WEBVTT 00:00:00.000 --> 00:00:02.339 Imagine, just for a sec, a world where you don't 00:00:02.339 --> 00:00:05.620 need to learn complex coding languages or spend 00:00:05.620 --> 00:00:08.480 years debugging lines of a pickier text just 00:00:08.480 --> 00:00:11.359 to create software. Like, you know, picture building 00:00:11.359 --> 00:00:14.160 entire applications just by talking to a computer 00:00:14.160 --> 00:00:17.059 in plain English. Yeah. That sounds pretty revolutionary, 00:00:17.100 --> 00:00:19.940 right? It really does. Well, today we're diving 00:00:19.940 --> 00:00:22.600 deep into exactly that. The fundamental shift 00:00:22.600 --> 00:00:24.780 happening right now in software development. 00:00:25.199 --> 00:00:28.579 It's all driven by large language models, LLMs, 00:00:28.719 --> 00:00:33.079 and many are calling it software 3 .0. That's 00:00:33.079 --> 00:00:35.140 the term, yeah, software 3 .0. We've got some 00:00:35.140 --> 00:00:37.060 incredible sources for this deep dive. We'll 00:00:37.060 --> 00:00:39.280 be pulling insights directly from Andrej Karpathy, 00:00:39.399 --> 00:00:41.799 who really helped define this whole concept. 00:00:42.039 --> 00:00:44.600 He really did. His perspective is key here. Plus, 00:00:44.619 --> 00:00:47.020 we've got some really solid industry reports 00:00:47.020 --> 00:00:50.640 and the very latest data on AI trends that are, 00:00:50.759 --> 00:00:53.549 well, they might surprise you. Yeah, some interesting 00:00:53.549 --> 00:00:55.590 numbers coming out. Our mission today is all 00:00:55.590 --> 00:00:58.130 about unpacking how LLMs are changing everything. 00:00:58.509 --> 00:01:00.909 I mean, from what it even means to be a programmer 00:01:00.909 --> 00:01:04.090 to how we search for information and even how 00:01:04.090 --> 00:01:07.209 our jobs are evolving. It touches so many areas. 00:01:07.409 --> 00:01:08.950 It's a good one, so let's jump right in. Let's 00:01:08.950 --> 00:01:11.730 do it. Okay, let's unpack this. Andrej Karpathy's 00:01:11.730 --> 00:01:16.599 vision of Software 3 .0. What is this exactly? 00:01:16.799 --> 00:01:19.579 Like what makes it so profoundly different from 00:01:19.579 --> 00:01:22.420 what came before? Right. So what's really key 00:01:22.420 --> 00:01:25.040 here is understanding the progression. Think 00:01:25.040 --> 00:01:28.299 about it. Software 1 .0 was traditional handwritten 00:01:28.299 --> 00:01:31.879 code. Yeah, the classic C++ to your Java. Exactly. 00:01:32.019 --> 00:01:34.840 You were like explicitly telling the computer 00:01:34.840 --> 00:01:39.180 every single step. Then software 2 .0 came along 00:01:39.180 --> 00:01:42.260 with neural networks. Okay. This was a huge shift. 00:01:42.379 --> 00:01:44.219 You weren't writing explicit code so much as 00:01:44.219 --> 00:01:46.980 focusing on data and weights. Those are the numerical 00:01:46.980 --> 00:01:49.140 parameters within the network that get adjusted 00:01:49.140 --> 00:01:51.700 during training to learn patterns. So the network 00:01:51.700 --> 00:01:54.280 learned from the data itself. Precisely. And 00:01:54.280 --> 00:01:56.500 now, software 3 .0, that's where you program 00:01:56.500 --> 00:01:59.019 with language, using LLMs. It's a fundamental 00:01:59.019 --> 00:02:02.159 leap from writing intricate code to just... Just 00:02:02.159 --> 00:02:03.859 talking to it. You know, it's kind of like instead 00:02:03.859 --> 00:02:05.599 of writing assembly language, you're just giving 00:02:05.599 --> 00:02:08.120 a verbal command like make me an app that tracks 00:02:08.120 --> 00:02:10.960 my running. It interprets the intent. So if you're 00:02:10.960 --> 00:02:13.860 programming with language, the sources mentioned 00:02:13.860 --> 