WEBVTT 00:00:00.000 --> 00:00:02.319 Imagine a world where your company's workforce 00:00:02.319 --> 00:00:06.759 could surge from just a handful of people to 00:00:06.759 --> 00:00:09.900 over 100 instantly, maybe just for a few hours. 00:00:10.199 --> 00:00:12.919 A world where headcount isn't this, you know, 00:00:12.919 --> 00:00:15.919 slow moving staircase, but more like a dynamic 00:00:15.919 --> 00:00:18.760 kind of spiky crypto chart. We're going to unpack 00:00:18.760 --> 00:00:21.800 how this is actually becoming our reality. Welcome 00:00:21.800 --> 00:00:24.600 to the Deep Dive. Today, we're peeling back the 00:00:24.600 --> 00:00:26.440 layers on something genuinely transformative 00:00:26.440 --> 00:00:30.320 in AI, the ChachiPT agent model. People are calling 00:00:30.320 --> 00:00:32.479 it the digital worker revolution. And honestly, 00:00:32.579 --> 00:00:34.280 it feels like a really profound shift in how 00:00:34.280 --> 00:00:36.439 we even think about work. Absolutely. Yeah, we've 00:00:36.439 --> 00:00:38.659 been digging into the latest research that guides 00:00:38.659 --> 00:00:41.219 around this whole new era. And it's, well, it's 00:00:41.219 --> 00:00:42.659 definitely clear this isn't just another tech 00:00:42.659 --> 00:00:44.679 update, right? Our mission today is to really 00:00:44.679 --> 00:00:47.280 get a handle on how these general purpose digital 00:00:47.280 --> 00:00:49.579 workers are fundamentally reshaping how businesses 00:00:49.579 --> 00:00:52.119 operate. And maybe more importantly, how you 00:00:52.119 --> 00:00:54.520 might engage with this new frontier as it unfolds. 00:00:54.750 --> 00:00:57.189 So we'll kick off by defining what these new 00:00:57.189 --> 00:00:58.869 agents actually are, how they stack up against 00:00:58.869 --> 00:01:01.270 older AI. Then we'll dive into that big debate, 00:01:01.390 --> 00:01:04.230 APIs versus computer vision. It's a fascinating 00:01:04.230 --> 00:01:07.510 strategic split for AI, and we'll see which path 00:01:07.510 --> 00:01:09.909 seems to be winning out. After that, we'll break 00:01:09.909 --> 00:01:11.930 down what this means strategically for businesses, 00:01:12.189 --> 00:01:14.629 look at the whole AI automation landscape as 00:01:14.629 --> 00:01:16.790 it's evolving, and finally, really dig into the 00:01:16.790 --> 00:01:19.870 big question for the future of work. Is it replacement? 00:01:20.349 --> 00:01:23.109 Yeah. Or is it augmentation? Okay, so let's start 00:01:23.109 --> 00:01:27.489 right there. OpenAI's chat GPT agent. This feels 00:01:27.489 --> 00:01:29.349 different. It really marks a profound shift, 00:01:29.489 --> 00:01:31.409 bringing us into an era of what you could call 00:01:31.409 --> 00:01:34.390 true digital workers. I mean, for years we've 00:01:34.390 --> 00:01:36.530 had these really specialized AI tools, but this. 00:01:36.969 --> 00:01:39.430 This feels like it bridges the gap to a more 00:01:39.430 --> 00:01:41.650 general purpose digital workforce. That's it 00:01:41.650 --> 00:01:44.049 exactly. And what's really helpful, I think, 00:01:44.049 --> 00:01:45.629 for getting your head around it is this analogy 00:01:45.629 --> 00:01:47.469 some of the sources use. They talk about the 00:01:47.469 --> 00:01:50.030 difference between a digital dishwasher and a 00:01:50.030 --> 00:01:53.579 digital optimist. Think about it. A digital dishwasher, 00:01:53.879 --> 00:01:56.870 it's super