WEBVTT 00:00:00.000 --> 00:00:03.319 AI can now win a gold medal at the Math Olympiad. 00:00:03.540 --> 00:00:06.480 Two sec silence. But it still fails to read an 00:00:06.480 --> 00:00:10.080 analog clock 50 % of the time. Beat. I find this 00:00:10.080 --> 00:00:12.960 contrast absolutely fascinating. We have built 00:00:12.960 --> 00:00:16.039 a machine capable of solving Olympic -level theoretical 00:00:16.039 --> 00:00:20.079 mathematics. Yet, it cannot consistently tell 00:00:20.079 --> 00:00:22.429 you the time on a wall clock. It is a profound 00:00:22.429 --> 00:00:25.550 paradox. We are dealing with an entirely new 00:00:25.550 --> 00:00:27.929 form of intelligence here. I mean, it does not 00:00:27.929 --> 00:00:30.449 learn the way human beings learn. The architecture 00:00:30.449 --> 00:00:33.570 scales brilliantly for complex logic, but it 00:00:33.570 --> 00:00:36.270 completely fractures on basic physical intuition. 00:00:36.590 --> 00:00:39.210 Welcome to today's deep dive. We are exploring 00:00:39.210 --> 00:00:41.670 the rapidly shifting balance of power in artificial 00:00:41.670 --> 00:00:43.670 intelligence. We are going to look closely at 00:00:43.670 --> 00:00:46.310 the tools you use every single day. We will explore 00:00:46.310 --> 00:00:48.570 the massive structural changes happening in corporate 00:00:48.570 --> 00:00:50.829 software right now. And then we are scaling. 00:00:50.890 --> 00:00:53.049 all the way up to a global geopolitical dead 00:00:53.049 --> 00:00:54.810 heat. Yeah, we have a lot of crucial ground to 00:00:54.810 --> 00:00:56.810 cover today. We are looking at a massive new 00:00:56.810 --> 00:00:59.170 leak from Anthropic. This development basically 00:00:59.170 --> 00:01:02.070 threatens to disrupt an entire $6 .6 billion 00:01:02.070 --> 00:01:05.109 industry. We're also going to unpack Microsoft's 00:01:05.109 --> 00:01:07.870 sudden code red panic. We will review some highly 00:01:07.870 --> 00:01:10.810 capable new tools available to you. And finally, 00:01:10.909 --> 00:01:15.010 we will examine Stanford's massive 2026 AI index 00:01:15.010 --> 00:01:17.689 report. Let us start right at the software layer 00:01:17.689 --> 00:01:20.780 today. We are witnessing a fundamental shift 00:01:20.780 --> 00:01:24.079 in the AI ecosystem. The creators of the foundational 00:01:24.079 --> 00:01:27.019 models are starting to claw back power. They 00:01:27.019 --> 00:01:28.920 are taking it directly from the nimble startups 00:01:28.920 --> 00:01:30.900 that built on top of them. To really understand 00:01:30.900 --> 00:01:33.760 this, we need to look at Lovable. They were the 00:01:33.760 --> 00:01:35.760 biggest success story of the last 12 months. 00:01:35.939 --> 00:01:39.780 They hit a staggering $6 .6 billion valuation. 00:01:40.200 --> 00:01:42.799 Which is wild. Right. They achieved that by building 00:01:42.799 --> 00:01:45.739 a beautiful, intuitive layer on top of Claude's 00:01:45.739 --> 00:01:49.400 API. That specific layer enabled a massive trend 00:01:49.400 --> 00:01:51.799 we call vibe coding. Let me define that specific 00:01:51.799 --> 00:01:53.879 term for you right away. Vibe coding is coding 00:01:53.879 --> 00:01:56.159 by simply describing what you want in plain English. 00:01:56.439 --> 00:01:58.920 Exactly. You do not need to know Python or JavaScript 00:01:58.920 --> 00:02:01.060 anymore. You just explain your vision and the 00:02:01.060 --> 00:02:03.599 software builds the interface. It made building 00:02:03.599 --> 00:02:05.680 custom software accessible to almost anyone. 