00:02:17.300 some truly transformative implications like building 00:02:17.300 --> 00:02:19.939 apps just using English prompts and this idea 00:02:19.939 --> 00:02:24.139 that the product is the prompt. Can you unpack 00:02:24.139 --> 00:02:26.080 what that really means? Yeah, it means exactly 00:02:26.080 --> 00:02:28.469 what it sounds like, really. Karpathy emphasizes 00:02:28.469 --> 00:02:31.590 this concept of immediate accessibility for billions 00:02:31.590 --> 00:02:34.889 of people. Billions. Wow. Yeah. If the prompt 00:02:34.889 --> 00:02:37.409 is the product and you can just write it in plain 00:02:37.409 --> 00:02:40.430 English, then suddenly everyone, almost anyone 00:02:40.430 --> 00:02:42.969 with an idea becomes a programmer. So no more 00:02:42.969 --> 00:02:46.689 gatekeeping by code. Exactly. This really opens 00:02:46.689 --> 00:02:49.389 the door for kids, for creatives, for pretty 00:02:49.389 --> 00:02:52.370 much any non -tech user. He calls it vibe coding. 00:02:52.610 --> 00:02:54.729 Vibe coding. I like that. You're just expressing 00:02:54.729 --> 00:02:56.909 what you want, you know, your vibe and the LLM. 00:02:56.969 --> 00:02:58.949 interprets it and builds it, or at least that's 00:02:58.949 --> 00:03:01.569 the vision. Wow. So anyone can code now. That's 00:03:01.569 --> 00:03:05.590 like truly huge. If my neighbor who's never touched 00:03:05.590 --> 00:03:07.409 a line of code can suddenly build an app just 00:03:07.409 --> 00:03:10.430 by talking to it, that's unprecedented. It feels 00:03:10.430 --> 00:03:14.610 that way, doesn't it? But are there any inherent 00:03:14.610 --> 00:03:18.090 limitations or potential downsides to this vibe 00:03:18.090 --> 00:03:20.550 coding approach that we should consider? I mean, 00:03:20.569 --> 00:03:22.810 it can't be that simple. Right? That's a great 00:03:22.810 --> 00:03:25.110 question. And no, it's not quite that simple 00:03:25.110 --> 00:03:28.710 yet. While the accessibility is profound, Karpathy 00:03:28.710 --> 00:03:31.409 also draws an interesting analogy. He says we're 00:03:31.409 --> 00:03:35.669 currently in the pre -PC era for LLMs. Pre -PC. 00:03:36.280 --> 00:03:39.020 Like the 1960s? Yeah. We're sort of stuck there 00:03:39.020 --> 00:03:41.139 in terms of the ecosystem, the underlying stack, 00:03:41.419 --> 00:03:43.699 the tools, the infrastructure. It's definitely 00:03:43.699 --> 00:03:46.379 forming. You've got things like Hugging Face 00:03:46.379 --> 00:03:48.860 becoming the new GitHub for these linguistic 00:03:48.860 --> 00:03:51.060 programs. Okay. So the pieces are starting to 00:03:51.060 --> 00:03:53.780 appear. Right. But he points out our current 00:03:53.780 --> 00:03:56.500 DevOps and infrastructure aren't really LLM friendly. 00:03:56.659 --> 00:03:58.680 They weren't built for this. We actually need 00:03:58.680 --> 00:04:00.900 to redesign the entire underlying infrastructure 00:04:00.900 --> 00:04:03.819 so that these AI agents can directly navigate 00:04:03.819 --> 00:04:06.840 and manipulate it rather than us humans translating 00:04:06.840 --> 00:04:09.520 for them. So the limitations right now are often 00:04:09.520 --> 00:04:11.539 in the plumbing. You know, the vision is there, 00:04:11.639 --> 00:04:14.759 but the execution needs a lot more work. That's 00:04:14.759 --> 00:04:16.519 a really interesting point about the infrastructure 00:04:16.519 --> 00:04:20.060 needing to catch up. And he also brought up this 00:04:20.060 --> 00:04:22.319 idea that this is the decade of Iron Man suits. 