efficient, right? Amazing at one specific 00:01:56.870 --> 00:01:59.510 repetitive job, like maybe a bot that updates 00:01:59.510 --> 00:02:02.870 your CRM or a content generator that just churns 00:02:02.870 --> 00:02:05.489 out posts on one topic. Powerful, yeah, but also 00:02:05.489 --> 00:02:07.829 kind of brittle. It leans heavily on rigid API 00:02:07.829 --> 00:02:10.729 integrations, like a preset wash cycle. It just 00:02:10.729 --> 00:02:13.430 can't adapt outside that very narrow task. Okay, 00:02:13.509 --> 00:02:15.789 I see. So the dishwasher is great, but only for 00:02:15.789 --> 00:02:18.610 dishes. But then you have this digital optimist. 00:02:18.689 --> 00:02:20.569 This is what the general purpose agent is aiming 00:02:20.569 --> 00:02:22.330 for. It's not built for just one job, is it? 00:02:22.330 --> 00:02:25.659 It's designed to, well. operate a computer pretty 00:02:25.659 --> 00:02:27.560 much like a person does. It literally uses computer 00:02:27.560 --> 00:02:29.939 vision, which is basically the AI understanding 00:02:29.939 --> 00:02:31.939 what's on the screen, like our eyes do to see 00:02:31.939 --> 00:02:34.219 the display. It navigates the visual stuff, clicks 00:02:34.219 --> 00:02:37.240 buttons, types and fields. It adapts to almost 00:02:37.240 --> 00:02:39.939 any software without needing special APIs for 00:02:39.939 --> 00:02:42.419 everything. It can actually execute complex multi 00:02:42.419 --> 00:02:45.599 -step workflows across different apps. It's pretty 00:02:45.599 --> 00:02:48.159 wild when you think about it. It really is. And 00:02:48.159 --> 00:02:50.879 this whole vision -based approach, it begs the 00:02:50.879 --> 00:02:53.960 question, how does it truly change what tasks 00:02:53.960 --> 00:02:56.979 AI can handle today? Well, fundamentally, it 00:02:56.979 --> 00:02:59.699 lets AI use the interfaces we already have, making 00:02:59.699 --> 00:03:01.819 it adaptable to pretty much any software right 00:03:01.819 --> 00:03:05.520 out of the box. And for years, the AI world was 00:03:05.520 --> 00:03:09.139 kind of caught up in this great debate. It was 00:03:09.139 --> 00:03:11.759 about two possible futures for how AI agents 00:03:11.759 --> 00:03:14.270 would interact with everything online. Yeah, 00:03:14.409 --> 00:03:16.789 it did feel a bit like arguing about national 00:03:16.789 --> 00:03:19.830 infrastructure, you know? Like Path 1, the API 00:03:19.830 --> 00:03:22.210 everything dream, that was like picturing a bullet 00:03:22.210 --> 00:03:25.449 train future. Super clean, super fast, all programmed 00:03:25.449 --> 00:03:28.430 connections. But the catch was huge. It required 00:03:28.430 --> 00:03:32.009 every single software company, every city in 00:03:32.009 --> 00:03:34.569 this analogy, to build a standard API station. 00:03:35.110 --> 00:03:38.250 That's like decades of work, massive coordination. 00:03:38.330 --> 00:03:39.969 Maybe it would never even happen universally. 00:03:40.310 --> 00:03:43.189 Exactly. A huge if. And then there was Path 2. 