLevel":"partial","fullyHighlighted":false,"matchedWords":["open"]},{"value":"imprisonment","matchLevel":"none","matchedWords":[]},{"value":"sovereignty","matchLevel":"none","matchedWords":[]},{"value":"surveillance","matchLevel":"partial","fullyHighlighted":true,"matchedWords":["surveillance"]},{"value":"privacy","matchLevel":"partial","fullyHighlighted":true,"matchedWords":["privacy"]},{"value":"control","matchLevel":"none","matchedWords":[]},{"value":"tracking","matchLevel":"partial","fullyHighlighted":true,"matchedWords":["tracking"]},{"value":"Turkey","matchLevel":"none","matchedWords":[]},{"value":"travel prices","matchLevel":"none","matchedWords":[]},{"value":"inflation","matchLevel":"none","matchedWords":[]},{"value":"slaughterbots","matchLevel":"none","matchedWords":[]},{"value":"AI - artificial intelligence","matchLevel":"partial","fullyHighlighted":false,"matchedWords":["artificial","intelligence"]},{"value":"drones","matchLevel":"none","matchedWords":[]},{"value":"Pete Hegseth","matchLevel":"none","matchedWords":[]},{"value":"Secretary of War","matchLevel":"none","matchedWords":[]},{"value":"Department of War - DOW","matchLevel":"none","matchedWords":[]},{"value":"DOD - department of defense","matchLevel":"none","matchedWords":[]},{"value":"DDP - drone dominance program","matchLevel":"none","matchedWords":[]},{"value":"military","matchLevel":"none","matchedWords":[]},{"value":"airforce","matchLevel":"none","matchedWords":[]},{"value":"Gauntlet","matchLevel":"none","matchedWords":[]},{"value":"Fort Benning","matchLevel":"none","matchedWords":[]},{"value":"combat drones","matchLevel":"none","matchedWords":[]},{"value":"Palantir","matchLevel":"none","matchedWords":[]},{"value":"data sovereignty","matchLevel":"partial","fullyHighlighted":false,"matchedWords":["data","data","data"]},{"value":"personal data","matchLevel":"partial","fullyHighlighted":true,"matchedWords":["data","personal","data","data"]},{"value":"genocide","matchLevel":"none","matchedWords":[]},{"value":"tech companies","matchLevel":"none","matchedWords":[]},{"value":"Apple","matchLevel":"none","matchedWords":[]},{"value":"pre-speech","matchLevel":"none","matchedWords":[]},{"value":"first amendment - 1st amendment","matchLevel":"none","matchedWords":[]},{"value":"freedom of speech","matchLevel":"none","matchedWords":[]},{"value":"Ring","matchLevel":"none","matchedWords":[]},{"value":"Flock","matchLevel":"none","matchedWords":[]},{"value":"Super Bowl ads","matchLevel":"none","matchedWords":[]},{"value":"data centers","matchLevel":"partial","fullyHighlighted":true,"matchedWords":["data","data","centers","data"]},{"value":"doorbells","matchLevel":"none","matchedWords":[]},{"value":"cameras","matchLevel":"none","matchedWords":[]},{"value":"XTEND","matchLevel":"none","matchedWords":[]},{"value":"security threats","matchLevel":"none","matchedWords":[]},{"value":"border","matchLevel":"none","matchedWords":[]},{"value":"CBP","matchLevel":"none","matchedWords":[]},{"value":"license plate readers","matchLevel":"none","matchedWords":[]},{"value":"censorship","matchLevel":"none","matchedWords":[]},{"value":"ICE","matchLevel":"none","matchedWords":[]},{"value":"immigration","matchLevel":"none","matchedWords":[]},{"value":"biometrics","matchLevel":"none","matchedWords":[]},{"value":"location data","matchLevel":"partial","fullyHighlighted":false,"matchedWords":["data","data","data"]},{"value":"LPR","matchLevel":"none","matchedWords":[]},{"value":"DEA","matchLevel":"none","matchedWords":[]},{"value":"fourth amendment - 4th amendment","matchLevel":"none","matchedWords":[]}],"showTitle":{"value":"Financial Rebellion with Catherine Austin Fitts","matchLevel":"partial","fullyHighlighted":false,"matchedWords":["catherine","austin","fitts"]}}}]}]