WEBVTT 00:00:00.000 --> 00:00:03.480 ChatGPT once looked pretty much invincible, didn't 00:00:03.480 --> 00:00:06.780 it? But now it finds itself, well, surrounded. 00:00:07.219 --> 00:00:11.140 It's this surprising, really intense shift happening 00:00:11.140 --> 00:00:13.720 right now in the AI landscape. We're talking 00:00:13.720 --> 00:00:18.280 new challengers, global power plays, and the 00:00:18.280 --> 00:00:21.199 real AI war. you know often unfolding right there 00:00:21.199 --> 00:00:23.140 on your phone screen welcome to the deep dive 00:00:23.140 --> 00:00:25.620 we're here to unpack the latest insights from 00:00:25.620 --> 00:00:28.480 your sources yeah and today yeah we're plunging 00:00:28.480 --> 00:00:30.879 into this ai frontier looking where the industry 00:00:30.879 --> 00:00:32.740 stands maybe where it's heading and it is moving 00:00:32.740 --> 00:00:35.780 fast we'll uh explore the intense competition 00:00:35.780 --> 00:00:38.920 you know from the tech giants we all know to 00:00:38.920 --> 00:00:41.500 these um stealthy startups you might not have 00:00:41.500 --> 00:00:43.359 even heard of yet And we'll also dig into some 00:00:43.359 --> 00:00:46.100 truly unexpected real world applications. And 00:00:46.100 --> 00:00:48.840 this is crucial, I think, advancements in AI 00:00:48.840 --> 00:00:52.039 safety and reliability. It's such a dynamic space. 00:00:52.960 --> 00:00:55.359 Lots to consider. So, OK, what does this all 00:00:55.359 --> 00:00:57.560 mean for you listening? We'll map out the current 00:00:57.560 --> 00:01:00.939 state of play, see how AI is truly evolving beyond 00:01:00.939 --> 00:01:04.540 just those few big names and what innovations 00:01:04.540 --> 00:01:07.420 might be coming next. So let's unpack this first 00:01:07.420 --> 00:01:11.579 part. The A16's report, that's Andreessen Horowitz, 00:01:11.659 --> 00:01:14.900 the VC firm, their latest report is, well, it's 00:01:14.900 --> 00:01:17.640 a bit of a game changer. It shows ChatGPT's dominant 00:01:17.640 --> 00:01:21.040 lead, which, you know, once felt kind of insurmountable, 00:01:21.079 --> 00:01:23.659 it's actually shrinking. It is. The competition 00:01:23.659 --> 00:01:26.519 is catching up and fast. What's truly fascinating 00:01:26.519 --> 00:01:28.719 here, I think, is just how quickly this landscape 00:01:28.719 --> 00:01:31.439 can shift. Like you blink and things look different. 00:01:31.709 --> 00:01:33.829 Gemini, for example. It's not just a player anymore. 00:01:34.049 --> 00:01:37.670 It's now the number two AI app on mobile. Number 00:01:37.670 --> 00:01:42.090 two. Wow. And it's captured, get this, 90 % of 00:01:42.090 --> 00:01:44.629 the Android market. 90! Think about that penetration. 00:01:44.829 --> 00:01:46.989 It's not just a big slice. It's like practically 00:01:46.989 --> 00:01:49.310 the whole pie on Android. Right. And Grok, you 00:01:49.310 --> 00:01:51.409 know, from x .ai. That isn't just some niche 00:01:51.409 --> 00:01:53.790 thing either. No. It's quickly become a top five 00:01:53.790 --> 00:01:56.709 web AI tool and number 23 on mobile with, what, 00:01:56.730 --> 00:01:59.370 over 20 million users already? Yeah, that's exploding 00:01:59.370 --> 00:02:02.079 growth. I mean, it's fueled by that integration 00:02:02.079 --> 00:02:04.819 into X, right? Access to real -time info. Exactly. 00:02:04.819 --> 00:02:08.400 It really shows the power of platform integration. 00:02:09.080 --> 00:02:12.699 Meanwhile, you see Claude, Anthropix AI, it's 00:02:12.699 --> 00:02:15.319 maintaining steady web growth. Solid performer 00:02:15.319 --> 00:02:19.400 there. Okay. But surprisingly, it's flat on mobile. 00:02:19.639 --> 00:02:21.960 Just hasn't taken off the same way. Interesting. 00:02:21.979 --> 00:02:25.189 And then there's Perplexity AI, kind of. quietly, 00:02:25.289 --> 00:02:28.250 almost stealthily climbing in both web and mobile. 