WEBVTT 00:00:02.239 --> 00:00:04.780 Welcome to the AI Chat Podcast. I'm your host, 00:00:04.860 --> 00:00:07.179 Jaden Schaefer. Every day I cover cutting -edge 00:00:07.179 --> 00:00:09.320 AI news and talk with the leaders behind it, 00:00:09.339 --> 00:00:11.300 breaking down what it means for your life and 00:00:11.300 --> 00:00:17.820 business. Insiders say that Anthropic is eyeing 00:00:17.820 --> 00:00:21.480 a $2 trillion valuation for their IPO, the biggest 00:00:21.480 --> 00:00:25.440 ever recorded. Nvidia is going to backstop $500 00:00:25.440 --> 00:00:29.399 billion in AI data centers, which will be guaranteed 00:00:29.399 --> 00:00:33.520 by GPU resale values. Shield font is going to. 00:00:33.920 --> 00:00:37.479 poison AI scrapers by swapping words inside of 00:00:37.479 --> 00:00:39.640 the text that humans can't see but the AI models 00:00:39.640 --> 00:00:43.759 will. Lovable has raised $400 million at a $13 00:00:43.759 --> 00:00:46.380 .3 billion valuation. Their annual recurring 00:00:46.380 --> 00:00:49.770 revenue has tripled. Thrive Holdings has raised 00:00:49.770 --> 00:00:53.170 $2 billion at a $12 billion valuation. They're 00:00:53.170 --> 00:00:55.009 going to be essentially buying companies and 00:00:55.009 --> 00:00:58.210 injecting open AI inside of them. AI models are 00:00:58.210 --> 00:01:02.070 flagging fatty liver disease early as conditions 00:01:02.070 --> 00:01:05.969 are hitting about 30 % of adults worldwide. If 00:01:05.969 --> 00:01:07.510 you've ever been working with Claude and needed 00:01:07.510 --> 00:01:10.030 to generate images, audio, or video and been 00:01:10.030 --> 00:01:11.890 frustrated that it can't, I'd love for you to 00:01:11.890 --> 00:01:14.569 check out the AI Box MCP. Basically, it's a connector 00:01:14.569 --> 00:01:17.310 that connects over 80 different AI models, all 00:01:17.310 --> 00:01:19.200 the top... AI models you're used to using, undoubtedly 00:01:19.200 --> 00:01:21.659 a bunch of them you already subscribe to, and 00:01:21.659 --> 00:01:23.620 pulls them all inside of Cloud. So when I'm using 00:01:23.620 --> 00:01:26.439 Cloud and I'm trying to generate new images for 00:01:26.439 --> 00:01:29.060 Facebook ads, for example, I just say use the 00:01:29.060 --> 00:01:31.819 AI box MCP to generate an image of, and I just 00:01:31.819 --> 00:01:34.019 put in the normal image prompt I would use anywhere. 00:01:34.140 --> 00:01:36.159 You can even create scheduled tasks inside of 00:01:36.159 --> 00:01:39.319 Cloud that draw on the abilities of the AI box 00:01:39.319 --> 00:01:41.980 MCP, and it can automatically pull an image, 00:01:42.079 --> 00:01:44.620 audio, and video. It feels magical if I'm being 00:01:44.620 --> 00:01:46.599 honest, because in Cloud, which I'm not used 00:01:46.599 --> 00:01:48.799 to... seen images generated, I'm seeing images 00:01:48.799 --> 00:01:51.000 that are pulled straight into the chat or the 00:01:51.000 --> 00:01:53.180 side panel, and you can visually see them all 00:01:53.180 --> 00:01:55.040 there. So if you want to give it a try, it is 00:01:55.040 --> 00:01:58.000 linked in the description, the AI box MCP, it 00:01:58.000 --> 00:02:01.780 is AI box .ai slash MCP. Before we get into a 00:02:01.780 --> 00:02:03.200 couple of the stories, I wanted to talk about 00:02:03.200 --> 00:02:05.180 a few other headlines today. There's just so 00:02:05.180 --> 00:02:07.260 much news that I didn't want to build, you know, 00:02:07.260 --> 00:02:09.219 an hour long podcast episode. But I wanted to 00:02:09.219 --> 00:02:11.479 give you guys a couple updates. One is that Uber 00:02:11.479 --> 00:02:13.639 and Pony AI, they're going to deploy about 2000 00:02:13.639 --> 00:02:16.680 robo taxis inside of Europe, just really wild. 00:02:16.879 --> 00:02:19.240 Google is going to make the visible watermarks 00:02:19.240 --> 00:02:21.500 on their AI images, video and audio optional. 