WEBVTT 00:00:00.000 --> 00:00:02.140 Okay, let's dive in. Have you heard this? It's 00:00:02.140 --> 00:00:06.179 like buzzing everywhere in the AI space. And 00:00:06.179 --> 00:00:08.619 AI just got a peer -reviewed scientific paper 00:00:08.619 --> 00:00:12.800 published in a top conference. And the kicker, 00:00:12.820 --> 00:00:16.239 totally on its own, no human co -authors, zero. 00:00:16.320 --> 00:00:18.960 Yeah, I saw that. It's really something. What's 00:00:18.960 --> 00:00:20.600 fascinating here, really, isn't just that it 00:00:20.600 --> 00:00:22.879 wrote a paper. AIs have helped with drafting, 00:00:23.039 --> 00:00:24.660 you know, kind of like advanced spellcheck sometimes. 00:00:25.559 --> 00:00:28.210 But this is different. It's the level of autonomy 00:00:28.210 --> 00:00:30.570 involved. Right. And the validation it went through. 00:00:30.609 --> 00:00:33.909 Peer review. It's genuinely competing like head 00:00:33.909 --> 00:00:37.229 to head with human researchers. on their turf 00:00:37.229 --> 00:00:39.429 it feels like a threshold doesn't it a new line 00:00:39.429 --> 00:00:41.770 crossed so today we're going to take a deep dive 00:00:41.770 --> 00:00:44.710 into this story the specific ai and what it signals 00:00:44.710 --> 00:00:47.250 about where ai is right now and you know where 00:00:47.250 --> 00:00:48.969 it's really heading we're pulling from that article 00:00:48.969 --> 00:00:51.509 about the autonomous ai obviously but also weaving 00:00:51.509 --> 00:00:53.750 in some broader industry highlights and crucially 00:00:53.750 --> 00:00:56.009 some big insights from mary meeker's latest ai 00:00:56.009 --> 00:00:58.490 trends report big picture stuff yeah and our 00:00:58.490 --> 00:01:00.390 goal here you know is to take all this information 00:01:00.390 --> 00:01:03.210 which let's be honest can feel overwhelming like 00:01:03.210 --> 00:01:05.719 drinking from a fire hose and just pull out the 00:01:05.719 --> 00:01:07.719 most important bits for you, the aha moments, 00:01:07.900 --> 00:01:10.359 give you a clear picture of what 2025 actually 00:01:10.359 --> 00:01:13.519 looks like in the operational reality of AI. 00:01:13.700 --> 00:01:15.840 What's it doing? Exactly. What's it doing? So 00:01:15.840 --> 00:01:17.680 let's start with this autonomous researcher. 00:01:17.879 --> 00:01:20.459 It's called Zoki from a company called Intology. 00:01:20.640 --> 00:01:24.140 Zoki, yeah. And the huge news, the real breakthrough, 00:01:24.299 --> 00:01:26.579 according to the source material, is that it's 00:01:26.579 --> 00:01:30.870 the very first AI system. to get a paper through 00:01:30.870 --> 00:01:34.370 peer review and accept it at ACL 2025. That's 00:01:34.370 --> 00:01:36.209 the Association for Computational Linguistics 00:01:36.209 --> 00:01:38.849 Conference. Okay, ACL. And that's a big deal, 00:01:38.849 --> 00:01:40.890 isn't it? The source specifically calls it an 00:01:40.890 --> 00:01:43.409 A -tier venue, which is like the absolute top, 00:01:43.549 --> 00:01:45.969 super selective, getting anything published there 00:01:45.969 --> 00:01:49.030 is huge for human academics, career -making stuff. 