WEBVTT 00:00:00.000 --> 00:00:04.519 Did you know that where you live might significantly 00:00:04.519 --> 00:00:07.440 predict how much you use artificial intelligence? 00:00:07.719 --> 00:00:10.880 Yeah, it's kind of surprising. And maybe even 00:00:10.880 --> 00:00:13.539 more surprisingly, who's really winning the global 00:00:13.539 --> 00:00:17.000 AI race? Well, it might just challenge what you 00:00:17.000 --> 00:00:20.719 think. Or imagine this, typing out messages easily 00:00:20.719 --> 00:00:23.399 just by thinking them. Wow. It's not science 00:00:23.399 --> 00:00:44.090 fiction anymore. It's actually happening. Today, 00:00:44.170 --> 00:00:46.950 we're looking at a really fascinating exploration 00:00:46.950 --> 00:00:50.649 of the very latest in AI. From all angles. Exactly. 00:00:50.850 --> 00:00:53.149 Yeah. From the uneven global landscape of adoption. 00:00:53.310 --> 00:00:55.429 Yeah. All the way to this kind of mind -bending, 00:00:55.450 --> 00:00:57.590 brain -like computing that could totally redefine 00:00:57.590 --> 00:00:59.649 our tech future. Yeah, some big shifts potentially. 00:00:59.909 --> 00:01:02.250 Lots to unpack there. Our mission, like always, 00:01:02.469 --> 00:01:05.530 is to pull out the most important nuggets, give 00:01:05.530 --> 00:01:07.609 you a shortcut to being really well -informed 00:01:07.609 --> 00:01:10.250 without all that info overload. Let's do it. 00:01:10.430 --> 00:01:13.379 So let's dive in. OK, we're kicking off with 00:01:13.379 --> 00:01:15.560 some really compelling data. This isn't just 00:01:15.560 --> 00:01:19.019 theory or predictions. Real world stuff. Anthropic. 00:01:19.099 --> 00:01:22.180 You know, the team behind the AI model, Claude, 00:01:22.219 --> 00:01:24.500 they just put out this massive usage report. 00:01:25.099 --> 00:01:28.400 It's based on over a million anonymized conversations 00:01:28.400 --> 00:01:32.299 from just August this year. So very recent data. 00:01:32.459 --> 00:01:34.060 Exactly. Yeah. This is what's actually happening. 00:01:34.079 --> 00:01:36.700 Who's using AI right now and how? And the findings? 00:01:36.879 --> 00:01:38.859 Well, they're pretty stark. The landscape is 00:01:38.859 --> 00:01:41.799 anything but uniform. Super uneven, actually. 00:01:41.879 --> 00:01:43.959 How so? If you look at the U .S., you've got 00:01:43.959 --> 00:01:46.120 these power users really clustered in places 00:01:46.120 --> 00:01:49.040 like California, D .C., Utah. Usual tech hubs, 00:01:49.159 --> 00:01:51.700 maybe? Sort of, yeah. But then many southern 00:01:51.700 --> 00:01:54.439 and plains states noticeably lagging behind. 00:01:54.700 --> 00:01:56.840 Okay. And it's not just a U .S. thing. Globally, 00:01:56.840 --> 00:01:58.959 the differences are even bigger. Countries like 00:01:58.959 --> 00:02:03.180 Israel, Singapore, Canada. They're way ahead 00:02:03.180 --> 00:02:05.780 in adopting AI. Way ahead. But then nations like 00:02:05.780 --> 00:02:10.400 India, Nigeria, and frankly, most lower income 00:02:10.400 --> 00:02:12.819 countries are trailing significantly behind. 00:02:13.360 --> 00:02:16.319 So it's definitely not a uniform way of washing 00:02:16.319 --> 00:02:18.560 over everyone. Not at all. It's a fascinating 00:02:18.560 --> 00:02:20.539 picture, isn't it? Yeah. And this data really 00:02:20.539 --> 00:02:24.099 matters because, well, where AI is heavily used 00:02:24.099 --> 00:02:28.300 today kind of previews. which tasks, maybe even 00:02:28.300 --> 00:02:30.819 jobs, it might start to augment or replace tomorrow. 