00:03:43.939 --> 00:03:46.800 The computer vision reality. The self -driving 00:03:46.800 --> 00:03:49.500 car approach. This is where agents use visual 00:03:49.500 --> 00:03:51.479 recognition to navigate interfaces, just like 00:03:51.479 --> 00:03:54.139 you or I do. It might not be quite as perfectly 00:03:54.139 --> 00:03:56.080 efficient as a direct API of the bullet train, 00:03:56.180 --> 00:03:59.219 sure. But, and this is key, it uses the existing 00:03:59.219 --> 00:04:01.199 roads, all the visual interfaces we already have. 00:04:01.340 --> 00:04:03.620 It can go anywhere. Today, without needing special 00:04:03.620 --> 00:04:06.020 permission slips from every software maker, the 00:04:06.020 --> 00:04:08.219 roads are already built. Right. And now with 00:04:08.219 --> 00:04:10.080 things like the ChatGPT agent, it feels like 00:04:10.080 --> 00:04:12.180 the verdict is kind of in, doesn't it? The vision 00:04:12.180 --> 00:04:14.599 -based way, it's just faster, more practical 00:04:14.599 --> 00:04:17.060 right now. It works straight away with older 00:04:17.060 --> 00:04:19.519 software, doesn't need custom integrations for 00:04:19.519 --> 00:04:22.560 each platform. It adapts automatically when UIs 00:04:22.560 --> 00:04:25.459 change, and it can scale across, well, basically 00:04:25.459 --> 00:04:28.439 an unlimited number of applications. Whoa. Hang 00:04:28.439 --> 00:04:32.040 on. Scaling across unlimited applications without 00:04:32.040 --> 00:04:35.519 waiting for a single API. That really is a game 00:04:35.519 --> 00:04:38.019 changer. Seriously. I mean, from an operational 00:04:38.019 --> 00:04:40.500 view, what's the single biggest advantage for 00:04:40.500 --> 00:04:43.100 businesses jumping on this vision first strategy? 00:04:43.379 --> 00:04:45.819 It's that immediate compatibility. They can use 00:04:45.819 --> 00:04:48.300 their existing systems now. No waiting for new 00:04:48.300 --> 00:04:50.160 integrations. OK, this is where it gets really 00:04:50.160 --> 00:04:52.600 interesting for, say, business leaders. This 00:04:52.600 --> 00:04:54.680 shift. powered by these general purpose agents, 00:04:54.819 --> 00:04:56.980 it really changes how companies think about labor, 00:04:57.040 --> 00:04:59.540 about capacity. Historically, headcount is this 00:04:59.540 --> 00:05:01.779 very slow kind of staircase function, right? 00:05:01.839 --> 00:05:04.220 Pretty flat. It only changes with slow, expensive 00:05:04.220 --> 00:05:07.139 hiring or, you know, the opposite, layoffs. Precisely. 00:05:07.139 --> 00:05:10.560 It's usually a static, almost fixed thing. But 00:05:10.560 --> 00:05:13.019 now with these agents, productive capacity becomes 00:05:13.019 --> 00:05:16.740 incredibly dynamic, elastic even. The sources 00:05:16.740 --> 00:05:19.870 actually call it a crypto chart headcount. which 00:05:19.870 --> 00:05:22.269 is evocative. Instead of that slow staircase, 00:05:22.629 --> 00:05:24.889 your capacity chart looks like Bitcoin's price 00:05:24.889 --> 00:05:27.129 history, right? Massive vertical spikes of productivity, 00:05:27.310 --> 00:05:31.050 totally on demand. Imagine a leadership team 00:05:31.050 --> 00:05:33.350 decides on a strategy in the morning. An hour 00:05:33.350 --> 00:05:36.470 later, maybe five human team members each deploy, 00:05:36.649 --> 00:05:39.889 I don't know, 20 AI agents. Suddenly, the effective 00:05:39.889 --> 00:05:42.689 headcount for that specific job spikes from five 00:05:42.689 --> 00:05:46.379 to over 100, just for a short burst. Then tasks 00:05:46.379 --> 00:05:49.399 done, it scales right back down instantly. It's 00:05:49.399 --> 00:05:51.079 not just a surge. It's like a surgical strike 00:05:51.079 --> 00:05:53.279 of productivity. And this power plant insight 00:05:53.279 --> 00:05:55.180 they mentioned is also really something. It helps 