00:02:21.620 --> 00:02:24.219 Thank freaking goodness. I have been complaining. 00:02:24.280 --> 00:02:26.360 And in fact, I had Logan Kilpatrick on a podcast 00:02:26.360 --> 00:02:28.360 a few months ago. And I asked him about that. 00:02:28.400 --> 00:02:30.139 I'm like, man, when is Google going to get rid 00:02:30.139 --> 00:02:32.379 of if you've ever used like nano banana, there 00:02:32.379 --> 00:02:35.300 is a watermark on the images inside of nano banana. 00:02:35.400 --> 00:02:37.340 It's super annoying because you can't really 00:02:37.340 --> 00:02:38.699 use them for anything without trying to crop 00:02:38.699 --> 00:02:41.240 them out or using another AI tool to go and remove 00:02:41.240 --> 00:02:44.280 that like chat GPT. So it just was it felt very 00:02:44.280 --> 00:02:46.969 not super. usable without cropping stuff. So 00:02:46.969 --> 00:02:48.889 happy that that's going to be a possibility. 00:02:49.509 --> 00:02:52.889 Databricks has just raised $5 billion at $190 00:02:52.889 --> 00:02:56.069 billion valuation. And finally, Cognition is 00:02:56.069 --> 00:02:59.469 in talks to raise $40 billion valuation just 00:02:59.469 --> 00:03:02.469 months after their $26 billion valuation round. 00:03:02.669 --> 00:03:04.349 Okay, let's talk about what's going on with Thrive 00:03:04.349 --> 00:03:06.530 Holdings because this is a really interesting 00:03:06.530 --> 00:03:08.569 company. And I think the strategy is something 00:03:08.569 --> 00:03:12.090 we might see with more PE moving forward in the 00:03:12.090 --> 00:03:13.870 future. So Thrive Holdings is a company that 00:03:13.870 --> 00:03:16.650 by traditional service businesses, and then it 00:03:16.650 --> 00:03:18.750 rebuilds them around AI. They just raised $2 00:03:18.750 --> 00:03:21.889 billion at a $12 billion valuation. They raised 00:03:21.889 --> 00:03:23.750 the money from SoftBank and a bunch of other 00:03:23.750 --> 00:03:27.069 investors, but they now own over 70 different 00:03:27.069 --> 00:03:29.370 accounting and IT companies, and they're opening 00:03:29.370 --> 00:03:31.810 a third division, which is focused on infrastructure 00:03:31.810 --> 00:03:35.449 permitting and compliance work. Current tax AI 00:03:35.449 --> 00:03:39.370 processed over 7 ,000 tax returns, and it did 00:03:39.370 --> 00:03:42.870 it with a 98 % accuracy, and it cut tax prep 00:03:42.870 --> 00:03:45.599 time by more than 30 % at the different firms 00:03:45.599 --> 00:03:48.879 that were using it. Shield AI's tools sped up 00:03:48.879 --> 00:03:53.240 IT help desk resolutions by 36x and it doubled 00:03:53.240 --> 00:03:56.379 custom AI agent deployments in the last month. 00:03:56.439 --> 00:03:59.060 And OpenAI took an ownership stake in December 00:03:59.060 --> 00:04:02.240 of 2025, and they embed staff directly into portfolio 00:04:02.240 --> 00:04:04.400 companies to help accelerate the AI rollout. 00:04:04.500 --> 00:04:06.379 So I think this is interesting because they're, 00:04:06.379 --> 00:04:07.840 you know, they're working with IT companies, 00:04:07.919 --> 00:04:10.020 they're working with tax companies. I mean, something 00:04:10.020 --> 00:04:12.259 like a 98 % accuracy is obviously impressive. 00:04:12.300 --> 00:04:14.740 So you can see. There's a lot of value there. 