00:02:28.409 --> 00:02:31.030 Ah, yeah, perplexity. Yeah, they're building 00:02:31.030 --> 00:02:34.090 a strong user base by focusing on answer accuracy 00:02:34.090 --> 00:02:35.889 and showing their sources. And then there's this 00:02:35.889 --> 00:02:38.409 global element, which is maybe surprising to 00:02:38.409 --> 00:02:40.949 some. Oh, yeah. This isn't just a Silicon Valley 00:02:40.949 --> 00:02:44.389 game anymore, is it? 22 of the top 50 mobile 00:02:44.389 --> 00:02:48.349 AI apps from China now. 22. That's almost half. 00:02:48.569 --> 00:02:51.270 Dubao from ByteDance, for instance, already number 00:02:51.270 --> 00:02:54.030 four on mobile. Quark from Alibaba is number 00:02:54.030 --> 00:02:56.770 nine on the web. It really signals a truly global 00:02:56.770 --> 00:02:59.810 arms race in AI. You've got diverse players bringing 00:02:59.810 --> 00:03:02.080 different strengths, different approaches. Exactly. 00:03:02.300 --> 00:03:04.960 And we're also seeing new kind of intriguing 00:03:04.960 --> 00:03:08.180 players pop up, like Lovable and Replit, making 00:03:08.180 --> 00:03:10.180 their first appearance on these lists. Lovable, 00:03:10.280 --> 00:03:12.400 that's the vibe coding startup. Yeah, helping 00:03:12.400 --> 00:03:15.319 users create code based on like emotional cues 00:03:15.319 --> 00:03:17.840 or aesthetics, making coding more intuitive. 00:03:18.139 --> 00:03:19.939 And Replit's the collaborative code editor with 00:03:19.939 --> 00:03:22.300 AI built in. Right, speeding up development. 00:03:22.419 --> 00:03:25.379 And others like PixAI and AI Mirror, focusing 00:03:25.379 --> 00:03:27.400 on images. They feel like they're right on the 00:03:27.400 --> 00:03:30.000 edge of breaking through, you know? Mainstream 00:03:30.000 --> 00:03:32.759 adoption feels... close for them so okay thinking 00:03:32.759 --> 00:03:34.620 about all this competition especially on mobile 00:03:34.620 --> 00:03:38.919 yeah what does that intense focus really signify 00:03:38.919 --> 00:03:41.520 for like how we'll interact with AI day to day 00:03:41.520 --> 00:03:44.819 well I think it shows that ubiquitous you know 00:03:44.819 --> 00:03:47.919 personalized AI it's rapidly becoming just an 00:03:47.919 --> 00:03:50.680 indispensable part of our daily digital lives 00:03:50.680 --> 00:03:53.439 mm -hmm yeah makes sense and if we connect this 00:03:53.439 --> 00:03:55.949 to the bigger picture AI isn't just for chatbots 00:03:55.949 --> 00:03:58.810 anymore or just generating text and images. Its 00:03:58.810 --> 00:04:01.770 reliability is being seriously tested now in 00:04:01.770 --> 00:04:05.310 really critical real world areas. The Washington 00:04:05.310 --> 00:04:07.990 Post, for instance, ran this huge test. Oh, yeah, 00:04:08.069 --> 00:04:12.349 I saw that. 900 answers across nine major AI 00:04:12.349 --> 00:04:15.590 tools just to see what actually works, you know, 00:04:15.610 --> 00:04:17.730 not just in theory, but under pressure. That's 00:04:17.730 --> 00:04:20.670 a genuine stress test way beyond the usual benchmarks. 00:04:21.500 --> 00:04:24.180 And speaking of real world impact, remember PJ 00:04:24.180 --> 00:04:27.709 Ace's viral AI ad for Kelshi? Oh, yeah. That 00:04:27.709 --> 00:04:30.009 was impressive. Well, he did it again, crafted 00:04:30.009 --> 00:04:32.649 another super realistic video, this time for 00:04:32.649 --> 00:04:35.689 David Beckham's company, IM8. Wow. These aren't 00:04:35.689 --> 00:04:38.430 just clever tricks anymore. They're showing AI's 00:04:38.430 --> 00:04:41.110 ability to create stuff that looks almost exactly 00:04:41.110 --> 00:04:44.290 like live action video. It's transforming content 00:04:44.290 --> 00:04:47.029 production. And what's maybe equally remarkable 00:04:47.029 --> 00:04:48.769 is the collaboration that's starting to happen 00:04:48.769 --> 00:04:50.930 between these competitors. Like Anthropic and 00:04:50.930 --> 00:04:53.910 OpenAI? Exactly. They just audited each other's 00:04:53.910 --> 00:04:56.519 AI models for safety risks. That's a first of 00:04:56.519 --> 00:04:58.740 its kind move in this industry. That is significant. 