00:05:55.180 --> 00:05:57.339 explain this multi -trillion dollar race we're 00:05:57.339 --> 00:06:00.060 seeing for data centers, for GPUs, for raw electrical 00:06:00.060 --> 00:06:02.660 power. We are pretty literally converting electricity 00:06:02.660 --> 00:06:04.879 and algorithms into on -demand white -collar 00:06:04.879 --> 00:06:07.220 labor. The implications for national infrastructure, 00:06:07.500 --> 00:06:10.680 strategic resources are huge. Wow, that's a pretty 00:06:10.680 --> 00:06:12.939 staggering thought, converting raw energy into 00:06:12.939 --> 00:06:15.459 intellectual output. But doesn't that also paint 00:06:15.459 --> 00:06:17.939 a picture of incredibly intense resource demand? 00:06:18.199 --> 00:06:20.240 Are we talking about a future where a country's 00:06:20.240 --> 00:06:22.699 compute power is as strategically vital as its 00:06:22.699 --> 00:06:25.420 oil reserves used to be? Oh, absolutely. That's 00:06:25.420 --> 00:06:27.819 precisely the point. The race isn't just about 00:06:27.819 --> 00:06:30.180 chips anymore. It's about the power grid behind 00:06:30.180 --> 00:06:32.779 them. We're seeing nations, big tech companies 00:06:32.779 --> 00:06:35.160 pouring billions, not just into data centers, 00:06:35.259 --> 00:06:38.199 but into new power plants. It absolutely shifts 00:06:38.199 --> 00:06:41.100 the geopolitical chessboard, making access to 00:06:41.100 --> 00:06:44.040 abundant, reliable and hopefully clean energy 00:06:44.040 --> 00:06:47.040 a massive strategic imperative. So getting back 00:06:47.040 --> 00:06:49.300 to the business level, how does this dynamic, 00:06:49.379 --> 00:06:51.740 almost liquid capacity truly reshape strategy 00:06:51.740 --> 00:06:54.420 and planning? Well, businesses can instantly 00:06:54.420 --> 00:06:57.240 scale up productivity for specific pushes and 00:06:57.240 --> 00:06:59.199 then just as quickly scale back down without 00:06:59.199 --> 00:07:02.160 the traditional overhead. OK, so to really understand 00:07:02.160 --> 00:07:05.800 where these new general purpose agents. fit in 00:07:05.800 --> 00:07:08.100 the grand scheme of things, it helps to think 00:07:08.100 --> 00:07:11.120 about the whole AI automation landscape maybe 00:07:11.120 --> 00:07:13.160 as a pyramid, right? A pyramid of capabilities. 00:07:13.680 --> 00:07:16.579 Moving from simple digital plumbing at the bottom 00:07:16.579 --> 00:07:19.540 up to these really autonomous workers at the 00:07:19.540 --> 00:07:21.899 top. Yeah, exactly. So at the very bottom, layer 00:07:21.899 --> 00:07:24.079 one, you've got the plumbing. Simple automations. 00:07:24.180 --> 00:07:27.399 Think tools like Zapier, Make .com. They connect 00:07:27.399 --> 00:07:29.860 different apps, shuttle data around, usually 00:07:29.860 --> 00:07:32.180 triggered by an event. There's not much actual 00:07:32.180 --> 00:07:34.279 AI thinking here. It's mostly just connecting 00:07:34.279 --> 00:07:37.439 the pipe. moving data from A to B, like digital 00:07:37.439 --> 00:07:39.899 Lego blocks. Right. Simple connections. Then 00:07:39.899 --> 00:07:41.879 moving up a level, layer two. That's what we 00:07:41.879 --> 00:07:43.980 could call the power tools. These are distinct 00:07:43.980 --> 00:07:46.279 AI tools, but they're still operated by a human. 