00:04:14.860 --> 00:04:16.860 But it's also interesting to me that OpenAI took 00:04:16.860 --> 00:04:20.139 a literal ownership stake in this and they're 00:04:20.139 --> 00:04:22.300 embedding their own staff to try to make this 00:04:22.300 --> 00:04:25.139 work. So it's beyond just OpenAI saying, hey, 00:04:25.160 --> 00:04:26.699 look, we're gonna make AI that works great for 00:04:26.699 --> 00:04:29.500 AI or for tax companies. By taking an ownership 00:04:29.500 --> 00:04:31.959 stake, like they're becoming a tax company, which 00:04:31.959 --> 00:04:35.060 is really fascinating to me. Thrive in particular, 00:04:35.180 --> 00:04:37.379 this company is betting that the real bottleneck 00:04:37.379 --> 00:04:39.860 for AI in business isn't just the technology, 00:04:40.139 --> 00:04:42.120 it's actually getting it implemented. And we've 00:04:42.120 --> 00:04:44.019 - seen a bunch of these AI implementation companies 00:04:44.019 --> 00:04:45.480 isn't the only one that opening eyes working 00:04:45.480 --> 00:04:47.660 with and anthropics also working with some it's 00:04:47.660 --> 00:04:49.740 usually they'll kind of go with some big investment 00:04:49.740 --> 00:04:53.000 firms and create these but by owning entire service 00:04:53.000 --> 00:04:55.199 firms rather than just selling the software to 00:04:55.199 --> 00:04:58.120 them they're capturing the full benefit of the 00:04:58.120 --> 00:05:00.759 productivity gains that AI can roll out and I 00:05:00.759 --> 00:05:03.579 mean this isn't anything super revolutionary 00:05:03.579 --> 00:05:05.560 if you think about it private equity firms oftentimes 00:05:05.560 --> 00:05:08.240 will go buy a company or a bunch of companies 00:05:08.240 --> 00:05:11.790 in the same space and then they'll consolidate 00:05:11.790 --> 00:05:13.589 some of the back office, they'll they'll be able 00:05:13.589 --> 00:05:15.310 to cut a bunch of staff because of this, and 00:05:15.310 --> 00:05:17.250 they'll make the companies more profitable because 00:05:17.250 --> 00:05:20.029 of it. So this is just a new element to it where 00:05:20.029 --> 00:05:22.949 we're using AI to do that. But it's it's really 00:05:22.949 --> 00:05:25.310 the same playbook that they've been doing for 00:05:25.310 --> 00:05:27.899 a very long time. Lovable, which is based in 00:05:27.899 --> 00:05:30.660 Sweden, is a platform. If you haven't heard it 00:05:30.660 --> 00:05:33.139 already, it turns... I've talked about it a ton 00:05:33.139 --> 00:05:34.779 on the show, so I would assume you have. It's 00:05:34.779 --> 00:05:37.100 not a company I'm actively using right now. But 00:05:37.100 --> 00:05:39.199 basically, if you want to build a web app or 00:05:39.199 --> 00:05:41.699 some sort of website or tool, it's a really easy 00:05:41.699 --> 00:05:43.420 way to get started with that. Today, I would 00:05:43.420 --> 00:05:47.860 just say go use Cloud Cowork or ChatGPT Work 00:05:47.860 --> 00:05:50.360 because they can do a lot of the same stuff. 00:05:50.819 --> 00:05:54.560 But Lovable just raised $400 million at a $13 00:05:54.560 --> 00:05:58.800 .3 billion value. Their ARR has about tripled. 00:05:58.800 --> 00:06:02.160 So their user base is massive. Projects built 00:06:02.160 --> 00:06:04.519 on lovable right now jumped from about 25 million 00:06:04.519 --> 00:06:07.699 to 60 million in eight months and monthly visits 00:06:07.699 --> 00:06:11.740 to those apps. quadrupled to about 900 million. 