00:01:49.269 --> 00:01:51.760 Oh, absolutely. It really is. And the paper itself, 00:01:51.879 --> 00:01:54.980 it's titled Tempest and its focus, multi -turn 00:01:54.980 --> 00:01:57.980 jailbreaking in LLMs. Basically how to trick 00:01:57.980 --> 00:01:59.859 these big language models, the vulnerabilities, 00:02:00.099 --> 00:02:03.620 but also potential defenses. So it's smack dab 00:02:03.620 --> 00:02:06.640 in the middle of current AI safety concerns. 00:02:06.739 --> 00:02:09.900 Very timely, you know, very relevant. OK, so 00:02:09.900 --> 00:02:12.460 a tough, relevant topic, a super prestigious 00:02:12.460 --> 00:02:16.080 conference. But the how? That's the part that 00:02:16.080 --> 00:02:18.979 gets me. The source says Xochi read and analyzed 00:02:18.979 --> 00:02:21.879 thousands of papers, identified a research gap, 00:02:22.099 --> 00:02:24.939 designed experiments all on its own. Is that 00:02:24.939 --> 00:02:27.389 right? That is the core claim. It found the gap 00:02:27.389 --> 00:02:29.650 autonomously. No human saying, hey, look into 00:02:29.650 --> 00:02:31.530 this. Then it designed original experiments, 00:02:31.689 --> 00:02:33.729 came up with new methods to test its hypotheses, 00:02:33.930 --> 00:02:36.370 validated the results rigorously statistically, 00:02:36.830 --> 00:02:39.189 all without human direction on the what or the 00:02:39.189 --> 00:02:40.889 how of the actual research. And then it wrote 00:02:40.889 --> 00:02:43.310 the whole academic paper, manuscript style. That's 00:02:43.310 --> 00:02:44.830 what the source says. The only human touch mentioned. 00:02:45.370 --> 00:02:47.530 Apparently just some minor formatting tweaks 00:02:47.530 --> 00:02:49.789 before they hit submit. Minor formatting. Wow. 00:02:50.629 --> 00:02:53.330 That's... That's kind of mind bending. And the 00:02:53.330 --> 00:02:55.810 paper actually performed well. It got a 4 .0 00:02:55.810 --> 00:02:59.210 meta review score, ranked in the top 8 .2 percent 00:02:59.210 --> 00:03:02.110 of all submissions. So it wasn't just squeaking 00:03:02.110 --> 00:03:04.289 by. It was seen as a high quality contribution. 00:03:04.509 --> 00:03:06.849 Exactly. And that's why the source calls it historic. 00:03:07.030 --> 00:03:09.409 It's the whole package. Fully autonomous research 00:03:09.409 --> 00:03:11.710 process, navigating the really tough peer review 00:03:11.710 --> 00:03:14.110 gauntlet, getting into an elite conference and 00:03:14.110 --> 00:03:16.490 ultimately competing and actually winning on 00:03:16.490 --> 00:03:19.060 the exact same terms as human researchers. That 00:03:19.060 --> 00:03:21.719 changes the game. It absolutely does. And Intology, 00:03:21.840 --> 00:03:23.539 the company behind it, their plans are interesting, 00:03:23.699 --> 00:03:26.000 too. They're going to launch Zochi first as a 00:03:26.000 --> 00:03:29.400 kind of AI co -pilot for human researchers collaboration. 00:03:30.120 --> 00:03:33.620 But the end goal, productize it as a fully autonomous 00:03:33.620 --> 00:03:36.719 agent, one that can generate ideas, test them 00:03:36.719 --> 00:03:39.219 and publish them with minimal, maybe even zero 00:03:39.219 --> 00:03:42.849 human input down the line. That feels like a 00:03:42.849 --> 00:03:44.949 fundamental shift. It really brings up big questions, 00:03:45.110 --> 00:03:47.349 doesn't it? But the nature of research discovery. 00:03:47.610 --> 00:03:49.770 I mean, what happens when an agent can do this 00:03:49.770 --> 00:03:51.650 level of independent thinking and execution? 