00:04:59.060 --> 00:05:00.860 Yeah. It signals maybe a maturing approach to 00:05:00.860 --> 00:05:03.060 development where safety isn't just internal. 00:05:03.199 --> 00:05:07.220 It's a shared, auditable thing. Building collective 00:05:07.220 --> 00:05:10.319 trust, maybe. It absolutely shows a maturing 00:05:10.319 --> 00:05:12.399 industry. I agree. And then in education, Anthropic 00:05:12.399 --> 00:05:16.420 analyzed, what, 74 ,000 educator chats with Claude? 00:05:16.480 --> 00:05:19.519 74 ,000. Yeah. That wasn't just a survey. It 00:05:19.519 --> 00:05:23.199 was a deep dive into how professors are kind 00:05:23.199 --> 00:05:26.670 of. quietly but significantly adapting to AI 00:05:26.670 --> 00:05:31.110 in higher ed. Integrating it into teaching, curriculum, 00:05:31.310 --> 00:05:34.089 often behind the scenes. And on a totally different 00:05:34.089 --> 00:05:37.290 scale, Google DeepMind's experimental weather 00:05:37.290 --> 00:05:41.379 AI. It just... nailed its first real -world test. 00:05:41.500 --> 00:05:43.360 Oh, the hurricane prediction. Yeah, predicted 00:05:43.360 --> 00:05:45.920 Hurricane Erin's path more accurately than the 00:05:45.920 --> 00:05:48.699 traditional methods by, you know, analyzing vast 00:05:48.699 --> 00:05:51.100 amounts of atmospheric data, spotting patterns 00:05:51.100 --> 00:05:53.300 maybe missed before. I mean, imagine scaling 00:05:53.300 --> 00:05:55.660 that. Right. Predicting extreme weather precisely. 00:05:56.269 --> 00:05:58.769 With that accuracy, it could literally save so 00:05:58.769 --> 00:06:01.110 many lives. Think about disaster preparedness. 00:06:01.449 --> 00:06:04.470 Yeah, the global impact is huge. Indeed. And 00:06:04.470 --> 00:06:06.310 on the funding side, we're seeing massive investments 00:06:06.310 --> 00:06:08.689 in these practical applications, too. Field AI 00:06:08.689 --> 00:06:12.730 just raised $405 million. Wow, $405 million. 00:06:12.810 --> 00:06:16.370 What's our goal? Get this. Build one universal 00:06:16.370 --> 00:06:19.709 robot brain. One brain that works across all 00:06:19.709 --> 00:06:22.730 robot types. Ambitious. Very. They're physics 00:06:22.730 --> 00:06:25.050 -based AI. It understands real -world forces, 00:06:25.269 --> 00:06:28.769 movement. It's already reducing risk in energy 00:06:28.769 --> 00:06:31.129 delivery construction. Okay, so when we look 00:06:31.129 --> 00:06:33.689 across all these different uses, weather, robotics, 00:06:33.870 --> 00:06:36.910 safety audits, what's the common thread here 00:06:36.910 --> 00:06:40.339 about... AI's readiness for, you know, real impact? 00:06:40.759 --> 00:06:43.720 I'd say AI is really proving its diverse utility 00:06:43.720 --> 00:06:46.639 now and it's increasing reliability across a 00:06:46.639 --> 00:06:48.600 whole bunch of critical sectors. Here's where 00:06:48.600 --> 00:06:50.319 it gets really interesting, I think, for people 00:06:50.319 --> 00:06:53.519 wanting to actually use AI themselves. Can GPC5, 00:06:53.680 --> 00:06:55.959 you know, one of the newer models, actually create 00:06:55.959 --> 00:06:58.800 profitable trading strategies? Ah, the million 00:06:58.800 --> 00:07:00.620 dollar question. Yeah. Or maybe billion dollar. 