00:06:11.819 --> 00:06:14.259 So not, not, you know, just visits exclusively 00:06:14.259 --> 00:06:16.199 to lovable, but the apps that are built on lovable 00:06:16.199 --> 00:06:19.079 are generating about 900 million visits, which 00:06:19.079 --> 00:06:21.860 is absolutely wild. So their annual revenue, 00:06:21.939 --> 00:06:23.740 like I mentioned, tripled, and that's just over 00:06:23.740 --> 00:06:25.459 eight months. So I mean, this is a company that 00:06:25.459 --> 00:06:28.160 is growing incredibly fast. lovable didn't say 00:06:28.160 --> 00:06:30.839 exactly what their annual revenue number was, 00:06:30.959 --> 00:06:32.699 but they're like, look, it tripled. So I'm not 00:06:32.699 --> 00:06:34.180 going to put as much stock in that until I know 00:06:34.180 --> 00:06:37.199 the actual number. Menlo ventures led this particular 00:06:37.199 --> 00:06:40.199 round. They also led the previous one so i think 00:06:40.199 --> 00:06:42.480 that they're very bullish on the company on the 00:06:42.480 --> 00:06:44.160 metrics and they think that it's got some strong 00:06:44.160 --> 00:06:46.060 growth numbers even if they're not telling them 00:06:46.060 --> 00:06:49.980 all to us Lovable in particular now handles security 00:06:49.980 --> 00:06:53.019 checks. They prompt users towards making safer 00:06:53.019 --> 00:06:55.040 decisions. So if you're non -technical, which 00:06:55.040 --> 00:06:57.439 is basically everyone using Lovable, you can 00:06:57.439 --> 00:06:59.319 ship stuff really fast and you're not really 00:06:59.319 --> 00:07:01.180 worried about the security angle. And when I 00:07:01.180 --> 00:07:03.139 was using Lovable, because honestly, Lovable 00:07:03.139 --> 00:07:05.980 to me was a gateway drug into vibe coding. I'm 00:07:05.980 --> 00:07:07.800 not, I wouldn't have considered myself a technical 00:07:07.800 --> 00:07:10.800 founder. And Lovable is how I got started building 00:07:10.800 --> 00:07:14.100 apps by myself and kind of coding stuff. It does 00:07:14.100 --> 00:07:16.410 like your backend. It does like your login. with 00:07:16.410 --> 00:07:18.850 Google, it does, you know, links to your Stripe 00:07:18.850 --> 00:07:20.949 payment. So it does a lot of this stuff for you. 00:07:20.990 --> 00:07:23.230 And it makes it pretty simple. I'll say if you're 00:07:23.230 --> 00:07:25.550 willing to invest like one or two days, you can 00:07:25.550 --> 00:07:27.769 get off of lovable completely and just build 00:07:27.769 --> 00:07:30.829 it from scratch with with something like Claude 00:07:30.829 --> 00:07:34.889 coworker, Claude, or chat GPT work. And using 00:07:34.889 --> 00:07:37.029 that all you need to do is set up a Vercel account, 00:07:37.350 --> 00:07:40.750 a GitHub account, and you need to make a super 00:07:40.750 --> 00:07:43.050 base account. So if you make accounts on those 00:07:43.050 --> 00:07:46.079 three websites, you can get chat GPT work or 00:07:46.079 --> 00:07:51.259 Claude co work to go get an API or an MCP link 00:07:51.259 --> 00:07:55.019 them into into their tools, and they can just 00:07:55.019 --> 00:07:57.040 build everything and work on everything for you. 00:07:57.079 --> 00:07:59.160 So if you if you want to spend a couple days, 00:07:59.379 --> 00:08:00.980 that's how I would do it. Otherwise, you can 00:08:00.980 --> 00:08:02.579 go to level my only downside with lovable if 00:08:02.579 --> 00:08:04.259 I'm being honest, which obviously they're they're 00:08:04.259 --> 00:08:07.019 making an insane amount of money is it was pretty 00:08:07.019 --> 00:08:10.399 expensive. And so I was just doing so much. And 00:08:10.399 --> 00:08:14.040 what you got to appreciate is that Claude and 00:08:14.120 --> 00:08:17.860 and ChatGPT both have like $20, $100, and $200 00:08:17.860 --> 00:08:20.620 tiers. And all of those tiers, the tokens are 00:08:20.620 --> 00:08:22.480 subsidized, especially if you're doing like a 00:08:22.480 --> 00:08:25.980 $200 a month tier. You can get a ton of building, 00:08:26.060 --> 00:08:28.540 a ton of coding done, and you get like thousands 00:08:28.540 --> 00:08:31.100 of dollars worth of coding for $200. If you're 00:08:31.100 --> 00:08:34.200 using lovable, it's all API based. And so you're 00:08:34.200 --> 00:08:35.779 going to have to pay the thousands of