00:04:22.639 --> 00:04:25.579 Small chuckle. Yeah, that's a great image. It's 00:04:25.579 --> 00:04:29.779 such a vivid image. Within that vision, what's 00:04:29.779 --> 00:04:33.160 the fundamental question he poses? Is it like, 00:04:33.180 --> 00:04:36.180 are we building these suits for humans or bots 00:04:36.180 --> 00:04:39.620 or both? Precisely. It raises an important question 00:04:39.620 --> 00:04:42.459 about the actual end user or maybe end agent. 00:04:42.680 --> 00:04:46.259 As these powerful tools emerge, the focus shifts. 00:04:46.730 --> 00:04:49.889 Who or what is the primary beneficiary? Are we 00:04:49.889 --> 00:04:52.490 empowering human users to do more amazing things? 00:04:52.769 --> 00:04:55.350 Or are we building intelligent agents that will 00:04:55.350 --> 00:04:57.689 mostly interact with each other, automating complex 00:04:57.689 --> 00:04:59.750 workflows behind the scenes? We're making a mix. 00:05:00.160 --> 00:05:02.259 It's likely a mix. Yeah, it's a complex, evolving 00:05:02.259 --> 00:05:04.100 question. But, you know, it's really important 00:05:04.100 --> 00:05:06.459 to remember his immediate caveat. He's very clear 00:05:06.459 --> 00:05:09.279 on this. He stresses that this stuff is not production 00:05:09.279 --> 00:05:11.860 ready without guardrails. While the vision is 00:05:11.860 --> 00:05:14.699 grand, the practical, reliable and safe application 00:05:14.699 --> 00:05:17.740 still requires significant development. He really 00:05:17.740 --> 00:05:19.980 underscores that safety and reliability aspect. 00:05:20.300 --> 00:05:22.259 That's a crucial point, the guardrails. Absolutely. 00:05:22.759 --> 00:05:25.480 So moving from the programming aspect, let's 00:05:25.480 --> 00:05:29.389 transition a bit. How is AI impacting? our jobs 00:05:29.389 --> 00:05:31.629 and daily tasks what does this all mean for how 00:05:31.629 --> 00:05:34.629 we work you know day to day well the sources 00:05:34.629 --> 00:05:36.990 point out something quite stark actually the 00:05:36.990 --> 00:05:41.810 idea is if your job can be measured ai can probably 00:05:41.810 --> 00:05:44.750 automate parts of it or maybe even all of it 00:05:44.750 --> 00:05:47.350 eventually measured how like Quantitatively. 00:05:47.370 --> 00:05:49.709 Both ways, apparently. It applies to both hard 00:05:49.709 --> 00:05:53.189 tasks like complex data analysis and what they 00:05:53.189 --> 00:05:55.750 call soft tasks, things like drafting communications 00:05:55.750 --> 00:05:58.649 or even generating creative content. Like writing 00:05:58.649 --> 00:06:00.970 emails, reports. Yeah, things like that. There 00:06:00.970 --> 00:06:04.110 are legitimate job displacement concerns. Of 00:06:04.110 --> 00:06:06.029 course, we can't ignore that. But the surprising 00:06:06.029 --> 00:06:08.209 counterpoint is that the sources also highlight 00:06:08.209 --> 00:06:11.009 22 new jobs AI could actually give back. Oh, 00:06:11.069 --> 00:06:12.949 interesting. So it's not just taking away. It's 00:06:12.949 --> 00:06:14.810 not just a one -way street of automation, no. 00:06:15.009 --> 00:06:16.790 And these aren't just mine. roles either. We're 00:06:16.790 --> 00:06:19.350 talking about entirely new categories like AI 00:06:19.350 --> 00:06:23.069 prompt engineers, AI ethicists, data curators 00:06:23.069 --> 00:06:26.250 for AI, and roles focused on human AI collaboration. 00:06:26.769 --> 00:06:29.910 Hmm. Overseeing the AI, basically. It underscores 00:06:29.910 --> 00:06:32.490 a shift, yeah. From manual execution to strategic 00:06:32.490 --> 00:06:34.850 oversight and creative direction alongside AI. 00:06:35.129 --> 00:06:38.310 Less doing, more directing. That's a really interesting 00:06:38.310 --> 00:06:41.629 shift in focus for jobs. And speaking of how 00:06:41.629 --> 00:06:44.930 AI operates, the sources presented some fascinating 00:06:44.930 --> 00:06:48.730 data on AI's quirks. One observation that truly 00:06:48.730 --> 00:06:51.990 stood out to me was this strange tendency for 00:06:51.990 --> 00:06:55.290 AIs to return the number 27 when asked for a 00:06:55.290 --> 00:06:57.529 random number. Chuckles. Yeah, the 27 thing. 