00:02:31.180 --> 00:02:33.400 Yeah, it's a leading indicator. And this feeds 00:02:33.400 --> 00:02:36.180 directly into that whole productivity gap discussion, 00:02:36.319 --> 00:02:38.719 right? That growing difference in output between 00:02:38.719 --> 00:02:41.800 people or companies using AI effectively and 00:02:41.800 --> 00:02:43.919 those who aren't. Absolutely. And what's really 00:02:43.919 --> 00:02:46.199 striking, maybe even counterintuitive, is that 00:02:46.199 --> 00:02:48.520 high skill workers seem to be benefiting the 00:02:48.520 --> 00:02:50.500 most right now. Not necessarily leveling the 00:02:50.500 --> 00:02:52.729 field then. Not yet, it seems. They're using 00:02:52.729 --> 00:02:55.969 AI to fill knowledge gaps faster, speed up decisions, 00:02:56.289 --> 00:02:59.129 basically multiply their output. So instead of 00:02:59.129 --> 00:03:01.030 closing the gap, it's actually widening it. It's 00:03:01.030 --> 00:03:03.310 amplifying the advantages for those already kind 00:03:03.310 --> 00:03:05.330 of ahead. That's quite a finding. Not the narrative 00:03:05.330 --> 00:03:07.949 we often hear. And I do appreciate Anthropic 00:03:07.949 --> 00:03:10.210 being transparent here. They didn't just publish 00:03:10.210 --> 00:03:12.289 the findings. They actually open sourced the 00:03:12.289 --> 00:03:14.479 data set. Oh, that's great for researchers. Yeah. 00:03:14.580 --> 00:03:16.539 And they introduced something called an AI usage 00:03:16.539 --> 00:03:20.259 index, basically a tool to track how AI spreads 00:03:20.259 --> 00:03:22.560 over time in different places. That'll be useful 00:03:22.560 --> 00:03:24.860 for policymakers, too, I bet. Definitely. Provides 00:03:24.860 --> 00:03:28.060 some real data to work with. So thinking about 00:03:28.060 --> 00:03:31.979 this really uneven global adoption, what does 00:03:31.979 --> 00:03:35.139 that tell us about AI's immediate impact on different 00:03:35.139 --> 00:03:38.319 economies? Well, it strongly suggests AI will 00:03:38.319 --> 00:03:40.819 probably deepen existing divides before it starts 00:03:40.819 --> 00:03:43.360 leveling things out. Yeah, that's a sobering 00:03:43.360 --> 00:03:45.159 thought, the whole divide thing. But while those 00:03:45.159 --> 00:03:47.080 big trends play out, let's pivot. Let's look 00:03:47.080 --> 00:03:49.319 at the sheer speed of innovation happening right 00:03:49.319 --> 00:03:52.360 now. Okay, some rapid fire stuff. Exactly. Our 00:03:52.360 --> 00:03:54.979 Today in AI segment, get ready for some pretty 00:03:54.979 --> 00:03:57.460 wild things. First up, a founder actually built 00:03:57.460 --> 00:04:02.900 a $100 media brand completely solo. Solo? How? 00:04:03.000 --> 00:04:05.539 Using an AI agent. Now, an AI agent, just so 00:04:05.539 --> 00:04:07.500 we're clear, is basically software that acts 00:04:07.500 --> 00:04:09.939 on its own to do tasks. Right, autonomously. 00:04:10.319 --> 00:04:13.800 This one scrapes news, writes content, makes 00:04:13.800 --> 00:04:16.680 videos, even posts everything. That's incredible. 00:04:16.779 --> 00:04:19.120 One person doing the work of a small team. Pretty 00:04:19.120 --> 00:04:21.100 much. It just completely changes the game for 00:04:21.100 --> 00:04:23.040 entrepreneurs, you know. It really does. Lowers 00:04:23.040 --> 00:04:25.579 that barrier to entry significantly. What else 00:04:25.579 --> 00:04:27.420 is catching your eye? Okay, then there's this. 00:04:28.480 --> 00:04:32.379 Almost telepathic wearable. Uh -oh, sci -fi alert. 