00:07:46.519 --> 00:07:48.800 Like AI content generators or those research 00:07:48.800 --> 00:07:51.139 tools that summarize articles. A human puts something 00:07:51.139 --> 00:07:53.420 in, the AI gives something back. Powerful, yeah. 00:07:54.009 --> 00:07:55.889 but they need that human touch for every single 00:07:55.889 --> 00:07:58.889 step. And then layer three, the top of our pyramid, 00:07:58.930 --> 00:08:01.870 this is the workers, autonomous AI agents. But 00:08:01.870 --> 00:08:03.970 this layer, it actually splits into three categories. 00:08:04.129 --> 00:08:07.730 First, 3A, the specialized human -operated agent. 00:08:07.870 --> 00:08:11.310 Think of like a co -pilot for a sales rep. The 00:08:11.310 --> 00:08:14.410 human is still directly in charge, kind of overseeing 00:08:14.410 --> 00:08:17.569 and guiding the AI. Okay. Then there's 3B, the 00:08:17.569 --> 00:08:21.089 specialized automated agent. Here, the AI is 00:08:21.089 --> 00:08:23.410 built right into a workflow. It makes its own 00:08:23.410 --> 00:08:25.850 decisions. It's triggered by events. It operates 00:08:25.850 --> 00:08:27.689 without a human directly commanding it in real 00:08:27.689 --> 00:08:30.129 time. It's automated, but still for a specific 00:08:30.129 --> 00:08:32.769 job, like an AI automatically sorting support 00:08:32.769 --> 00:08:35.049 tickets or something. Right. And finally, 3C. 00:08:35.549 --> 00:08:37.429 The general purpose computer operating agent. 00:08:37.470 --> 00:08:39.009 This is the new frontier we've been talking about. 00:08:39.070 --> 00:08:41.690 The digital optimist. Yeah. You give a high level 00:08:41.690 --> 00:08:43.809 goal like research competitors and summarize 00:08:43.809 --> 00:08:46.309 findings and it figures out how to do it. It 00:08:46.309 --> 00:08:48.269 independently navigates different software interfaces 00:08:48.269 --> 00:08:50.789 using that computer vision to get the job done. 00:08:50.929 --> 00:08:53.610 This is the real autonomous digital worker that 00:08:53.610 --> 00:08:55.870 acts like a human using a computer. Looking at 00:08:55.870 --> 00:08:58.049 that pyramid, it's clear there's a whole spectrum 00:08:58.049 --> 00:09:00.830 of automation happening. Where would you say 00:09:00.830 --> 00:09:03.070 most companies are actually operating right now? 00:09:03.399 --> 00:09:05.960 You know, I think many are still firmly planted 00:09:05.960 --> 00:09:09.379 in those bottom layers, the plumbing or maybe 00:09:09.379 --> 00:09:12.159 just starting to use the power tools. The agent 00:09:12.159 --> 00:09:14.480 layer is still pretty new for most. So when it 00:09:14.480 --> 00:09:16.240 comes to actually using this stuff successfully, 00:09:16.460 --> 00:09:18.820 it's not just about having agents, right? It 00:09:18.820 --> 00:09:21.100 seems critical to know which type of agent to 00:09:21.100 --> 00:09:23.960 use for which job. It's like a master craftsman 00:09:23.960 --> 00:09:26.299 choosing exactly the right tool. You wouldn't 00:09:26.299 --> 00:09:28.519 use a sledgehammer for fine carving. You've got 00:09:28.519 --> 00:09:30.519 your high precision power tool, that's the specialized 00:09:30.519 --> 00:09:32.720 agent, and then the versatile multi -tool, the 00:09:32.720 --> 00:09:34.480 general purpose one. You know, I have to admit, 00:09:34.600 --> 00:09:37.120 I still sometimes struggle with picking the perfect 00:09:37.120 --> 00:09:40.000 tool myself. It's easy to just grab what's familiar. 