dollars, 00:08:35.820 --> 00:08:37.379 which I was when I was using lovable. And I was, 00:08:37.460 --> 00:08:39.559 you know, once, once I think I spent $2 ,000 00:08:39.559 --> 00:08:41.919 in a month on credits, I was like, okay, I need 00:08:41.919 --> 00:08:43.840 to figure out a cheaper way to do this. And that's 00:08:43.840 --> 00:08:46.340 when I moved to Claude. But you can build a really 00:08:46.340 --> 00:08:49.129 impressive and really cool stuff there. NVIDIA 00:08:49.129 --> 00:08:52.590 is backing a $500 billion data center financing 00:08:52.590 --> 00:08:55.830 deal with Apollo, BlackRock, Blackstone, Brookfield, 00:08:55.990 --> 00:08:58.690 Goldman Sachs, and KKR. And they are guaranteed 00:08:58.690 --> 00:09:03.590 to cover 25 % of any losses if the GPUs use, 00:09:03.730 --> 00:09:07.149 and they're using the GPUs as collateral for 00:09:07.149 --> 00:09:10.409 any sort of loss in value. So that basically 00:09:10.409 --> 00:09:12.110 is letting those six different companies fund 00:09:12.110 --> 00:09:15.110 massive AI infrastructure build outs. And NVIDIA 00:09:15.110 --> 00:09:17.769 is going to gain a channel to sell billions. 00:09:18.000 --> 00:09:19.919 in GPU sales. They're not going to use their 00:09:19.919 --> 00:09:21.799 own cash for it. They're just kind of like a 00:09:21.799 --> 00:09:24.879 backstop on it. But they're going to make a ton 00:09:24.879 --> 00:09:27.500 of money from this. This is really interesting 00:09:27.500 --> 00:09:31.500 because NVIDIA has committed about $750 billion 00:09:31.500 --> 00:09:34.580 in these type of deals just this summer. They 00:09:34.580 --> 00:09:36.980 have deals with OpenAI, Anthropic, CoreWeave, 00:09:36.980 --> 00:09:38.700 and a bunch of other AI companies. Bloomberg 00:09:38.700 --> 00:09:41.870 was kind of reporting on this earlier. But the 00:09:41.870 --> 00:09:44.730 guarantee basically creates a quote unquote wrong 00:09:44.730 --> 00:09:48.669 way risk. So NVIDIA's obligation to pay grows 00:09:48.669 --> 00:09:52.289 biggest exactly when GPU demand might get weak 00:09:52.289 --> 00:09:55.070 and NVIDIA's own revenue would slow. So that's 00:09:55.070 --> 00:09:57.970 kind of a dangerous moment, right? They're like, 00:09:58.049 --> 00:09:59.870 hey, look, like if for some reason everything 00:09:59.870 --> 00:10:02.110 crashes, like we'll backstop you. But like if 00:10:02.110 --> 00:10:03.529 everything's crashing and no one's buying the 00:10:03.529 --> 00:10:06.289 GPUs, NVIDIA, like their ability to backstop 00:10:06.289 --> 00:10:09.700 them is going to be. Definitely strange. So it's 00:10:09.700 --> 00:10:12.620 a tricky moment. The CEO of Microsoft, Satya 00:10:12.620 --> 00:10:15.360 Nadella, recently recommended a book about railroad 00:10:15.360 --> 00:10:18.279 era financial engineering that ended in collapse. 00:10:18.759 --> 00:10:21.779 And so, you know, potentially there's other people 00:10:21.779 --> 00:10:23.399 that are kind of looking at this and saying, 00:10:23.539 --> 00:10:24.639 hey, maybe this is kind of like what happened 00:10:24.639 --> 00:10:27.159 to the railroad as far as infrastructure spending 00:10:27.159 --> 00:10:29.299 goes and saying that there are some dangerous 00:10:29.299 --> 00:10:32.139 things happening. But in any case, if NVIDIA's... 00:10:32.759 --> 00:10:35.919 residual value guarantee works, it could create 00:10:35.919 --> 00:10:39.200 a kind of a really liquid secondhand market for 00:10:39.200 --> 00:10:41.320 older GPU chips. And that would let startups 00:10:41.320 --> 00:10:44.740 and smaller companies buy some of the older generations 00:10:44.740 --> 00:10:46.600 of hardware, it would be a lot cheaper instead 00:10:46.600 --> 00:10:48.200 of just always having to buy the newest stuff. 