00:06:57.709 --> 00:07:00.029 Why do you think that happens? It feels almost 00:07:00.029 --> 00:07:03.550 superstitious or like an inside joke for the 00:07:03.550 --> 00:07:06.529 AI. It's a fascinating glimpse into the black 00:07:06.529 --> 00:07:09.430 box of AI, isn't it? It's not superstition, though, 00:07:09.449 --> 00:07:11.310 not really. There are a few theories floating 00:07:11.310 --> 00:07:14.230 around. Some suggest it's due to subtle biases 00:07:14.230 --> 00:07:17.430 in the training data itself. Maybe 27 appeared 00:07:17.430 --> 00:07:20.170 more frequently in example sets meant to demonstrate 00:07:20.170 --> 00:07:23.430 randomness, ironically. Others think it's a quirk 00:07:23.430 --> 00:07:25.750 in the underlying statistical models or the way 00:07:25.750 --> 00:07:28.389 the model initializes its internal states when 00:07:28.389 --> 00:07:30.350 asked for something random. It kind of defaults 00:07:30.350 --> 00:07:32.170 somewhere. So it's like a red it falls into. 00:07:32.389 --> 00:07:35.230 Sort of. It reminds us that even when we ask 00:07:35.230 --> 00:07:37.670 for randomness, these models are determined. 00:07:42.009 --> 00:07:44.910 It's a subtle but important lesson in how AI 00:07:44.910 --> 00:07:49.069 actually thinks or processes information. It's 00:07:49.069 --> 00:07:52.199 not... truly random in the human sense. That's 00:07:52.199 --> 00:07:53.819 a great insight, you know, understanding that 00:07:53.819 --> 00:07:56.360 deterministic nature. Okay, here's where it gets 00:07:56.360 --> 00:07:58.360 really interesting for day -to -day life, though. 00:07:58.759 --> 00:08:01.939 How can AI specifically help us with those mundane 00:08:01.939 --> 00:08:04.959 tasks we all face, saving us time and effort? 00:08:05.160 --> 00:08:07.079 Right, the practical stuff. Well, the sources 00:08:07.079 --> 00:08:10.279 list nine mundane tasks that ChatGPT, for example, 00:08:10.279 --> 00:08:12.620 can handle in seconds, tasks that would otherwise 00:08:12.620 --> 00:08:14.939 take you hours. Like what? Give me some examples. 00:08:15.500 --> 00:08:18.519 Things like drafting emails, summarizing long 00:08:18.519 --> 00:08:21.120 documents or articles, brainstorming ideas for 00:08:21.120 --> 00:08:23.620 a project, maybe outlining a presentation or 00:08:23.620 --> 00:08:26.180 even planning a complex travel itinerary. OK, 00:08:26.300 --> 00:08:27.839 yeah, I can see that saving time. Definitely. 00:08:27.920 --> 00:08:31.399 It really makes the case that we maybe need to 00:08:31.399 --> 00:08:33.600 chat GPT more, leveraging it for all those small, 00:08:33.759 --> 00:08:36.039 time -consuming administrative burdens that just 00:08:36.039 --> 00:08:38.080 eat up our day, freeing us up for more important 00:08:38.080 --> 00:08:41.279 things. And it's not just chat GPT either, right? 00:08:42.250 --> 00:08:45.169 The sources highlight some truly powerful new 00:08:45.169 --> 00:08:49.070 AI tools that are changing how we approach creative 00:08:49.070 --> 00:08:51.649 and administrative tasks. Oh, yeah. There's a 00:08:51.649 --> 00:08:54.330 whole ecosystem popping up. Like take Higgs field 00:08:54.330 --> 00:08:56.750 canvas for image editing. Imagine transforming 00:08:56.750 --> 00:08:59.750 a simple sketch into a photorealistic image in 00:08:59.750 --> 00:09:03.830 seconds. That's wild. It is. Or Y -code AI, which 00:09:03.830 --> 00:09:06.250 lets you build an entire website just by describing 00:09:06.250 --> 00:09:08.129 what you want. Just describing it. No coding. 