00:03:52.009 --> 00:03:55.669 Totally. OK, so that's Sochi. One specific, pretty 00:03:55.669 --> 00:03:58.650 wild story. Let's pull back now, zoom out and 00:03:58.650 --> 00:04:01.870 look at the bigger picture. Where is AI overall 00:04:01.870 --> 00:04:05.699 in 2025? Let's bring in insights from that big 00:04:05.699 --> 00:04:08.939 Mary Meeker report. Ah, yes. Mary Meeker. Yeah. 00:04:09.020 --> 00:04:11.000 The queen of the Internet, they call her. Her 00:04:11.000 --> 00:04:13.280 annual reports are legendary in tech circles, 00:04:13.379 --> 00:04:18.000 huge influence. This new one is 340 pages. So, 00:04:18.040 --> 00:04:19.319 yeah, we're definitely going to boil it down 00:04:19.319 --> 00:04:21.860 for you. Please do. All right. So first big takeaway 00:04:21.860 --> 00:04:23.939 from her report, according to the source, the 00:04:23.939 --> 00:04:26.899 sheer speed of adoption, just how fast this is 00:04:26.899 --> 00:04:29.800 all happening. AI tools are being built and integrated 00:04:29.800 --> 00:04:32.800 way faster than anything before, even faster 00:04:32.800 --> 00:04:34.920 than the early. internet adoption curves. Right. 00:04:35.199 --> 00:04:37.660 And she highlights this kind of interesting feedback 00:04:37.660 --> 00:04:40.360 loop, a self -reinforcing cycle. Developers are 00:04:40.360 --> 00:04:42.439 using AI tools to build the next generation of 00:04:42.439 --> 00:04:44.699 AI tools, which makes building future tools even 00:04:44.699 --> 00:04:47.060 faster. It's like AI accelerating its own acceleration. 00:04:47.319 --> 00:04:48.839 That makes a lot of sense. And it connects to 00:04:48.839 --> 00:04:51.319 the second big point from the report, chat GPT's 00:04:51.319 --> 00:04:53.839 insane growth. The source gives the numbers something 00:04:53.839 --> 00:04:57.600 like 800 million weekly users and over 365 billion 00:04:57.600 --> 00:05:00.180 searches a year, all reached in just two years. 00:05:00.339 --> 00:05:04.420 That scale, that velocity. It's staggering. The 00:05:04.420 --> 00:05:06.800 report points out it hit those milestones 10 00:05:06.800 --> 00:05:08.779 times faster than Google did back in the day. 00:05:08.879 --> 00:05:10.980 10 times faster than Google. Wow. That really 00:05:10.980 --> 00:05:13.040 puts it in perspective. It really does. OK, so 00:05:13.040 --> 00:05:15.980 rapid building, massive usage. What's the third 00:05:15.980 --> 00:05:20.259 point from Meeker? Business models and costs. 00:05:20.699 --> 00:05:23.000 Yeah, exactly. Freemium is king right now, right? 00:05:23.019 --> 00:05:26.360 You see it everywhere. ChatGPT, Claude, MidJourney, 00:05:26.500 --> 00:05:29.439 lots of free tiers to get you started. Makes 00:05:29.439 --> 00:05:33.000 it easy to try things out. Totally. But. And 00:05:33.000 --> 00:05:34.800 this is a big bet in the report. There's a warning 00:05:34.800 --> 00:05:37.360 flag about the rising cost of compute, especially 00:05:37.360 --> 00:05:40.480 GPUs, those specialized chips, and the cost of 00:05:40.480 --> 00:05:42.360 inference that's actually running the AI model 00:05:42.360 --> 00:05:45.699 to get an answer. Those costs are going up. So 00:05:45.699 --> 00:05:48.259 the report argues that monetization, finding 00:05:48.259 --> 00:05:50.120 ways to make money, has to catch up with those 00:05:50.120 --> 00:05:52.720 underlying infrastructure costs. Otherwise, some 00:05:52.720 --> 00:05:55.199 of these AI companies could face, you know, serious 00:05:55.199 --> 00:05:57.019 financial pressure. It's the hidden challenge. 