00:10:48.379 --> 00:10:52.039 But that also assumes that GPU demand is going 00:10:52.039 --> 00:10:54.059 to be really strong, right. And so I think Nvidia 00:10:54.059 --> 00:10:57.029 is trying to like ensure it's strong by backstopping 00:10:57.029 --> 00:10:59.490 these big firms to go build out all the new stuff. 00:10:59.610 --> 00:11:01.529 And then the old stuff is still gonna be valuable 00:11:01.529 --> 00:11:03.629 to other people. But you know, NVIDIA's liability 00:11:03.629 --> 00:11:06.669 is going to be massive in in the case of anything 00:11:06.669 --> 00:11:09.269 going wrong. Okay, this is an interesting story. 00:11:09.450 --> 00:11:12.190 It's called shield font. And basically what it 00:11:12.190 --> 00:11:15.730 does is it poisons AI scrapers by swapping out 00:11:15.730 --> 00:11:19.289 different words. So in the HTML, when a scraper 00:11:19.289 --> 00:11:21.129 is going and scraping your website or your blog, 00:11:21.350 --> 00:11:24.899 it's going to show random words, mixed. in there 00:11:24.899 --> 00:11:27.460 into a human you want to be able to see it but 00:11:27.460 --> 00:11:30.460 in the html it's going to mess it all up and 00:11:30.460 --> 00:11:32.500 so human readers go to your site they get everything 00:11:32.500 --> 00:11:35.080 they want but if an ai model is trying to scrape 00:11:35.080 --> 00:11:37.299 it it's going to get really jumbled up text and 00:11:37.299 --> 00:11:39.940 it's going to mess up their llms for you know 00:11:39.940 --> 00:11:42.960 producing high quality content. So they actually 00:11:42.960 --> 00:11:45.559 tested it against six major AI scraping pipelines 00:11:45.559 --> 00:11:48.759 and over 90 % of pages that were using shield 00:11:48.759 --> 00:11:52.240 font were rejected by the scrapers quality filters. 00:11:52.279 --> 00:11:55.919 So basically it's going to be, um, it's a technical 00:11:55.919 --> 00:11:58.700 defense against AI models coming and trying to 00:11:58.700 --> 00:12:00.860 train on your, on your data without being authorized 00:12:00.860 --> 00:12:03.759 to do it. I love this concept if I'm being honest. 00:12:03.919 --> 00:12:06.120 And of course I'm very bullish on AI and I think 00:12:06.120 --> 00:12:08.360 like, you know, accelerate AI and I'm like very 00:12:08.360 --> 00:12:10.419 pro that. But at the same time, I think these 00:12:10.419 --> 00:12:13.090 AI models need to compensate people and people 00:12:13.090 --> 00:12:15.730 take measures to ensure that that happens i'm 00:12:15.730 --> 00:12:18.649 thrilled by it one thing that i've done is with 00:12:18.649 --> 00:12:21.190 even like this podcast i've licensed my podcast 00:12:21.190 --> 00:12:25.169 audio to different ai models that train um They 00:12:25.169 --> 00:12:27.789 train models on, you know, just how people talk. 00:12:27.929 --> 00:12:29.389 So it's general purpose. I'm not like cloning 00:12:29.389 --> 00:12:31.870 my voice or anything. And it's been an awesome 00:12:31.870 --> 00:12:33.629 revenue source for me. It's kind of random, but 00:12:33.629 --> 00:12:35.870 it actually makes a ton of money. And I've seen 00:12:35.870 --> 00:12:37.490 a lot of other people doing this with audio, 00:12:37.629 --> 00:12:40.830 with video, with music. And by the way, if you're 00:12:40.830 --> 00:12:42.590 interested in doing that and you have some sort 00:12:42.590 --> 00:12:47.149 of data set, go to my website, fiund .com. I'll 00:12:47.149 --> 00:12:49.490 leave a link in the description. It's fiund .com 00:12:49.490 --> 00:12:52.450 where I will give you, I'll let you know what 00:12:52.450 --> 00:12:54.659 your podcast or your YouTube. YouTube channel 00:12:54.659 --> 00:12:57.759 or whatever your data set might be