00:02:06.239 --> 00:02:08.960 But Anthropic just signaled a massive change 00:02:08.960 --> 00:02:10.819 in the weather. They're tired of letting their 00:02:10.819 --> 00:02:13.189 partners have all the fun. Yeah, we have seen 00:02:13.189 --> 00:02:15.669 leaked screenshots showing a massive shift in 00:02:15.669 --> 00:02:18.250 their strategy. Anthropic is building a native 00:02:18.250 --> 00:02:21.509 full -stack app builder directly inside Claude. 00:02:21.650 --> 00:02:23.590 This is no longer just a simple text generator. 00:02:23.849 --> 00:02:26.430 Not at all. Until now, Claude could just show 00:02:26.430 --> 00:02:28.389 you a nice little preview window. You would ask 00:02:28.389 --> 00:02:30.349 for a website, and you'd get a simple React component. 00:02:30.650 --> 00:02:33.490 It was visually pleasing, but functionally hollow. 00:02:33.789 --> 00:02:37.129 Right. Now, they're completely changing the underlying 00:02:37.129 --> 00:02:40.189 architecture. Quad handles the front -end user 00:02:40.189 --> 00:02:43.490 interface and the complex backend logic. It sets 00:02:43.490 --> 00:02:45.449 up your database connections and storage buckets 00:02:45.449 --> 00:02:48.069 in one seamless motion. The leak shows dedicated 00:02:48.069 --> 00:02:50.449 tabs for critical infrastructure right on the 00:02:50.449 --> 00:02:52.710 dashboard. You have Auth and Secrets Management 00:02:52.710 --> 00:02:56.289 built right in. Let's explain why that matters 00:02:56.289 --> 00:02:59.030 to you. Auth isn't just a basic login screen. 00:02:59.229 --> 00:03:02.009 No. It is the complex security required to manage 00:03:02.009 --> 00:03:05.330 user identities safely. Claude also handles security 00:03:05.330 --> 00:03:06.909 scanning for your applications automatically. 00:03:07.409 --> 00:03:09.949 It is clearly designed for robust apps that people 00:03:09.949 --> 00:03:12.610 actually use. It features one -click deployment. 00:03:12.990 --> 00:03:15.349 With completely integrated logs and monitoring, 00:03:15.469 --> 00:03:18.889 too. This effectively turns Claude into a serious, 00:03:18.990 --> 00:03:22.030 heavyweight rival for platforms like Vercel or 00:03:22.030 --> 00:03:24.770 Riplet. It appears to be a very strategic marriage 00:03:24.770 --> 00:03:28.030 of two distinct tools. It combines Claude artifacts, 00:03:28.409 --> 00:03:31.969 which expertly handles the visual UI, and it 00:03:31.969 --> 00:03:34.590 connects that directly to the raw power of Claude 00:03:34.590 --> 00:03:36.789 code. Right. And Claude code is fundamentally 00:03:36.789 --> 00:03:39.830 an agentic engine. Meaning? AI that can plan 00:03:39.830 --> 00:03:43.219 and execute tasks on its own. Beep. Honestly, 00:03:43.319 --> 00:03:45.639 I still wrestle with prompt drift myself when 00:03:45.639 --> 00:03:48.139 building apps. You start with a clear goal and 00:03:48.139 --> 00:03:51.259 the AI slowly loses the plot. Yeah, we all do. 00:03:51.520 --> 00:03:54.280 The idea of an AI reliably managing the entire 00:03:54.280 --> 00:03:57.039 stack architecture is wild. It is like a car 00:03:57.039 --> 00:04:00.120 engine manufacturer making a sudden pivot. Imagine 00:04:00.120 --> 00:04:02.060 if Ford suddenly decided they didn't just want 00:04:02.060 --> 00:04:04.180 to sell engines anymore. They decide to build 00:04:04.180 --> 00:04:06.020 the actual cars themselves. They decide to earn 00:04:06.020 --> 00:04:08.180 all the local dealerships. And they start running 00:04:08.180 --> 00:04:10.159 the neighborhood gas stations, too. They want 00:04:10.159 --> 00:04:12.360 to control the entire vertical pipeline, right? 