00:04:32.720 --> 00:04:35.379 Feels like it. It's being called the world's 00:04:35.379 --> 00:04:38.160 first wearable that lets you speak without moving 00:04:38.160 --> 00:04:41.290 your mouth. Wait, what? or type messages literally 00:04:41.290 --> 00:04:46.110 at the speed of thought, beat. Whoa. Just imagine 00:04:46.110 --> 00:04:48.389 the possibilities there, especially for communication 00:04:48.389 --> 00:04:51.370 accessibility. It's a huge leap, right? And yeah, 00:04:51.410 --> 00:04:53.600 early access is already open for it. That feels 00:04:53.600 --> 00:04:55.240 like something straight out of a movie. Okay. 00:04:55.560 --> 00:04:57.319 Bypassing traditional input methods entirely. 00:04:57.660 --> 00:04:59.319 Totally. You wonder about the learning curve, 00:04:59.399 --> 00:05:01.959 the practicalities. Right. But the potential 00:05:01.959 --> 00:05:04.819 is just massive. Yeah. The rollout will be fascinating. 00:05:04.899 --> 00:05:08.480 And for creators, VE'd just launched Fabric 1 00:05:08.480 --> 00:05:12.480 .0. VE, the video platform. Yep. It's an AI talking 00:05:12.480 --> 00:05:15.279 video model. It can make these emotional one 00:05:15.279 --> 00:05:17.879 -minute videos. Apparently seven times faster 00:05:17.879 --> 00:05:20.560 just from a single photo. From one photo. How 00:05:20.560 --> 00:05:22.800 realistic does it look? Surprisingly realistic. 00:05:23.060 --> 00:05:26.100 It's really blurring those lines. Also, gamma 00:05:26.100 --> 00:05:29.279 that AI presentation tool. Oh, yeah. Heard a 00:05:29.279 --> 00:05:32.319 lot about them. They just hit $50 million in 00:05:32.319 --> 00:05:35.829 annual recurring revenue. In under two years. 00:05:35.930 --> 00:05:38.029 Wow, that's fast growth. And their co -founders 00:05:38.029 --> 00:05:39.910 even shared their growth playbook, which is pretty 00:05:39.910 --> 00:05:42.750 cool. Nice. Then switching gears to robotics, 00:05:43.069 --> 00:05:47.029 Dyna Robotics raised a huge $120 million from 00:05:47.029 --> 00:05:50.529 big names like NVIDIA and Amazon. For what specifically? 00:05:50.810 --> 00:05:53.629 For their AI -powered robot arms. It's part of 00:05:53.629 --> 00:05:56.610 this massive robotics boon, like $12 .1 billion 00:05:56.610 --> 00:05:59.170 invested just this year. It's clear the money 00:05:59.170 --> 00:06:00.910 and the innovation are just pouring in everywhere, 00:06:01.050 --> 00:06:03.750 isn't it? Absolutely exploding. We also saw OpenAI. 00:06:03.949 --> 00:06:06.610 And Harvard put out that big report on chat GPT 00:06:06.610 --> 00:06:08.970 usage. Yeah, the largest public one ever, I think. 00:06:08.990 --> 00:06:11.170 Right. Showing common queries, how people really 00:06:11.170 --> 00:06:16.069 use it. And Google, DeepMind's CEO, hinted at 00:06:16.069 --> 00:06:18.329 the number one skill needed for this AI world. 00:06:18.529 --> 00:06:20.649 What was it? Didn't say explicitly, but the hint 00:06:20.649 --> 00:06:23.470 points towards not just using AI, but really 00:06:23.470 --> 00:06:26.170 effectively directing it. Judgment, prompt crafting, 00:06:26.449 --> 00:06:28.610 that kind of thing. So all these different innovations. 00:06:29.740 --> 00:06:31.899 How do they reshape our daily interactions, our 00:06:31.899 --> 00:06:34.259 work lives? They're pushing us towards faster, 00:06:34.439 --> 00:06:38.259 more intuitive, highly automated ways of doing 00:06:38.259 --> 00:06:41.199 things, demands new skills. Building on that, 00:06:41.339 --> 00:06:44.959 let's shift focus now to practical applications. 00:06:45.740 --> 00:06:49.139 Powerful new tools for automation and... Maybe 00:06:49.139 --> 00:06:51.240 more excitingly, creative work. Yeah, where the 00:06:51.240 --> 00:06:53.220 rubber meets the road for a lot of us. Exactly. 