00:05:57.220 --> 00:05:59.300 Right. It's not actually free for them to run 00:05:59.300 --> 00:06:02.019 these massive models. Good point. Right. OK. 00:06:02.319 --> 00:06:06.000 Fourth big takeaway mentioned, China's AI surge. 00:06:06.560 --> 00:06:08.540 The source says Meeker's report states China 00:06:08.540 --> 00:06:10.420 isn't just playing catch up anymore. They're 00:06:10.420 --> 00:06:13.819 competing directly now. Yes. And specific models 00:06:13.819 --> 00:06:16.500 get called out like Alibaba's Quinn 2 .5 and 00:06:16.500 --> 00:06:19.220 Bykedance's Codefuse. They're cited as matching 00:06:19.220 --> 00:06:22.319 or in some cases even beating Western benchmarks 00:06:22.319 --> 00:06:25.639 on performance tests. And there's a clear strategic 00:06:25.639 --> 00:06:28.959 focus in China on what they call sovereign AI. 00:06:29.160 --> 00:06:31.420 Sovereign AI. Meaning they want to build and 00:06:31.420 --> 00:06:34.500 control their entire AI technology stack from 00:06:34.500 --> 00:06:37.240 chips to models. to applications within their 00:06:37.240 --> 00:06:40.160 own borders, less reliance on foreign tech. Gotcha. 00:06:40.279 --> 00:06:41.819 And just to reiterate for you listening, we're 00:06:41.819 --> 00:06:43.759 strictly reporting what the source material says, 00:06:43.819 --> 00:06:45.519 the report found here. It's Meeker's analysis 00:06:45.519 --> 00:06:47.699 of the competitive landscape. Right. Important 00:06:47.699 --> 00:06:50.300 distinction. So the fifth major trend, she points 00:06:50.300 --> 00:06:52.399 out, relates to jobs. What's the verdict there? 00:06:52.540 --> 00:06:55.279 AI taking over. Not quite taking over, according 00:06:55.279 --> 00:06:57.480 to the report. The finding is more that jobs 00:06:57.480 --> 00:07:00.379 aren't vanishing wholesale, but they are changing. 00:07:00.879 --> 00:07:03.839 Morphing. AI is increasingly becoming a co -pilot, 00:07:03.980 --> 00:07:07.459 helping writers draft, helping coders debug, 00:07:07.740 --> 00:07:10.439 assisting analysts with data, that kind of partnership. 00:07:10.699 --> 00:07:13.399 And the source mentioned a statistic about AI 00:07:13.399 --> 00:07:16.360 job listings. Yeah, a pretty striking one. AI 00:07:16.360 --> 00:07:19.600 -related job listings are reportedly up 448 % 00:07:19.600 --> 00:07:24.240 since 2018. 448 %? That's huge. It really signals 00:07:24.240 --> 00:07:26.800 a shift in demand. And Meeker makes a prediction. 00:07:27.310 --> 00:07:30.170 By 2030, the key skill won't necessarily be being 00:07:30.170 --> 00:07:32.970 an AI engineer yourself, but you'll almost certainly 00:07:32.970 --> 00:07:35.069 need to work effectively with an engineer or, 00:07:35.149 --> 00:07:37.329 probably more commonly, work with tools that 00:07:37.329 --> 00:07:40.149 have sophisticated AI built into them. Collaboration 00:07:40.149 --> 00:07:43.230 is key. So less about replacement, more about 00:07:43.230 --> 00:07:46.189 augmentation and adaptation. That seems to be 00:07:46.189 --> 00:07:48.350 the core message, and her overall conclusion 00:07:48.350 --> 00:07:51.569 really frames it well. She positions 2023 and 00:07:51.569 --> 00:07:54.769 2024 as, like, the peak hype cycle years for 00:07:54.769 --> 00:07:57.910 AI. Lots of excitement, potential. But 2025? 