00:04:12.419 --> 00:04:15.080 And they capture all that lost downstream revenue. 00:04:15.479 --> 00:04:18.639 If a startup can get a mass evaluation just wrapping 00:04:18.639 --> 00:04:22.040 an API, Anthropic has every financial incentive 00:04:22.040 --> 00:04:24.399 to reclaim that immense value for themselves. 00:04:24.740 --> 00:04:26.899 Makes sense. The vibe -coding crown is moving 00:04:26.899 --> 00:04:29.680 rapidly back to the original source. That brings 00:04:29.680 --> 00:04:31.399 up a really crucial question for the broader 00:04:31.399 --> 00:04:34.560 industry. Will this completely stifle startup 00:04:34.560 --> 00:04:37.980 innovation in the AI wrapper space? Startups 00:04:37.980 --> 00:04:40.060 will definitely be forced to change their entire 00:04:40.060 --> 00:04:43.360 approach. Simply building a pretty user interface 00:04:43.360 --> 00:04:46.740 is no longer a viable defensive moat. They are 00:04:46.740 --> 00:04:49.180 going to be forced to innovate on unique proprietary 00:04:49.180 --> 00:04:52.139 data. They have to solve highly specific complex 00:04:52.139 --> 00:04:54.839 workflow problems now. So startups must innovate 00:04:54.839 --> 00:04:58.500 beyond just wrapping existing AI APIs. Pete, 00:04:58.500 --> 00:05:00.579 this aggressive move isn't just threatening smaller 00:05:00.579 --> 00:05:03.339 startups, though. It is sending massive shockwaves 00:05:03.339 --> 00:05:05.000 through the biggest tech giants in the world. 00:05:05.040 --> 00:05:07.779 Absolutely. It is triggering real panic and some 00:05:07.779 --> 00:05:11.230 very weird new corporate innovations. After these 00:05:11.230 --> 00:05:13.509 recent moves, Microsoft has officially triggered 00:05:13.509 --> 00:05:16.189 a co -pilot code red. Let's unpack the logic 00:05:16.189 --> 00:05:19.069 behind that specific corporate panic. Microsoft 00:05:19.069 --> 00:05:21.930 CEO Satya Nadella is personally pushing major 00:05:21.930 --> 00:05:25.050 system upgrades. He wants to fix glaring performance 00:05:25.050 --> 00:05:28.029 concerns immediately. They are desperately trying 00:05:28.029 --> 00:05:30.389 to stay ahead of Anthropic's rapidly expanding 00:05:30.389 --> 00:05:32.629 capabilities. You can see the corporate AI arms 00:05:32.629 --> 00:05:34.939 race everywhere you look right now. Claude now 00:05:34.939 --> 00:05:37.360 works directly inside Microsoft Word. It does 00:05:37.360 --> 00:05:39.639 not just write text for you. Right. It can edit, 00:05:39.740 --> 00:05:42.259 analyze, and even answer document comments directly. 00:05:42.779 --> 00:05:45.620 You can save repeatable writing workflows as 00:05:45.620 --> 00:05:48.180 reusable custom skills. Meanwhile, Salesforce 00:05:48.180 --> 00:05:50.800 is piloting a totally different strategic approach. 00:05:51.079 --> 00:05:53.800 They want their specific merchants sitting directly 00:05:53.800 --> 00:05:56.740 inside ChatGPT. Yeah, early partners for this 00:05:56.740 --> 00:05:59.079 unique pilot include major brands like Crocs 00:05:59.079 --> 00:06:01.319 and Passin. The goal is to help their physical 00:06:01.319 --> 00:06:03.899 products show up in AI search directly. They 00:06:03.899 --> 00:06:05.800 do not want to rely solely on the traditional 00:06:05.800 --> 00:06:08.639 Google search index anymore. AI search is rapidly 00:06:08.639 --> 00:06:11.100 becoming the new digital storefront. Exactly. 