00:06:53.500 --> 00:06:55.800 Okay, so for folks deep into automation, there 00:06:55.800 --> 00:07:00.319 are these new NAN workflow guides. NAN. Remind 00:07:00.319 --> 00:07:02.519 us what a workflow is in this context. Sure. 00:07:02.680 --> 00:07:05.000 An NAN workflow is basically an automated sequence 00:07:05.000 --> 00:07:07.500 of tasks. It connects different apps and services 00:07:07.500 --> 00:07:10.019 together to do something automatically. Got it. 00:07:10.019 --> 00:07:12.740 Like Zapier, but maybe more complex. Yeah, often 00:07:12.740 --> 00:07:14.819 more customizable. So there's a guide on evaluating 00:07:14.819 --> 00:07:17.639 these workflows using actual test data to optimize 00:07:17.639 --> 00:07:19.620 them. No more guesswork. That sounds useful. 00:07:19.779 --> 00:07:22.339 Data -driven automation. And for the finance 00:07:22.339 --> 00:07:24.399 crowd, there's another guide showing how to build 00:07:24.399 --> 00:07:29.040 a no -code NAN workflow to analyze crypto, forex, 00:07:29.079 --> 00:07:32.410 stocks. All automatically. Bringing automation 00:07:32.410 --> 00:07:35.029 to personal finance analysis. Interesting. Yeah, 00:07:35.069 --> 00:07:38.209 bringing precision. You know, honestly, I still 00:07:38.209 --> 00:07:40.629 wrestle with prompt drift myself sometimes. Oh, 00:07:40.649 --> 00:07:43.149 yeah. Explain it. It's that subtle thing where 00:07:43.149 --> 00:07:45.629 the AI's understanding of your instructions kind 00:07:45.629 --> 00:07:47.850 of shifts over time, even if you use the same 00:07:47.850 --> 00:07:50.810 prompt, keeping it consistent. It's a constant 00:07:50.810 --> 00:07:52.910 effort. So crafting those prompts and optimizing 00:07:52.910 --> 00:07:55.470 the workflow is key. It's a skill we all need 00:07:55.470 --> 00:07:58.100 to keep honing. Definitely. Like trying to get 00:07:58.100 --> 00:08:00.439 a puppy to consistently fetch the right stick. 00:08:00.480 --> 00:08:02.550 That's a good analogy. It really does highlight 00:08:02.550 --> 00:08:05.990 that shift from just using AI to like mastering 00:08:05.990 --> 00:08:08.850 the art of directing it precisely. And speaking 00:08:08.850 --> 00:08:10.709 of crafting things, Google's apparently got a 00:08:10.709 --> 00:08:13.509 new AI image tool that's making pro photo editors 00:08:13.509 --> 00:08:15.889 a bit nervous. Oh, yeah. This one's a potential 00:08:15.889 --> 00:08:18.709 game changer. It's incredibly powerful. How so? 00:08:18.870 --> 00:08:21.529 You can basically go from a simple selfie to 00:08:21.529 --> 00:08:23.430 something looking like a polished magazine cover 00:08:23.430 --> 00:08:26.990 just by typing commands. Seriously, like make 00:08:26.990 --> 00:08:29.110 my hair look better or change the background 00:08:29.110 --> 00:08:32.059 to Paris. Pretty much, but much more sophisticated. 00:08:32.399 --> 00:08:35.600 It interprets complex text commands to build 00:08:35.600 --> 00:08:38.679 and manipulate visuals. High -level editing becomes 00:08:38.679 --> 00:08:41.639 super simple. Wow. So it really democratizes 00:08:41.639 --> 00:08:43.820 that kind of high -end visual work. Absolutely. 00:08:43.919 --> 00:08:46.320 But, you know, also raises questions about traditional 00:08:46.320 --> 00:08:48.519 photo editing roles. And we're also seeing a 00:08:48.519 --> 00:08:51.059 flood of what we're calling new empowered AI 00:08:51.059 --> 00:08:53.899 tools just popping up constantly. Yeah, a whole 00:08:53.899 --> 00:08:56.799 ecosystem. Things like admin tools offers over 00:08:56.799 --> 00:09:00.750 125 administrative tools. in one place. Handy. 