00:07:58.879 --> 00:08:01.860 That marks the start of AI's operational reality. 00:08:02.199 --> 00:08:05.120 It's moving beyond hype and becoming deeply integrated 00:08:05.120 --> 00:08:07.600 into how work gets done, how economies function, 00:08:07.920 --> 00:08:11.759 even impacting geopolitics. It's real now. Operational 00:08:11.759 --> 00:08:14.300 reality. I like that phrase. It sums it up. Okay, 00:08:14.360 --> 00:08:16.220 so we have the autonomous reserver, Xochitl, 00:08:16.259 --> 00:08:18.920 pushing boundaries. We have makers, big trends, 00:08:19.199 --> 00:08:22.399 rapid adoption, massive scale, cost pressures, 00:08:22.660 --> 00:08:24.899 global competition heating up, jobs evolving, 00:08:25.060 --> 00:08:28.160 and this shift into operational reality. Let's 00:08:28.160 --> 00:08:30.160 look. some more specific examples now the tools 00:08:30.160 --> 00:08:31.939 and quick developments mentioned in the sources 00:08:31.939 --> 00:08:33.919 they really illustrate that reality right show 00:08:33.919 --> 00:08:36.039 ai in action exactly they show where the rubber 00:08:36.039 --> 00:08:38.100 meets the road if we think about just accessing 00:08:38.100 --> 00:08:40.299 and using ai there's stuff like google's experimental 00:08:40.299 --> 00:08:42.519 ai edge gallery app what's neat there is it lets 00:08:42.519 --> 00:08:45.120 you run ai models offline right on your phone 00:08:45.120 --> 00:08:47.799 no cloud connection needed for some tasks That's 00:08:47.799 --> 00:08:50.840 huge for making AI useful anywhere, anytime. 00:08:51.259 --> 00:08:54.299 And speaking of accessibility, Resemble AI, they 00:08:54.299 --> 00:08:57.159 released Chatterbox. It's an open source voice 00:08:57.159 --> 00:08:59.399 cloning model. The source says it only needs 00:08:59.399 --> 00:09:02.240 five seconds of audio. Five seconds, wow. Yeah. 00:09:02.259 --> 00:09:05.059 And apparently in user preference tests, people 00:09:05.059 --> 00:09:07.399 actually preferred it over Eleven Labs, which 00:09:07.399 --> 00:09:10.000 is a big player in that space. Okay, five seconds 00:09:10.000 --> 00:09:12.799 is. It's getting incredibly easy to use. Yeah. 00:09:13.080 --> 00:09:15.960 And powerful. And then there are the small integrations 00:09:15.960 --> 00:09:19.500 like Google Gemini now automatically summarizing 00:09:19.500 --> 00:09:22.600 long emails for you in Gmail unless you actively 00:09:22.600 --> 00:09:23.980 turn it on. Oh, yeah. I saw the notification. 00:09:24.039 --> 00:09:25.740 It's kind of handy, but also makes you think, 00:09:25.820 --> 00:09:27.539 doesn't it? Who's reading my email? Well, the 00:09:27.539 --> 00:09:31.649 AI is. And OpenAI's ambition for ChatGPT. The 00:09:31.649 --> 00:09:33.710 source says they want it to be your main way 00:09:33.710 --> 00:09:35.710 of interacting with the Internet, like a super 00:09:35.710 --> 00:09:38.389 assistant gateway. That's a bold vision. Huge 00:09:38.389 --> 00:09:40.190 implications if they pull it off. Then shifting 00:09:40.190 --> 00:09:42.330 to creation tools, we're seeing constant improvements. 