00:07:00.699 --> 00:07:02.699 Right. A review on TradingView put it through 00:07:02.699 --> 00:07:05.819 a pretty brutal test, real market data. So we're 00:07:05.819 --> 00:07:07.980 asking where does it shine and where does it 00:07:07.980 --> 00:07:10.600 completely fall flat in finance? Well, this brings 00:07:10.600 --> 00:07:13.079 up such an important question. How do you even 00:07:13.079 --> 00:07:14.980 tell an AI what you want it to do, especially 00:07:14.980 --> 00:07:17.060 when money is involved? Yeah, the prompt matters. 00:07:17.459 --> 00:07:20.279 Exactly. That's prompt engineering. It's the 00:07:20.279 --> 00:07:23.399 art, maybe the science, of crafting really precise, 00:07:23.500 --> 00:07:27.019 clear instructions, like writing a super detailed 00:07:27.019 --> 00:07:30.740 recipe for a chef, you know. Good analogy. It's 00:07:30.740 --> 00:07:33.160 becoming absolutely key to making money with 00:07:33.160 --> 00:07:35.779 AI, to consistently getting something valuable 00:07:35.779 --> 00:07:37.959 out of it. I have to admit, I still wrestle with 00:07:37.959 --> 00:07:40.259 prompt drift myself sometimes, you know, getting 00:07:40.259 --> 00:07:42.279 exactly what I need consistently. It's a learning 00:07:42.279 --> 00:07:44.259 curve for everyone, I think, even us. Oh, for 00:07:44.259 --> 00:07:46.500 sure. But for those looking to start, there's 00:07:46.500 --> 00:07:48.720 actually a seven -day AI business launch plan 00:07:48.720 --> 00:07:51.019 out there now. A full roadmap. Yeah, complete 00:07:51.019 --> 00:07:53.579 step -by -step from just an idea to a launch 00:07:53.579 --> 00:07:55.579 company. Covers market research, deployment, 00:07:55.680 --> 00:07:57.720 the whole thing. We're also seeing these highly 00:07:57.720 --> 00:08:01.040 specialized AI tools emerge, redefining what's 00:08:01.040 --> 00:08:04.120 possible for creators. For example, Gemini 2 00:08:04.120 --> 00:08:06.939 .5 Flash Image. The image generator. Yeah, it 00:08:06.939 --> 00:08:09.339 creates images with incredible detail, down to 00:08:09.339 --> 00:08:12.319 like what they call a nano -banana. A nano -banana. 00:08:12.379 --> 00:08:14.699 Right. This is a SOTA image model, state -of 00:08:14.699 --> 00:08:17.220 -the -art. The absolute cutting edge allows for 00:08:17.220 --> 00:08:20.060 hyper -detailed product stuff or, you know, microscopic 00:08:20.060 --> 00:08:22.740 scientific images. Then there's Codex, which 00:08:22.740 --> 00:08:25.620 is pretty amazing too. Builds web apps in minutes. 00:08:26.490 --> 00:08:29.189 For free. Using like 27 different AI models, 00:08:29.269 --> 00:08:30.910 right? Yeah, handling different parts of development. 00:08:31.069 --> 00:08:34.210 And one part turns content into interactive chat 00:08:34.210 --> 00:08:37.769 pages. No code needed at all. Wow. And Screenshot 00:08:37.769 --> 00:08:40.269 Reports creates branded reports just from screenshots, 00:08:40.649 --> 00:08:42.309 streamlining something that's usually pretty 00:08:42.309 --> 00:08:45.289 tedious. Okay, so thinking about all these tools, 00:08:45.730 --> 00:08:48.710 prompt engineering, no code builders, specialized 00:08:48.710 --> 00:08:52.029 models, what's the fundamental shift happening 00:08:52.029 --> 00:08:54.600 for creators and entrepreneurs here? I think 00:08:54.600 --> 00:08:57.519 AI is really democratizing innovation. It's allowing 00:08:57.519 --> 00:08:59.799 pretty much anyone to build and create, often 00:08:59.799 --> 00:09:02.980 with absolutely no code required. So, okay, how 00:09:02.980 --> 00:09:06.220 do we make AI feel more human, more relatable 00:09:06.220 --> 00:09:08.200 when we interact with it? There's this viral 00:09:08.200 --> 00:09:10.620 system prompt for ChatGPT going around. Oh, yeah. 00:09:10.720 --> 00:09:12.860 What does it do? It basically teaches the AI 00:09:12.860 --> 00:09:15.840 to speak more like a real person, injects conversational 00:09:15.840 --> 00:09:18.980 nuances, idioms, even a bit of personality. Makes 00:09:18.980 --> 00:09:21.399 it feel less robotic. Huh. Interesting. Like 00:09:21.399 --> 