00:09:01.049 --> 00:09:03.250 Cliptery, which claims to turn a script straight 00:09:03.250 --> 00:09:05.570 into a YouTube video. Script to video, getting 00:09:05.570 --> 00:09:07.850 better all the time. Summa AI for generating 00:09:07.850 --> 00:09:10.629 personalized LinkedIn posts based on news. Keeping 00:09:10.629 --> 00:09:12.929 that feed active. And Prompt Prism, this one 00:09:12.929 --> 00:09:16.480 takes your, like... messy ideas and turns them 00:09:16.480 --> 00:09:19.220 into detailed prompts for advanced video models 00:09:19.220 --> 00:09:22.200 like VO3. So helping with the prompt crafting 00:09:22.200 --> 00:09:25.179 itself. Exactly. That's the core of prompt engineering, 00:09:25.399 --> 00:09:27.919 right? Crafting those precise instructions to 00:09:27.919 --> 00:09:31.379 get the AI to do exactly what you want. So how 00:09:31.379 --> 00:09:34.480 do these tools democratize these complex AI capabilities? 00:09:34.820 --> 00:09:38.139 How do they empower everyday users? They basically 00:09:38.139 --> 00:09:40.500 make sophisticated tasks much more accessible, 00:09:40.620 --> 00:09:42.440 even if you don't have deep technical skills. 00:09:42.740 --> 00:09:45.940 Okay, let's do a quick run through some AI quick 00:09:45.940 --> 00:09:48.679 hits, get a feel for the industry pulse, maybe 00:09:48.679 --> 00:09:51.080 the human side of things, too. Sounds good. What's 00:09:51.080 --> 00:09:53.740 buzzing? Well, Sam Altman, OpenAI's CEO, admitted 00:09:53.740 --> 00:09:56.120 recently he hasn't slept well since GPT first 00:09:56.120 --> 00:09:58.480 launched. Wow. Gives you a sense of the pressure, 00:09:58.539 --> 00:10:00.700 huh? Totally. The immense stakes involved there. 00:10:00.820 --> 00:10:03.879 And speaking of open AI, China's Tencent just 00:10:03.879 --> 00:10:05.600 poached one of their prominent researchers. Oh, 00:10:05.600 --> 00:10:08.059 ooh, the talent wars continue. Fiercely competitive. 00:10:08.759 --> 00:10:11.600 Then, on a lighter note, Gemini had this funny 00:10:11.600 --> 00:10:14.340 moment failing to make a seahorse emoji. Made 00:10:14.340 --> 00:10:17.320 it seem like it briefly lost its mind. Happens 00:10:17.320 --> 00:10:19.919 to the best of us, I guess. Even AI needs a moment. 00:10:20.179 --> 00:10:22.399 Right. But then, kind of surprisingly, Gemini 00:10:22.399 --> 00:10:25.120 actually overtook ChatGPT as the top iOS app 00:10:25.120 --> 00:10:28.059 for a bit. Really? Mostly due to this weirdly 00:10:28.059 --> 00:10:32.179 viral banana moment it had. You just can't predict 00:10:32.179 --> 00:10:34.500 what captures the public's attention sometimes. 00:10:34.919 --> 00:10:37.179 The Internet is strange. Tell me about it. And 00:10:37.179 --> 00:10:39.240 for the developers out there, OpenAI upgraded 00:10:39.240 --> 00:10:42.299 GPT -5 codecs. That coding model. Yeah. Enhanced 00:10:42.299 --> 00:10:44.679 capabilities for writing code, debugging, all 00:10:44.679 --> 00:10:47.919 that jazz. So continuous progress under the hood. 00:10:48.059 --> 00:10:50.600 Looking at these little snippets, what do they 00:10:50.600 --> 00:10:53.899 reveal about? the human side of AI development 00:10:53.899 --> 00:10:56.600 beyond just the tech. They really show both the 00:10:56.600 --> 00:11:00.139 intense pressures and the sometimes unexpected, 00:11:00.419 --> 00:11:03.320 almost quirky challenges in this whole AI race. 00:11:03.559 --> 00:11:05.779 Okay, now here's where things get, I think, really 00:11:05.779 --> 00:11:09.259 interesting. Potentially paradigm shifting. We've 00:11:09.259 --> 00:11:11.059 talked a lot about current AI models, the ones 00:11:11.059 --> 00:11:13.360 most people know, but there's this fascinating 00:11:13.360 --> 00:11:16.440 AI view coming out of China. Spiking Brain 1 00:11:16.440 --> 00:11:19.399 .0. Ah, yes, this is huge. And it seems to directly 00:11:19.399 --> 00:11:22.379 challenge the whole NVIDIA stack dominance we 00:11:22.379 --> 00:11:24.139 see everywhere else. It absolutely does. This 00:11:24.139 --> 00:11:25.740 could be a monumental development, seriously. 