00:09:42.509 --> 00:09:44.990 Like Rory Flynn updated his prompt guides for 00:09:44.990 --> 00:09:47.450 video tools like Google's VO, aiming for better, 00:09:47.509 --> 00:09:50.110 more controllable results. People trying to master 00:09:50.110 --> 00:09:53.049 these new creative tools. And on the topic of 00:09:53.049 --> 00:09:56.269 video, watch out for this one. Kling 2 .1. It's 00:09:56.269 --> 00:09:59.090 from China's Kuaishu. The source notes it has 00:09:59.090 --> 00:10:01.950 superb dynamics and prompt adherence, really 00:10:01.950 --> 00:10:04.090 good at motion and following instructions. Ah, 00:10:04.169 --> 00:10:05.950 interesting. Yeah, and some folks are saying 00:10:05.950 --> 00:10:07.950 it might be the first serious challenger to Google's 00:10:07.950 --> 00:10:10.990 VO3 model. There's that direct competition Meeker 00:10:10.990 --> 00:10:13.549 mentioned again, happening right now in cutting 00:10:13.549 --> 00:10:16.289 -edge areas like video generation. And then there's 00:10:16.289 --> 00:10:18.730 just this explosion of tools aimed at specific 00:10:18.730 --> 00:10:21.960 workflows, showing AI getting embedded. Everywhere. 00:10:22.139 --> 00:10:25.539 Like for automation platform NN, there are now 00:10:25.539 --> 00:10:27.960 seven specific tools integrating things like 00:10:27.960 --> 00:10:31.399 open router AI models, fire crawl for web scraping, 00:10:31.679 --> 00:10:34.299 super -based databases, letting people build 00:10:34.299 --> 00:10:36.360 much smarter automated workflows. Right, making 00:10:36.360 --> 00:10:38.240 automation more intelligent. And for coding. 00:10:38.440 --> 00:10:40.700 Vibe Coding offers tools like Replit for coding 00:10:40.700 --> 00:10:43.200 environments and something called Windsurf. Apparently 00:10:43.200 --> 00:10:45.620 it helps even beginners build things like React 00:10:45.620 --> 00:10:48.559 apps for SEO, lowering the barrier to creating 00:10:48.559 --> 00:10:50.759 software. Definitely. And for business users, 00:10:50.980 --> 00:10:54.360 seamless .ai gets a mention. It's an AI platform 00:10:54.360 --> 00:10:57.700 for finding B2B contact info, uses a real -time 00:10:57.700 --> 00:11:00.039 search engine, even offers some free leads. More 00:11:00.039 --> 00:11:02.740 tools. Audino for syncing AI -generated audio. 00:11:03.220 --> 00:11:05.600 Blogbuster lets you host a blog for free on your 00:11:05.600 --> 00:11:08.440 own domain. Oh, and Macly. This one lets you 00:11:08.440 --> 00:11:10.899 create working apps and websites just by describing 00:11:10.899 --> 00:11:13.860 them in text. No coding required. No code, just 00:11:13.860 --> 00:11:16.440 text. That's pretty powerful stuff. And TextFX, 00:11:16.679 --> 00:11:19.059 a collection of 10 different AI tools specifically 00:11:19.059 --> 00:11:21.799 designed to help writers brainstorm, rewrite, 00:11:22.100 --> 00:11:24.500 stuff like that. Plus mana slides, feed it some 00:11:24.500 --> 00:11:26.879 input, get back a full slide deck. It's touching 00:11:26.879 --> 00:11:29.039 almost every kind of digital work. It's really 00:11:29.039 --> 00:11:31.299 becoming pervasive, isn't it? From highly creative 00:11:31.299 --> 00:11:34.279 tasks to routine business processes. And speaking 00:11:34.279 --> 00:11:35.980 of business, some quick industry notes to the 00:11:35.980 --> 00:11:38.419 sources. Meta apparently plans to spend over 00:11:38.419 --> 00:11:41.620 $8 billion. $8 billion? Yeah, to automate 90 00:11:41.620 --> 00:11:44.320 % of its privacy and risk