00:09:24.139 giving a conversational coaching. Exactly. And 00:09:24.139 --> 00:09:26.279 on the video side, things are moving fast too. 00:09:26.700 --> 00:09:30.500 Alibaba just updated its video AI, film quality 00:09:30.500 --> 00:09:32.559 avatars, they're calling them. Making AI characters 00:09:32.559 --> 00:09:35.139 look almost real. Stunningly real, yeah. Almost 00:09:35.139 --> 00:09:37.860 indistinguishable from live actors. And Google 00:09:37.860 --> 00:09:40.960 Workspace now has a free AI video tool called 00:09:40.960 --> 00:09:44.220 Google Vids. Puts easy video creation into more 00:09:44.220 --> 00:09:46.840 hands. But AI isn't all positive, right? There 00:09:46.840 --> 00:09:49.340 are challenges. It's a pretty stark one. Definitely. 00:09:49.639 --> 00:09:52.480 Stanford University recently found AI actually 00:09:52.480 --> 00:09:56.100 killed 13 percent of jobs for 22 to 25 year olds, 00:09:56.220 --> 00:09:58.940 specifically in coding and support roles. 13 00:09:58.940 --> 00:10:01.600 percent. That's significant. It's a stark reminder 00:10:01.600 --> 00:10:04.100 of the disruptive power affecting real people, 00:10:04.360 --> 00:10:07.480 highlighting that need for workforce adaptation. 00:10:08.000 --> 00:10:09.799 And sadly, we've seen a darker side already. 00:10:10.019 --> 00:10:12.980 A hacker recently used Claude Anthropics AI to 00:10:12.980 --> 00:10:15.519 run a full scale cybercrime operation. Seriously? 00:10:15.779 --> 00:10:18.590 God. It hits 17 different organizations. It just 00:10:18.590 --> 00:10:20.929 clearly shows the critical need for robust security, 00:10:21.129 --> 00:10:23.610 ethical safeguards, preventing these powerful 00:10:23.610 --> 00:10:25.850 tools from being weaponized. Which really raises 00:10:25.850 --> 00:10:29.549 the question, how do we properly test these incredibly 00:10:29.549 --> 00:10:32.570 powerful AI agents, test them in a way that reflects 00:10:32.570 --> 00:10:35.110 real world performance? Well, AI2, the Allen 00:10:35.110 --> 00:10:37.950 Institute for AI. They just dropped AstaBench. 00:10:38.029 --> 00:10:40.129 AstaBench. Okay, what is it? It's a new benchmark 00:10:40.129 --> 00:10:42.629 suite. But a really rigorous one. It's built 00:10:42.629 --> 00:10:46.149 specifically to test scientific AI agents. Uses 00:10:46.149 --> 00:10:49.789 over 2 ,400 real research tasks. Wow, 2 ,400 00:10:49.789 --> 00:10:52.350 real tasks. That's way beyond simple chatbot 00:10:52.350 --> 00:10:55.070 stuff. Oh, yeah. And AstaBench focuses on what 00:10:55.070 --> 00:10:57.330 they call five brutal truths about agent testing. 00:10:57.509 --> 00:10:59.990 Okay, like what? Well, first, tasks have to be 00:10:59.990 --> 00:11:03.039 real world and genuinely hard. Second, tool use 00:11:03.039 --> 00:11:05.639 needs careful control. Third, you've got to track 00:11:05.639 --> 00:11:07.860 the cost accuracies and everything if it's bankrupting 00:11:07.860 --> 00:11:10.480 you. Fourth, standardization is key for comparing 00:11:10.480 --> 00:11:12.820 apples to apples. And fifth, you can't claim 00:11:12.820 --> 00:11:15.000 progress unless you know exactly what you've 00:11:15.000 --> 00:11:17.080 actually beaten. Really practical fundamental 00:11:17.080 --> 00:11:20.000 points. And they tested a lot of agents. 57 agents 00:11:20.000 --> 00:11:22.679 across 22 different architectures. Huge sample. 00:11:22.860 --> 00:11:27.299 Their own agent, Asta V0, scored 53%. React plus 00:11:27.299 --> 00:11:30.919 G5 got 43 .3%. Okay, so still room for improvement. 