00:11:26.019 --> 00:11:28.320 Chinese researchers unveiled this completely 00:11:28.320 --> 00:11:31.820 new neural -inspired AI system. Neural -inspired, 00:11:31.919 --> 00:11:34.799 meaning? Right. It literally mimics how our actual 00:11:34.799 --> 00:11:37.580 brains process information. And here's the kicker. 00:11:37.659 --> 00:11:40.799 Okay. It runs entirely on Chinese Medic X chips. 00:11:41.179 --> 00:11:44.840 Zero reliance on NVIDIA hardware. None. Wow. 00:11:45.360 --> 00:11:48.659 So completely independent silicon. Exactly. That's 00:11:48.659 --> 00:11:51.559 not just an engineering achievement. It's a massive 00:11:51.559 --> 00:11:54.120 strategic play for technological sovereignty. 00:11:54.480 --> 00:11:56.759 So how is it fundamentally different? How does 00:11:56.759 --> 00:12:00.080 it work compared to, say, GBT -4 or Gemini. OK, 00:12:00.179 --> 00:12:03.620 so most current models work token by token, right? 00:12:03.759 --> 00:12:06.759 They process information sequentially, like reading 00:12:06.759 --> 00:12:09.019 word by word. Right, piece by piece. Spiking 00:12:09.019 --> 00:12:11.779 brain 1 .0 doesn't activate the whole model for 00:12:11.779 --> 00:12:15.000 every token. Instead, it mimics how neurons in 00:12:15.000 --> 00:12:18.039 our brain fire selectively. Only the needed parts 00:12:18.039 --> 00:12:21.139 activate. Ah, like only using the relevant brain 00:12:21.139 --> 00:12:23.419 cells for a specific thought. Precisely. It's 00:12:23.419 --> 00:12:25.559 called a neuromorphic approach, computing inspired 00:12:25.559 --> 00:12:27.879 directly by brain structure and function. And 00:12:27.879 --> 00:12:29.700 they've already trained pretty big models, 7 00:12:29.700 --> 00:12:32.740 billion and 76 billion parameters, using this 00:12:32.740 --> 00:12:34.980 architecture. That's impressive scale for such 00:12:34.980 --> 00:12:37.139 a different approach. It's not just a lab experiment, 00:12:37.240 --> 00:12:39.100 then. Definitely not. And the efficiency gains 00:12:39.100 --> 00:12:42.759 are just staggering. How efficient are we talking? 00:12:42.940 --> 00:12:46.320 Get this. It used less than 2 % of the training 00:12:46.320 --> 00:12:48.980 data standard models need, but it matched their 00:12:48.980 --> 00:12:51.399 performance on core language tasks. Less than 00:12:51.399 --> 00:12:54.059 2%. That's incredible. And it was stable. Ran 00:12:54.059 --> 00:12:56.750 for weeks without crashing. But here's the real 00:12:56.750 --> 00:13:00.029 headline number. Go on. On a massive 4 million 00:13:00.029 --> 00:13:03.740 token prompt. That's a huge amount of text. It 00:13:03.740 --> 00:13:06.460 ran 100 times faster than traditional models. 00:13:06.679 --> 00:13:10.879 100 times faster. 100 times. Whoa. Just imagine 00:13:10.879 --> 00:13:13.679 scaling that kind of efficiency. It's like stacking 00:13:13.679 --> 00:13:16.820 Lego blocks of data compared to, I don't know, 00:13:16.919 --> 00:13:19.639 building with bricks. It completely changes the 00:13:19.639 --> 00:13:23.000 economics and speed of AI. 100Xs. That's almost 00:13:23.000 --> 00:13:24.759 hard to comprehend in this field. Yeah. That's 00:13:24.759 --> 00:13:26.740 a generational leap, not an incremental one. 00:13:26.960 --> 00:13:28.940 And I heard there's a public version people can 00:13:28.940 --> 00:13:30.940 try. Yes. Incredibly, they released a public. 