worth, if 00:12:57.759 --> 00:13:00.159 you're interested in that. But yeah, so I love 00:13:00.159 --> 00:13:01.919 this. And I think this is a great way to help 00:13:01.919 --> 00:13:04.559 people get paid and kind of fight back against 00:13:04.559 --> 00:13:08.000 scraping that isn't authorized per se. I mean, 00:13:08.000 --> 00:13:09.620 if people want to opt into it, fantastic. If 00:13:09.620 --> 00:13:10.740 you want to opt out of it, I think there should 00:13:10.740 --> 00:13:12.539 be a way to opt out of it. And sometimes even 00:13:12.539 --> 00:13:14.299 when you have like a do not scrape, they still 00:13:14.299 --> 00:13:16.539 go scrape you. So I think this is this is the 00:13:16.539 --> 00:13:18.720 next best thing and they, you know, they get 00:13:18.720 --> 00:13:21.509 what they pay for. AI models that are trained 00:13:21.509 --> 00:13:24.750 on routine blood tests and chest x -rays can 00:13:24.750 --> 00:13:27.570 now spot fatty liver disease early. So this is 00:13:27.570 --> 00:13:30.629 a condition that affects about 30 % of adults 00:13:30.629 --> 00:13:33.830 worldwide. And if you can catch it early, it 00:13:33.830 --> 00:13:36.830 helps because liver damage is reversible if you 00:13:36.830 --> 00:13:38.509 have like treatment or some lifestyle changes. 00:13:38.649 --> 00:13:41.570 But yeah, three quarters of patients only get 00:13:41.570 --> 00:13:44.870 diagnosed when it is already life threatening. 00:13:45.009 --> 00:13:48.620 So there was an OSCA. Metropolitan University 00:13:48.620 --> 00:13:50.980 model that identified fatty liver diseases from 00:13:50.980 --> 00:13:54.139 chest x -rays. It had an 82 % accuracy, so it's 00:13:54.139 --> 00:13:58.139 able to spot spotting the liver fat in really 00:13:58.139 --> 00:14:01.460 small slices of organs that are visible on like 00:14:01.460 --> 00:14:05.019 routine scans. And Evito's LiverPro algorithm, 00:14:05.299 --> 00:14:09.059 which was built from age and nine blood biomarkers, 00:14:09.120 --> 00:14:13.960 outperformed the standard FIB4 test across 470 00:14:13.960 --> 00:14:16.519 ,000 middle -aged patients, and it's now being 00:14:16.519 --> 00:14:19.639 commercialized. by Roche. So I love that because 00:14:19.639 --> 00:14:22.580 we kind of had the standard test, and they came 00:14:22.580 --> 00:14:24.480 up with a new algorithm and it and it beats it. 00:14:24.539 --> 00:14:26.580 And so that's what's actually going to be commercialized 00:14:26.580 --> 00:14:28.940 and rolled out for everyone. The liver actually 00:14:28.940 --> 00:14:32.000 can regenerate and it can also reverse fibrosis 00:14:32.000 --> 00:14:34.179 if you catch it early. So if you know this kind 00:14:34.179 --> 00:14:37.299 of treatment is lifestyle stuff, I think we know 00:14:37.299 --> 00:14:38.779 the things that aren't good for our liver, but 00:14:38.779 --> 00:14:40.419 your doctor can definitely tell you more about 00:14:40.419 --> 00:14:43.519 what what's entitled there. But yeah, I think 00:14:43.519 --> 00:14:45.100 this is phenomenal. And this is something that 00:14:45.100 --> 00:14:48.100 is really exciting to me. I love it anytime that 00:14:48.100 --> 00:14:50.539 AI can help with healthcare can help cure diseases 00:14:50.539 --> 00:14:52.860 can help solve diseases. And so this is something 00:14:52.860 --> 00:14:54.700 that I'm really excited about. Guys, thanks so 00:14:54.700 --> 00:14:56.480 much for tuning into the podcast. If you enjoyed 00:14:56.480 --> 00:14:59.059 it, and you want to get access to over 80 different 00:14:59.059 --> 00:15:02.019 AI models inside of clot or anywhere else, go 00:15:02.019 --> 00:15:05.580 check out AI box .ai. And I hope to see you all 00:15:05.580 --> 00:15:06.399 on the next episode.