00:06:11.339 --> 00:06:14.279 Then we get to the truly bizarre corporate applications 00:06:14.279 --> 00:06:17.160 of this technology. Meta is currently building 00:06:17.160 --> 00:06:20.220 an AI Mark Zuckerberg. They are meticulously 00:06:20.220 --> 00:06:24.980 training a digital CEO on his voice and his specific 00:06:24.980 --> 00:06:28.759 mannerisms. The core idea is that employees can 00:06:28.759 --> 00:06:31.459 actually talk to him. Wow. It is supposed to 00:06:31.459 --> 00:06:33.839 help remote workers feel more deeply connected 00:06:33.839 --> 00:06:36.709 to corporate leadership. I really have to pause 00:06:36.709 --> 00:06:38.870 and look at the dissonance here. We have the 00:06:38.870 --> 00:06:41.029 corporate budget to build a digital Zuckerberg 00:06:41.029 --> 00:06:44.069 for employee morale. Yet the backbone of these 00:06:44.069 --> 00:06:46.470 complex systems still relies heavily on human 00:06:46.470 --> 00:06:49.529 labor. Yeah. Gig workers at Scale .ai are speaking 00:06:49.529 --> 00:06:51.589 out about their working conditions. They report 00:06:51.589 --> 00:06:54.250 being paid pennies to label kids' faces all day. 00:06:54.410 --> 00:06:57.230 They are forced to review explicit audio and 00:06:57.230 --> 00:07:00.209 deeply disturbing images. It is a stark observation 00:07:00.209 --> 00:07:03.660 of the global AI supply chain. The industry still 00:07:03.660 --> 00:07:06.060 relies heavily on underpaid humans for critical 00:07:06.060 --> 00:07:08.720 data labeling. The models need human feedback 00:07:08.720 --> 00:07:10.879 to understand what is safe and what is toxic. 00:07:11.120 --> 00:07:13.939 It is a massive structural gap between the glossy 00:07:13.939 --> 00:07:16.339 corporate front end and the human backbone. It 00:07:16.339 --> 00:07:18.360 feels incredibly disconnected from the immense 00:07:18.360 --> 00:07:21.220 wealth being generated. This disconnect is actually 00:07:21.220 --> 00:07:23.480 creating an entirely new economy in the middle. 00:07:23.680 --> 00:07:26.220 LinkedIn is currently testing a brand new AI 00:07:26.220 --> 00:07:30.079 job marketplace. People can earn up to $150 an 00:07:30.079 --> 00:07:33.529 hour. They are being paid simply to train and 00:07:33.529 --> 00:07:36.410 refine chatbots. AI trainers might quickly become 00:07:36.410 --> 00:07:39.230 one of the hottest new online jobs. We are also 00:07:39.230 --> 00:07:42.189 seeing massive strategic moves in physical infrastructure. 00:07:42.750 --> 00:07:46.389 Rapidus just raised an astonishing $1 .7 billion 00:07:46.389 --> 00:07:49.850 for chip manufacturing. They are focusing entirely 00:07:49.850 --> 00:07:53.550 on building two nanometer AI chips. The primary 00:07:53.550 --> 00:07:56.250 goal is to ensure a highly reliable domestic 00:07:56.250 --> 00:07:59.560 supply chain. Everyone wants independence from 00:07:59.560 --> 00:08:02.019 foreign technology dependencies right now. Hardware 00:08:02.019 --> 00:08:04.220 is just as critical as the software layer. But 00:08:04.220 --> 00:08:06.040 let's loop back to that Microsoft panic for a 00:08:06.040 --> 00:08:08.319 second. Why is Microsoft suddenly hitting the 00:08:08.319 --> 00:08:10.959 code red panic button now? Well, Microsoft's 00:08:10.959 --> 00:08:13.180 entire empire is built on dominating enterprise 00:08:13.180 --> 00:08:16.079 software tools. If Claude natively handles the 00:08:16.079 --> 00:08:18.980 whole stack, Microsoft loses its core developer 00:08:18.980 --> 00:08:21.660 ecosystem. They are terrified of Anthropic eating 00:08:21.660 --> 00:08:25.069 their enterprise software dominance. Welcome 00:08:25.069 --> 00:08:27.829 back to our deep dive. We've explored how billionaires 00:08:27.829 --> 00:08:29.810 and gig workers are battling over the underlying 00:08:29.810 --> 00:08:32.529 infrastructure. But everyday users really just 00:08:32.529 --> 00:08:34.570 want tools that make their daily lives easier. 00:08:34.730 --> 00:08:37.250 Yeah, we are seeing that powerful agentic shift 00:08:37.250 --> 00:08:40.230 trickle down to everyday workflows. Let's look 00:08:40.230 --> 00:08:42.990 at a few empowered AI tools making serious waves 00:08:42.990 --> 00:08:45.360 today. The first one we need to look at is the 00:08:45.360 --> 00:08:48.220 Crisp Accent Converter. It is a free extension 00:08:48.220 --> 00:08:51.539 that alters digital audio in real time. You can 00:08:51.539 --> 00:08:54.659 hear any accented English on YouTube much more 00:08:54.659 --> 00:08:57.440 clearly. The underlying mechanism here is actually 00:08:57.440 --> 00:08:59.980 quite brilliant. It does not just turn up the 00:08:59.980 --> 00:09:02.659 volume or filter background noise. It dynamically 00:09:02.659 --> 00:09:05.440 reconstructs the audio waveforms locally on your 00:09:05.440 --> 00:09:08.120 specific device. That is fascinating. It works 00:09:08.120 --> 00:09:11.460 seamlessly for complex courses, podcasts, interviews, 00:09:11.720 --> 00:09:14.919 or any streaming video. It is a massive breakthrough 00:09:14.919 --> 00:09:17.620 for global education and basic accessibility. 00:09:18.000 --> 00:09:20.559 The next fascinating tool is called Luma Agents. 00:09:20.679 --> 00:09:23.059 This one is aimed directly at creative professionals 00:09:23.059 --> 00:09:25.799 and digital marketers. It essentially acts as 00:09:25.799 --> 00:09:27.860 an entire production studio in your browser. 00:09:28.019 --> 00:09:30.700 It lets you plan, generate, and meticulously 00:09:30.700 --> 00:09:33.480 iterate in one continuous pipeline. You can manage 00:09:33.480 --> 00:09:36.240 complex video, high resolution images, and dynamic 00:09:36.240 --> 00:09:39.029 audio seamlessly together. It is absolutely perfect 00:09:39.029 --> 00:09:41.789 for managing broad brand campaigns or complex 00:09:41.789 --> 00:09:44.870 social media ads. Then we have Clio Labs tackling 00:09:44.870 --> 00:09:48.610 a seemingly very boring problem. But it is a 00:09:48.610 --> 00:09:50.850 problem that costs international businesses millions 00:09:50.850 --> 00:09:54.629 of dollars. They use AI to automate global compliance 00:09:54.629 --> 00:09:57.269 for selling physical products. Think about what 00:09:57.269 --> 00:09:59.370 it actually takes to ship a physical product 00:09:59.370 --> 00:10:01.840 globally. You cross one international border 00:10:01.840 --> 00:10:04.580 and suddenly you are hit with a hundred new safety 00:10:04.580 --> 00:10:06.840 regulations. Right. Those complex regulations 00:10:06.840 --> 00:10:09.799 change constantly from one specific country to 00:10:09.799 --> 00:10:13.159 the next. Clio Labs uses agentic AI to read those 00:10:13.159 --> 00:10:16.100 local laws in real time. It automatically updates 00:10:16.100 --> 00:10:18.460 your compliance paperwork before the product 00:10:18.460 --> 00:10:20.799 even hits the customs checkpoint. Finally, we 00:10:20.799 --> 