00:09:08.490 --> 00:09:11.570 Apparently so. But those are just two examples 00:09:11.570 --> 00:09:14.629 of this whole new suite of digital superpowers 00:09:14.629 --> 00:09:18.049 emerging. There's also Hylou 002 for video, Vanny 00:09:18.049 --> 00:09:20.309 for video answers, and 8Coder for workflows. 00:09:20.669 --> 00:09:23.350 The list keeps growing. And what's really striking 00:09:23.350 --> 00:09:25.809 is how these individual tools connect to broader 00:09:25.809 --> 00:09:27.830 innovations. For instance, there's something 00:09:27.830 --> 00:09:30.970 called the Model Context Protocol, or MCP. MCP. 00:09:31.070 --> 00:09:33.049 What's that? It's being developed to save potentially 00:09:33.049 --> 00:09:36.429 90 % of manual AI integration work. 90 %? That's 00:09:36.429 --> 00:09:38.539 huge. It is. Because it's going to let you turn 00:09:38.539 --> 00:09:40.919 chat bots into true AI assistants and connect 00:09:40.919 --> 00:09:43.000 AI directly to your existing tools and workflows 00:09:43.000 --> 00:09:45.019 much more easily. Less friction. Making them 00:09:45.019 --> 00:09:47.139 talk to each other properly. Right. Exactly. 00:09:47.399 --> 00:09:50.299 And speaking of big developments, there's been 00:09:50.299 --> 00:09:53.559 significant AI grant news recently, like Scale 00:09:53.559 --> 00:09:57.460 AI's massive deal, reportedly over $10 billion 00:09:57.460 --> 00:10:00.360 from Meta. Wow, $10 billion. These aren't just 00:10:00.360 --> 00:10:02.679 small startups playing around. These are huge 00:10:02.679 --> 00:10:05.159 investments that are fundamentally shaping the 00:10:05.159 --> 00:10:07.820 future of AI infrastructure and capability. The 00:10:07.820 --> 00:10:10.120 money is serious. Okay, now let's shift gears 00:10:10.120 --> 00:10:12.759 from the user experience and zoom out to the 00:10:12.759 --> 00:10:14.679 corporate landscape. Let's focus on the major 00:10:14.679 --> 00:10:16.639 players shaping this future. You know, what's 00:10:16.639 --> 00:10:18.720 OpenAI up to? They seem to be at the center of 00:10:18.720 --> 00:10:20.980 a lot of this. Absolutely. OpenAI is definitely 00:10:20.980 --> 00:10:23.460 a central figure. Sam Altman has revealed their 00:10:23.460 --> 00:10:25.980 roadmap. And it's really all about making ChatGPT 00:10:25.980 --> 00:10:28.820 simple again, focusing on usability and power. 00:10:28.960 --> 00:10:31.120 Simple again. What does that mean? Focusing on 00:10:31.120 --> 00:10:34.000 core capabilities, making it easier to use, more 00:10:34.000 --> 00:10:36.740 reliable. And the expectation is that GPT -5 00:10:36.740 --> 00:10:39.740 should probably arrive this summer, which they 00:10:39.740 --> 00:10:43.250 seem to define as summer 2025. GPT -5. OK. This 00:10:43.250 --> 00:10:46.129 next iteration promises faster AI training and 00:10:46.129 --> 00:10:48.909 significantly enhanced capabilities. We don't 00:10:48.909 --> 00:10:50.610 know the specifics yet, but the anticipation 00:10:50.610 --> 00:10:52.549 is high. And they've got the funding for it, 00:10:52.570 --> 00:10:54.490 right? Oh, absolutely. It's attracted massive 00:10:54.490 --> 00:10:57.250 funding, $13 billion from Microsoft, and there 00:10:57.250 --> 00:11:00.309 was talk of a potential $40 billion from SoftBank. 00:11:00.409 --> 00:11:03.049 That kind of investment really speaks volumes 00:11:03.049 --> 00:11:06.269 about the perceived future impact and, you know, 00:11:06.289 --> 00:11:09.049 the intense race to build these next -gen models. 