00:11:31.389 --> 00:11:34.110 Definitely. And interestingly, data analysis, 00:11:34.350 --> 00:11:36.750 still pretty weak across the board. No one scored 00:11:36.750 --> 00:11:40.389 above 34 % there. That's a major area needing 00:11:40.389 --> 00:11:43.470 work. And what about GPT -5's impact in those 00:11:43.470 --> 00:11:46.350 tests? Also interesting. It actually helped general 00:11:46.350 --> 00:11:48.610 agents, like the ones using that React framework 00:11:48.610 --> 00:11:52.850 for reasoning and action. But it oddly hurt specialized 00:11:52.850 --> 00:11:56.039 ones. which suggests maybe it's tuned specifically 00:11:56.039 --> 00:11:57.980 for certain workflows like those React -style 00:11:57.980 --> 00:12:00.440 operations. This is definitely the benchmark 00:12:00.440 --> 00:12:02.659 to watch if you're building AI for critical fields, 00:12:02.779 --> 00:12:05.840 research, medicine, engineering, where trust 00:12:05.840 --> 00:12:08.059 and reproducibility are everything. So boiling 00:12:08.059 --> 00:12:10.279 it down, what's the biggest takeaway from Ask 00:12:10.279 --> 00:12:12.139 the Bench for you and maybe for anyone relying 00:12:12.139 --> 00:12:15.100 on AI for really important tasks? I think it's 00:12:15.100 --> 00:12:18.139 that rigorous, truly real -world benchmarks are 00:12:18.139 --> 00:12:20.059 absolutely essential if we're going to build 00:12:20.059 --> 00:12:23.320 trustworthy and reliable AI, period. So if we 00:12:23.320 --> 00:12:25.100 try to connect all of this, tie it together, 00:12:25.320 --> 00:12:27.480 three things really stand out from today's deep 00:12:27.480 --> 00:12:30.700 dive, I think. First, the AI competition is just 00:12:30.700 --> 00:12:33.000 incredibly intense now, and it's truly global. 00:12:33.320 --> 00:12:35.779 Chad GPT isn't alone at the top anymore. And 00:12:35.779 --> 00:12:39.440 that battleground for dominance, it's decisively 00:12:39.440 --> 00:12:42.379 shifted to mobile devices. That's where the action 00:12:42.379 --> 00:12:45.320 is. Second, AI's practical applications are just 00:12:45.320 --> 00:12:47.620 expanding so rapidly. They're touching almost 00:12:47.620 --> 00:12:49.759 every part of our lives in really tangible ways 00:12:49.759 --> 00:12:53.220 now. From predicting hurricanes, transforming 00:12:53.220 --> 00:12:55.399 education, providing these powerful business 00:12:55.399 --> 00:12:57.320 tools. It's integrated. It's everywhere. It's 00:12:57.320 --> 00:12:59.879 not just niche tech anymore. Exactly. And third, 00:13:00.000 --> 00:13:03.019 maybe most importantly, the need for robust testing, 00:13:03.220 --> 00:13:06.419 industry -wide safety audits, clear ethical guidelines. 00:13:06.759 --> 00:13:09.460 Yeah. It's more critical than ever before. Yeah. 00:13:09.789 --> 00:13:12.669 As AI gets more powerful, more deeply integrated 00:13:12.669 --> 00:13:15.769 into our lives, we just need smarter ways to 00:13:15.769 --> 00:13:19.350 ensure it works reliably, safely, and responsibly 00:13:19.350 --> 00:13:22.440 for everyone. So the big question then, as AI 00:13:22.440 --> 00:13:24.320 keeps evolving so quickly, how do we make sure 00:13:24.320 --> 00:13:26.799 we're building trustworthy, responsible systems? 00:13:27.120 --> 00:13:29.200 Systems that actually benefit humanity rather 00:13:29.200 --> 00:13:32.159 than introducing new risks. And maybe related. 00:13:32.720 --> 00:13:35.000 What's your role in understanding this rapidly 00:13:35.000 --> 00:13:37.519 changing landscape? How will you adapt your own 00:13:37.519 --> 00:13:39.659 skills, your own expectations as this continues? 00:13:39.940 --> 00:13:42.059 We really encourage you to keep exploring, keep 00:13:42.059 --> 00:13:45.080 asking questions. This deep dive into your sources 00:13:45.080 --> 00:13:47.879 clearly shows the future of AI while it's still 00:13:47.879 --> 00:13:50.269 being written. And it's a story we're all a part 00:13:50.269 --> 00:13:52.850 of, really. Thank you for joining us on this 00:13:52.850 --> 00:13:55.070 deep dive. We'll be back soon with more insights. 00:13:55.490 --> 00:13:57.889 Until then, keep digging. Out Hero Music.