00:10:23.840 have a great free diagnostic tool called Skills 00:10:23.840 --> 00:10:26.720 Janitor. It is basically an automated auditor 00:10:26.720 --> 00:10:29.500 for your cloud code habits. It closely analyzes 00:10:29.500 --> 00:10:31.539 which AI skills you actually use on a regular 00:10:31.539 --> 00:10:33.679 basis. It shows you exactly what is fundamentally 00:10:33.679 --> 00:10:36.440 broken in your daily workflow. It finds overlapping 00:10:36.440 --> 00:10:38.620 redundant prompts and highlights powerful features 00:10:38.620 --> 00:10:41.559 you completely ignore. It is like having a brilliant 00:10:41.559 --> 00:10:44.720 personal assistant silently clean up your digital 00:10:44.720 --> 00:10:47.820 workspace. But looking closely at all these highly 00:10:47.820 --> 00:10:52.059 specific fragmented tools, do these highly specific 00:10:52.059 --> 00:10:54.860 fragmented tools eventually consolidate into 00:10:54.860 --> 00:10:57.740 one platform? History shows us that heavy consolidation 00:10:57.740 --> 00:11:01.090 in tech is practically inevitable. The big foundational 00:11:01.090 --> 00:11:03.669 models are closely tracking these successful 00:11:03.669 --> 00:11:06.309 niche features. They will absolutely try to swallow 00:11:06.309 --> 00:11:08.450 these capabilities into their native platforms. 00:11:08.750 --> 00:11:11.070 Eventually, yes, the big players will absorb 00:11:11.070 --> 00:11:15.110 these niche tools. We have zoomed in on the underlying 00:11:15.110 --> 00:11:17.590 code and the corporate battles. We have explored 00:11:17.590 --> 00:11:19.750 the practical consumer tools you can use right 00:11:19.750 --> 00:11:23.409 now. Now it is time to zoom all the way out. 00:11:23.529 --> 00:11:26.269 We need to look at the massive global macro picture 00:11:26.269 --> 00:11:28.940 shaping your future. Stanford University just 00:11:28.940 --> 00:11:32.820 dropped their massive 2026 AI index report. It 00:11:32.820 --> 00:11:34.840 formally confirms that we are officially hitting 00:11:34.840 --> 00:11:37.299 a jagged frontier. The term jagged frontier perfectly 00:11:37.299 --> 00:11:39.860 describes the current strange capability gap. 00:11:40.059 --> 00:11:42.820 AI is now consistently outperforming humans in 00:11:42.820 --> 00:11:45.600 PhD level scientific research. Yet, as you mentioned 00:11:45.600 --> 00:11:47.500 earlier in the show, it struggles with basic 00:11:47.500 --> 00:11:50.759 physical context. It still fails to read an analog 00:11:50.759 --> 00:11:54.000 clock 50 % of the time. The Google Gemini DeepThink 00:11:54.000 --> 00:11:56.720 model is a perfect example of this. It can easily 00:11:56.720 --> 00:11:59.019 win a gold medal at the Global Math Olympiad. 00:11:59.220 --> 00:12:01.960 But the jagged nature of these specific models 00:12:01.960 --> 00:12:05.240 is highly unpredictable. It fails nearly one 00:12:05.240 --> 00:12:08.480 in three basic computer navigation tasks. It 00:12:08.480 --> 00:12:12.240 is incredibly computationally capable, yet shockingly 00:12:12.240 --> 00:12:15.379 brittle at the exact same time. It is like hiring 00:12:15.379 --> 00:12:17.500 a brilliant Nobel laureate for your engineering 00:12:17.500 --> 00:12:20.039 company. But they occasionally forget how to 00:12:20.039 --> 00:12:23.440 tie their own shoes. Two sec silence. The Stanford 00:12:23.440 --> 00:12:25.799 report also gave us a stark look at the global 00:12:25.799 --> 00:12:29.100 geopolitical race. We are just reporting Stanford's 00:12:29.100 --> 00:12:31.059 impartial data here without taking any sides. 