00:11:09.330 --> 00:11:11.450 But there's also some. A bit of drama there, 00:11:11.549 --> 00:11:14.090 right? I saw something about OpenAI files being 00:11:14.090 --> 00:11:16.970 dropped by watchdogs, something about receipts 00:11:16.970 --> 00:11:20.730 on sketchy CEO moves, safety mess -ups, and an 00:11:20.730 --> 00:11:22.950 organizational structure described as highly 00:11:22.950 --> 00:11:25.889 convoluted. What's that all about? Sounds messy. 00:11:26.049 --> 00:11:27.529 Yeah, there's been some reporting around that. 00:11:27.820 --> 00:11:29.700 I think a key takeaway here, stepping back from 00:11:29.700 --> 00:11:32.080 the specifics, is that these AGI kings, these 00:11:32.080 --> 00:11:34.539 companies striving for artificial general intelligence, 00:11:34.820 --> 00:11:37.659 systems that theoretically can perform any intellectual 00:11:37.659 --> 00:11:39.840 task a human can while they're getting audited. 00:11:40.059 --> 00:11:42.700 Public scrutiny is increasing. Exactly. It really 00:11:42.700 --> 00:11:45.200 shows the intense scrutiny around these powerful 00:11:45.200 --> 00:11:47.940 entities, given their immense potential influence. 00:11:48.259 --> 00:11:50.700 It highlights the growing need for transparency 00:11:50.700 --> 00:11:53.379 and oversight as these models become more capable 00:11:53.379 --> 00:11:57.539 and more central to our lives. watching how they 00:11:57.539 --> 00:12:00.039 manage their power and their internal structures, 00:12:00.159 --> 00:12:02.399 safety protocols, all of it. It certainly does. 00:12:02.539 --> 00:12:04.820 Okay, so let's shift a little and look at this 00:12:04.820 --> 00:12:07.379 AI chart that compares chat GPT and Google search. 00:12:07.720 --> 00:12:11.059 What does all this mean for how we find information? 00:12:11.139 --> 00:12:12.879 You know, how we search for things online now. 00:12:13.240 --> 00:12:16.720 Is Google doomed? Shuckles slightly, the is Google 00:12:16.720 --> 00:12:19.500 doomed question. Well, the data is pretty clear 00:12:19.500 --> 00:12:21.399 for now. Google's daily search count is still 00:12:21.399 --> 00:12:24.620 immense, about 13 .7 billion searches a day. 00:12:24.840 --> 00:12:28.179 13 billion, wow. Yeah. ChatGPT, by comparison, 00:12:28.440 --> 00:12:30.960 pulls in around 1 billion daily interactions 00:12:30.960 --> 00:12:33.139 that look like searches or queries. So you're 00:12:33.139 --> 00:12:35.299 looking at Google being roughly 13 times more 00:12:35.299 --> 00:12:37.879 active in that traditional search sense. Okay, 00:12:37.899 --> 00:12:40.539 so Google's still dominant by volume. By sheer 00:12:40.539 --> 00:12:43.240 volume, yes. What's interesting, though, is that 00:12:43.240 --> 00:12:46.139 ChatGPT now apparently matches TikTok in raw 00:12:46.139 --> 00:12:49.460 query volume. But other platforms like Amazon, 00:12:49.820 --> 00:12:53.399 LinkedIn, and even Pinterest still have significantly 00:12:53.399 --> 00:12:56.519 higher engagement within their specific domains. 00:12:57.460 --> 00:13:01.360 So, like, is ChatGPT ever going to pass Google? 00:13:02.519 --> 00:13:05.159 That's the big question, right? Everyone's talking 00:13:05.159 --> 00:13:07.100 about it, wondering if the search engine as we 00:13:07.100 --> 00:13:09.769 know it is going away. It's probably not a simple 00:13:09.769 --> 00:13:11.830 replacement scenario. The answer is nuanced. 00:13:12.210 --> 00:13:14.230 You know, chat GPT is being used differently. 00:13:14.470 --> 00:13:18.269 It's more like a tutor or an assistant or a content 00:13:18.269 --> 00:13:21.029 generator, not just a traditional link aggregator. 