00:13:31.049 --> 00:13:33.289 version called shunxi it's free to test online 00:13:33.289 --> 00:13:36.049 and it runs on fully hosted on those chinese 00:13:36.049 --> 00:13:38.809 medic x chips it's a direct real world demonstration 00:13:38.809 --> 00:13:41.889 look we can do this entirely without us tech 00:13:41.889 --> 00:13:44.440 that's a powerful statement It really is. It 00:13:44.440 --> 00:13:47.620 feels like maybe the mouse moment for brain -like 00:13:47.620 --> 00:13:49.840 computing, the very beginning of something potentially 00:13:49.840 --> 00:13:52.940 huge. And it makes you ask, where are Meta's 00:13:52.940 --> 00:13:55.960 or OpenAI's neuromorphic projects? You have to 00:13:55.960 --> 00:13:57.539 assume they're working on something similar, 00:13:57.700 --> 00:14:00.120 right? So how might this kind of brain -like 00:14:00.120 --> 00:14:03.539 computing, like spiking brain 1 .0, fundamentally 00:14:03.539 --> 00:14:06.779 alter the AI development landscape and maybe 00:14:06.779 --> 00:14:10.840 global tech power? It really promises vastly 00:14:10.840 --> 00:14:13.480 more efficient, maybe even self -sustaining AI 00:14:13.480 --> 00:14:15.860 that could dramatically shift the geopolitical 00:14:15.860 --> 00:14:18.610 tech balance. Sponsor. Mid -roll sponsor read. 00:14:18.690 --> 00:14:20.950 So as we wrap up this deep dive, it seems abundantly 00:14:20.950 --> 00:14:24.309 clear AI is transforming things incredibly rapidly. 00:14:24.509 --> 00:14:27.730 But it's crucial to see it's not a uniform way 00:14:27.730 --> 00:14:30.009 washing over everyone equally. Definitely not. 00:14:30.149 --> 00:14:32.389 We've seen these stark divides in global adoption. 00:14:32.450 --> 00:14:34.409 It highlights who's benefiting most right now 00:14:34.409 --> 00:14:37.470 and worryingly where existing gaps might actually 00:14:37.470 --> 00:14:39.850 be getting wider. The productivity gap issue. 00:14:39.970 --> 00:14:42.549 Exactly. Yet at the same time, there's this absolute 00:14:42.549 --> 00:14:45.909 explosion of innovation. Across everything from 00:14:45.909 --> 00:14:48.029 personal tools, letting one person build a media 00:14:48.029 --> 00:14:51.049 empire. Yeah, the solo founder example. To entirely 00:14:51.049 --> 00:14:53.909 new computing architectures like China's spiking 00:14:53.909 --> 00:14:56.490 brain 1 .0, challenging the very foundations 00:14:56.490 --> 00:14:58.750 of how we've been building AI. It's a fascinating 00:14:58.750 --> 00:15:02.070 contrast. It really is. The global AI race clearly 00:15:02.070 --> 00:15:04.570 isn't just about features anymore. It's about 00:15:04.570 --> 00:15:08.230 fundamental approaches, efficiency, and increasingly 00:15:08.230 --> 00:15:11.830 national technological sovereignty. It's complex, 00:15:12.070 --> 00:15:15.129 dynamic, and just moving incredibly fast. The 00:15:15.129 --> 00:15:17.210 pace really is immense, isn't it? We're seeing 00:15:17.210 --> 00:15:19.470 breakthroughs almost daily that make you rethink 00:15:19.470 --> 00:15:22.070 what's even possible. Forces you to constantly 00:15:22.070 --> 00:15:24.429 recalibrate. Constantly. And it makes you wonder, 00:15:24.570 --> 00:15:27.389 right, if AI can mimic our own brain's processes 00:15:27.389 --> 00:15:30.889 this efficiently, what new insights might that 00:15:30.889 --> 00:15:33.309 give us into human intelligence itself? That's 00:15:33.309 --> 00:15:35.649 a deep question. What comes next? When we really 00:15:35.649 --> 00:15:37.230 understand how to build intelligence inspired 00:15:37.230 --> 00:15:40.759 by biology, not just simulate it. Lots to think 00:15:40.759 --> 00:15:43.240 about. Indeed. Thanks for diving deep with us 00:15:43.240 --> 00:15:44.440 today on Tiro Music.