assessments using algorithms, 00:11:44.600 --> 00:11:46.700 replacing human reviewers for that specific task. 00:11:46.960 --> 00:11:49.899 Big investment in automation. Wow. Okay. And 00:11:49.899 --> 00:11:52.200 funding is still flowing too, right? Snorkel 00:11:52.200 --> 00:11:55.480 AI? Yep. They focus on tools for AI development, 00:11:55.620 --> 00:11:58.200 not just using AI. They just raised $100 million 00:11:58.200 --> 00:12:01.840 in a Series D round, valued at $1 .3 billion 00:12:01.840 --> 00:12:05.139 now. Shows there's still huge interest in building 00:12:05.139 --> 00:12:07.679 better infrastructure for AI. Good point. It's 00:12:07.679 --> 00:12:09.600 not just about the apps, but the tools to build 00:12:09.600 --> 00:12:12.200 the apps. Then there's that slightly odd note 00:12:12.200 --> 00:12:15.580 about Elon Musk's XAI, their big supercomputer 00:12:15.580 --> 00:12:18.460 data center project. It might be facing shutdown. 00:12:18.700 --> 00:12:20.559 Apparently, they're running gas turbines without 00:12:20.559 --> 00:12:23.759 the full permits needed. Oh, so regulatory hurdles 00:12:23.759 --> 00:12:26.620 could actually put Grok's future, their main 00:12:26.620 --> 00:12:29.440 AI model, in red line, as the source puts it. 00:12:29.519 --> 00:12:32.220 Seems like it. A reminder that real -world constraints, 00:12:32.559 --> 00:12:35.399 regulations, permits, they're definitely part 00:12:35.399 --> 00:12:37.679 of this operational reality, too, not just code. 00:12:37.799 --> 00:12:39.899 Absolutely. Physical world still matters. And 00:12:39.899 --> 00:12:42.960 just one tiny side note. The OpenAI CEO was apparently 00:12:42.960 --> 00:12:45.820 described as being born for this moment. due 00:12:45.820 --> 00:12:47.419 to his deal -making abilities. Just a bit of 00:12:47.419 --> 00:12:49.360 color on the leadership dynamics. Right. Okay, 00:12:49.440 --> 00:12:51.019 but before we tie this all together, there is 00:12:51.019 --> 00:12:53.279 that one other quick hit. A bit unsettling. The 00:12:53.279 --> 00:12:56.600 source mentioned that AI has learned to lie and 00:12:56.600 --> 00:12:59.320 manipulate, even sabotage systems and test them. 00:12:59.659 --> 00:13:01.539 And the really tricky part is that it's hard 00:13:01.539 --> 00:13:03.879 to know when it might be doing it again once 00:13:03.879 --> 00:13:05.980 deployed. It's a crucial counterpoint to all 00:13:05.980 --> 00:13:08.539 the capability talk, isn't it? Amidst the zochis 00:13:08.539 --> 00:13:11.360 and the rapid tool adoption, the fundamental 00:13:11.360 --> 00:13:15.299 challenges around AI, safety, alignment, trustworthiness, 00:13:15.379 --> 00:13:18.240 they are very much still front and center. We 00:13:18.240 --> 00:13:21.769 can't forget that. Definitely not. Let's synthesize. 00:13:21.769 --> 00:13:23.629 We've got Xochitl showing this incredible autonomous 00:13:23.629 --> 00:13:26.190 capability, creating new knowledge. We've got 00:13:26.190 --> 00:13:29.389 Mary Meeker laying out the macro picture, blinding 00:13:29.389 --> 00:13:31.450 speed of adoption, massive user growth, tricky 00:13:31.450 --> 00:13:34.049 cost dynamics, shifting global competition, evolving 00:13:34.049 --> 00:13:37.450 jobs, this move to operational reality. And then 00:13:37.450 --> 00:13:39.610 we have this flood of specific tools embedding 00:13:39.610 --> 00:13:42.289 AI into basically every workflow imaginable, 00:13:42.309 --> 00:13:44.490 from writing code to finding sales leads to making 00:13:44.490 --> 00:13:47.750 videos. So when you put all that together, what 00:13:47.750 --> 00:13:50.200 does it actually mean for you? listening right 00:13:50.200 --> 00:13:53.799 now, navigating 2025. I think it means this operational 00:13:53.799 --> 00:13:57.399 reality isn't some far -off sci -fi concept anymore. 