00:13:21.330 --> 00:13:23.649 People are often looking for deep synthesized 00:13:23.649 --> 00:13:26.149 answers or creative help rather than just a list 00:13:26.149 --> 00:13:27.850 of websites. Right. Getting the answer directly 00:13:27.850 --> 00:13:30.230 instead of links to answers. Exactly. Users seem 00:13:30.230 --> 00:13:32.330 to be shifting to what the sources call a hybrid 00:13:32.330 --> 00:13:34.710 habit. They might go to Google for quick facts, 00:13:34.809 --> 00:13:37.129 news headlines or product searches areas where 00:13:37.129 --> 00:13:39.370 AI has. hasn't fully broken in yet or where speed 00:13:39.370 --> 00:13:42.450 is key. But then they'll go to ChatGPT or similar 00:13:42.450 --> 00:13:45.210 tools for depth for more comprehensive explanations 00:13:45.210 --> 00:13:47.370 or for help with creative tasks like writing 00:13:47.370 --> 00:13:49.929 or coding. And Google's not standing still either. 00:13:50.350 --> 00:13:52.429 Definitely not. Google is already integrating 00:13:52.429 --> 00:13:56.129 AI directly into search, like with Search Live 00:13:56.129 --> 00:13:59.250 and AI Overview. So they're evolving, too. It's 00:13:59.250 --> 00:14:01.330 really becoming about different tools for different 00:14:01.330 --> 00:14:04.509 types of information needs, you know, not just 00:14:04.509 --> 00:14:07.350 one tool for everything. OK, so to quickly recap 00:14:07.350 --> 00:14:10.049 this deep dive, then we've gone from this truly 00:14:10.049 --> 00:14:13.950 revolutionary software 3 .0 concept where you 00:14:13.950 --> 00:14:16.909 can literally vibe code just by talking to a 00:14:16.909 --> 00:14:19.899 computer. Which is still. Mind -blowing. It really 00:14:19.899 --> 00:14:22.740 is. To the surprising ways AI is impacting our 00:14:22.740 --> 00:14:25.399 jobs, both taking tasks and creating new roles 00:14:25.399 --> 00:14:28.399 and simplifying those dually mundane tasks. Getting 00:14:28.399 --> 00:14:30.399 rid of the busy work. Right. And then, you know, 00:14:30.399 --> 00:14:32.399 we looked at the evolving dynamics between the 00:14:32.399 --> 00:14:35.159 AI giants like OpenAI, the scrutiny they're under, 00:14:35.299 --> 00:14:37.299 and how they stack up against traditional search 00:14:37.299 --> 00:14:39.580 engines like Google. It's been a lot to unpack, 00:14:39.799 --> 00:14:42.700 but really, really insightful. And if we connect 00:14:42.700 --> 00:14:45.799 this to the bigger picture, it's clear we're 00:14:45.799 --> 00:14:48.899 not just looking at a few new... cool tools or 00:14:48.899 --> 00:14:51.820 apps. It's much bigger than that. How do you 00:14:51.820 --> 00:14:54.139 see it? It feels like a profound shift in how 00:14:54.139 --> 00:14:57.059 we think about computing itself and certainly 00:14:57.059 --> 00:14:59.679 human machine interaction. The very nature of 00:14:59.679 --> 00:15:01.700 how we create, how we search, how we consume 00:15:01.700 --> 00:15:04.580 digital experiences is fundamentally changing 00:15:04.580 --> 00:15:07.019 right underneath our feet. So here's a thought 00:15:07.019 --> 00:15:09.480 to leave everyone with. If the product really 00:15:09.480 --> 00:15:12.539 is the prompt and almost anyone can vibe code, 00:15:12.820 --> 00:15:16.809 what kind of people spirits? as Carpathy might 00:15:16.809 --> 00:15:19.549 put it, are we really inviting into our digital 00:15:19.549 --> 00:15:22.450 world through these prompts? That's deep. And 00:15:22.450 --> 00:15:24.549 what responsibilities come with that immense 00:15:24.549 --> 00:15:26.730 power to create complex things with just words? 00:15:27.149 --> 00:15:29.230 Yeah, that really raises an important question 00:15:29.230 --> 00:15:31.830 for all of us, doesn't it? How will we navigate 00:15:31.830 --> 00:15:34.669 this rapidly evolving landscape? And what role 00:15:34.669 --> 00:15:37.450 do we want to play in shaping it responsibly? 00:15:37.590 --> 00:15:39.350 Something to really think about. Definitely food 00:15:39.350 --> 00:15:40.850 for thought. Until next time.