00:12:31.379 --> 00:12:34.000 The U .S. technological lead over China has effectively 00:12:34.000 --> 00:12:36.600 evaporated completely. Anthropic's top foundational 00:12:36.600 --> 00:12:39.440 model currently leads China's absolute best model. 00:12:39.539 --> 00:12:42.779 But the lead is a razor thin 2 .7 percent margin. 00:12:42.980 --> 00:12:45.659 That is a statistical dead heat in the fast paced 00:12:45.659 --> 00:12:47.889 technology world. The U .S. still completely 00:12:47.889 --> 00:12:50.470 dominates in building massive physical data centers, 00:12:50.669 --> 00:12:54.009 and they lead heavily in private corporate financial 00:12:54.009 --> 00:12:57.769 investment. They have poured a staggering $285 00:12:57.769 --> 00:13:02.610 .9 billion into AI. But China is aggressively 00:13:02.610 --> 00:13:05.610 leading in several other crucial long -term metrics. 00:13:05.870 --> 00:13:08.370 They currently lead the world in new AI patents 00:13:08.370 --> 00:13:11.429 and academic research citations. They're also 00:13:11.429 --> 00:13:13.990 heavily dominating the physical deployment of 00:13:13.990 --> 00:13:16.740 industrial robotics. They are building the actual 00:13:16.740 --> 00:13:19.200 physical manifestation of artificial intelligence. 00:13:19.440 --> 00:13:21.779 Exactly. Despite the massive financial investment, 00:13:22.019 --> 00:13:24.620 the U .S. has a critical structural vulnerability. 00:13:25.139 --> 00:13:27.840 They are facing a severely worsening human talent 00:13:27.840 --> 00:13:30.620 crisis right now. The number of top AI researchers 00:13:30.620 --> 00:13:33.220 moving to the States has absolutely plummeted. 00:13:33.220 --> 00:13:36.559 We have seen an 89 percent drop since 2017. And 00:13:36.559 --> 00:13:38.279 the short -term data trend is actually looking 00:13:38.279 --> 00:13:41.220 much worse. There was an 80 % drop in just the 00:13:41.220 --> 00:13:44.259 last year alone. Beat. We also need to critically 00:13:44.259 --> 00:13:47.340 examine the growing global trust gap. Stanford's 00:13:47.340 --> 00:13:50.299 massive data on global public perception is really 00:13:50.299 --> 00:13:53.820 revealing. Only 31 % of Americans currently trust 00:13:53.820 --> 00:13:56.840 their government to regulate AI. That is currently 00:13:56.840 --> 00:13:59.820 the absolute lowest level of public trust globally. 00:14:00.399 --> 00:14:02.759 Meanwhile, the European Union has emerged with 00:14:02.759 --> 00:14:05.620 a very different public reputation. The EU is 00:14:05.620 --> 00:14:07.899 currently the most trusted entity for effective 00:14:07.899 --> 00:14:10.799 regulatory oversight. It is fascinating to see 00:14:10.799 --> 00:14:12.779 how different global regions view the fundamental 00:14:12.779 --> 00:14:15.659 safety of this technology. But regardless of 00:14:15.659 --> 00:14:18.480 varying public trust, the adoption speed is simply 00:14:18.480 --> 00:14:21.759 staggering. Generative AI officially reached 00:14:21.759 --> 00:14:26.049 53 % global population adoption. It achieved 00:14:26.049 --> 00:14:28.350 that incredible milestone in just three short 00:14:28.350 --> 00:14:30.990 years. Which is completely unheard of. Whoa. 00:14:31.570 --> 00:14:34.529 Imagine 53 % of the global population adopting 00:14:34.529 --> 00:14:37.230 a tech in just three years. That scale is unprecedented. 