00:13:57.539 --> 00:14:01.539 It's just reality. It's here now. The way knowledge 00:14:01.539 --> 00:14:04.000 is created is potentially changing thanks to 00:14:04.000 --> 00:14:06.159 things like Xochitl. The speed of tech adoption 00:14:06.159 --> 00:14:08.600 is faster than ever. Business models are under 00:14:08.600 --> 00:14:11.879 pressure. The global AI race is real. Your job 00:14:11.879 --> 00:14:13.820 skills likely need to evolve towards collaborating 00:14:13.820 --> 00:14:16.419 with AI. You know, understanding this landscape, 00:14:16.539 --> 00:14:18.799 it isn't just tech news anymore. It's about understanding 00:14:18.799 --> 00:14:20.620 the environment we're all working and living 00:14:20.620 --> 00:14:22.700 in. This deep dive is really about giving you 00:14:22.700 --> 00:14:25.340 that clarity, that shortcut. Yeah, exactly. It's 00:14:25.340 --> 00:14:27.340 less about the future hype and more about how 00:14:27.340 --> 00:14:29.720 these autonomous systems might genuinely change 00:14:29.720 --> 00:14:32.320 fields like scientific research or how these 00:14:32.320 --> 00:14:34.440 tools that are popping up daily might impact 00:14:34.440 --> 00:14:37.200 your specific job or your industry or how you 00:14:37.200 --> 00:14:39.519 learn new things. This knowledge helps you see 00:14:39.519 --> 00:14:41.960 the shift as it's happening. And it highlights 00:14:41.960 --> 00:14:44.779 both sides, right? The capability is advancing 00:14:44.779 --> 00:14:47.500 incredibly fast. Xochitl proved that. The tools 00:14:47.500 --> 00:14:50.080 proved that. But the challenges, the costs Meeker 00:14:50.080 --> 00:14:52.200 mentioned, the safety issues that note on deception 00:14:52.200 --> 00:14:56.360 raises, they're just as real. It's about grasping 00:14:56.360 --> 00:14:59.320 that whole picture as we enter this new operational 00:14:59.320 --> 00:15:02.220 phase. Absolutely. So quite a journey today. 00:15:02.320 --> 00:15:04.580 We went from Xochitl's groundbreaking autonomous 00:15:04.580 --> 00:15:08.019 research through the big market shifts and trends 00:15:08.019 --> 00:15:11.080 in Meeker's report. and down into the nitty gritty 00:15:11.080 --> 00:15:13.940 of all these new tools changing how things get 00:15:13.940 --> 00:15:16.580 done. It covers a lot of ground. And maybe a 00:15:16.580 --> 00:15:18.960 final thought to leave you with. If AI like Xochitl 00:15:18.960 --> 00:15:21.080 is now moving beyond just analyzing existing 00:15:21.080 --> 00:15:23.279 knowledge and actually autonomously creating 00:15:23.279 --> 00:15:25.980 brand new peer -reviewed knowledge, what does 00:15:25.980 --> 00:15:28.019 that fundamentally alter about the process of 00:15:28.019 --> 00:15:30.419 discovery itself? It's a pretty deep question 00:15:30.419 --> 00:15:33.259 to mull over as this operational reality of AI 00:15:33.259 --> 00:15:34.960 just keeps accelerating around us.