00:14:37.730 --> 00:14:40.309 The U .S. consumer economic value of these tools 00:14:40.309 --> 00:14:42.429 is absolutely massive. It is currently estimated 00:14:42.429 --> 00:14:46.330 at $172 billion annually. We've never seen a 00:14:46.330 --> 00:14:48.570 new technology integrate into daily human life 00:14:48.570 --> 00:14:50.789 this quickly. Not the advent of the Internet, 00:14:50.970 --> 00:14:53.610 not mobile smartphones. Absolutely nothing compares 00:14:53.610 --> 00:14:56.100 to this speed. Let me ask you about the implications 00:14:56.100 --> 00:14:59.340 of that severe talent drain data. Can the U .S. 00:14:59.379 --> 00:15:02.240 maintain its technological edge with this severe 00:15:02.240 --> 00:15:05.220 talent drain? Massive capital and sprawling data 00:15:05.220 --> 00:15:07.799 centers are incredibly important structural advantages. 00:15:08.159 --> 00:15:11.080 But eventually you need elite human brains pushing 00:15:11.080 --> 00:15:14.960 the actual algorithmic boundaries. If top global 00:15:14.960 --> 00:15:17.759 researchers stop immigrating, the domestic innovation 00:15:17.759 --> 00:15:20.299 curve will eventually flatten. Without foreign 00:15:20.299 --> 00:15:22.919 researchers, the U .S. relies purely on compute 00:15:22.919 --> 00:15:26.960 power. We have covered a massive amount of complex 00:15:26.960 --> 00:15:29.399 ground today. Let's pull all these diverse threads 00:15:29.399 --> 00:15:31.620 together for you. We are clearly living on a 00:15:31.620 --> 00:15:33.820 very jagged technological frontier right now. 00:15:33.940 --> 00:15:36.379 The underlying technology is immensely capable, 00:15:36.600 --> 00:15:39.539 but it remains surprisingly brittle. The corporate 00:15:39.539 --> 00:15:42.259 software war is officially hitting a code red 00:15:42.259 --> 00:15:45.019 panic level. The nimble startup landscape is 00:15:45.019 --> 00:15:47.639 being rapidly swallowed by the massive API providers, 00:15:47.879 --> 00:15:50.759 and the global geopolitical race is demonstrably 00:15:50.759 --> 00:15:53.460 tighter than it has ever been. It leaves us with 00:15:53.460 --> 00:15:55.799 a really provocative philosophical question for 00:15:55.799 --> 00:15:58.399 you to consider. Think about AI successfully 00:15:58.399 --> 00:16:00.980 handling all the basic scaffolding of our digital 00:16:00.980 --> 00:16:04.480 lives. It writes the complex full stack code 00:16:04.480 --> 00:16:07.639 and completely automates global regulatory compliance. 00:16:08.120 --> 00:16:11.720 It even accurately mimic our corporate CEOs for 00:16:11.720 --> 00:16:14.740 remote town halls. If a machine does all the 00:16:14.740 --> 00:16:16.980 heavy digital lifting, what actually changes? 00:16:17.039 --> 00:16:19.519 What becomes the new uniquely human defining 00:16:19.519 --> 00:16:22.860 skill in this new world? What exactly happens 00:16:22.860 --> 00:16:25.700 to human intuition when the daily cognitive struggle 00:16:25.700 --> 00:16:28.620 is permanently removed? That is something incredibly 00:16:28.620 --> 00:16:31.120 profound to seriously mull over this week. We 00:16:31.120 --> 00:16:32.899 invite you to pay close attention to your own 00:16:32.899 --> 00:16:35.360 daily workflows. Watch out for those jagged, 00:16:35.360 --> 00:16:37.620 unpredictable interactions with AI in your own 00:16:37.620 --> 00:16:39.720 life. Notice exactly where it acts like a profound 00:16:39.720 --> 00:16:42.019 genius and where it bizarrely stumbles. Thank 00:16:42.019 --> 00:16:44.460 you for joining our deep dive today. OTRO Music.