00:09:40.620 --> 00:09:43.240 even if it's not the absolute best fit. It really 00:09:43.240 --> 00:09:45.559 is an art. But yet, to be clear, the specialized 00:09:45.559 --> 00:09:48.259 agent is that high -precision power tool. You 00:09:48.259 --> 00:09:50.720 use it for deep, data -heavy tasks where speed, 00:09:50.899 --> 00:09:53.659 accuracy, volume, that stuff matters most. Doing 00:09:53.659 --> 00:09:56.039 one specific thing really, really well. The gold 00:09:56.039 --> 00:09:58.059 standard use cases they mention are things like 00:09:58.059 --> 00:10:00.580 competitor research, maybe using Claude's advanced 00:10:00.580 --> 00:10:03.779 research feature, or GBT5 Pro for very structured 00:10:03.779 --> 00:10:05.799 deep outputs, or high -volume lead generation 00:10:05.799 --> 00:10:08.419 scraping websites, mining databases at scale. 00:10:08.700 --> 00:10:10.559 They handle that specific data crunching reliably. 00:10:11.039 --> 00:10:13.620 Right, the focus tool. And then the general purpose 00:10:13.620 --> 00:10:16.159 agent is your adaptable multi -tool, like a Swiss 00:10:16.159 --> 00:10:18.639 Army knife. This is what you pull out for tasks 00:10:18.639 --> 00:10:20.840 that need flexibility, tasks that cross lots 00:10:20.840 --> 00:10:23.799 of different, maybe unpredictable systems, especially 00:10:23.799 --> 00:10:26.600 systems that don't have clean APIs. These are 00:10:26.600 --> 00:10:28.679 better for those multi -app admin jobs where 00:10:28.679 --> 00:10:30.940 being adaptable and navigating tricky interfaces 00:10:30.940 --> 00:10:33.799 is more important than just raw speed on one 00:10:33.799 --> 00:10:36.559 task. We're talking gold standard uses like creating 00:10:36.559 --> 00:10:38.539 a slideshow, where an agent might handle the 00:10:38.539 --> 00:10:40.600 whole flow research online, grab content. content, 00:10:40.600 --> 00:10:42.840 format slides and presentation software, maybe 00:10:42.840 --> 00:10:46.720 even find images across multiple tools or something 00:10:46.720 --> 00:10:49.980 like really dynamic LinkedIn outreach. The agent 00:10:49.980 --> 00:10:52.100 works right in the LinkedIn interface, adapting 00:10:52.100 --> 00:10:54.580 to changes, maybe even dodging anti automation 00:10:54.580 --> 00:10:57.100 stuff in real time. It's all about navigating 00:10:57.100 --> 00:11:00.120 the actual digital environment. So when a business 00:11:00.120 --> 00:11:02.179 is standing there trying to decide between these 00:11:02.179 --> 00:11:04.779 two types of agents, what's the absolute critical 00:11:04.779 --> 00:11:07.210 first step they need to take? First off they 00:11:07.210 --> 00:11:09.129 really need to analyze the task itself. What 00:11:09.129 --> 00:11:11.889 does it actually need more? Raw speed and precision 00:11:11.889 --> 00:11:14.610 or flexibility and adaptability across systems. 00:11:14.889 --> 00:11:17.570 Looking ahead, it feels pretty clear that today's 00:11:17.570 --> 00:11:19.850 ChatGPT agent models are really just the beginning, 00:11:19.990 --> 00:11:22.350 right? Like a public beta test. The enterprise 00:11:22.350 --> 00:11:24.590 grade platforms are coming and they'll bring 00:11:24.590 --> 00:11:27.490 crucial features like custom prompts, connecting 00:11:27.490 --> 00:11:29.409 to company knowledge bases, managing multiple 00:11:29.409 --> 00:11:32.649 users, better security, all that stuff. And this 00:11:32.649 --> 00:11:34.570 is where agencies, I think, have a huge opportunity 00:11:34.570 --> 00:11:37.950 to become these key expert implementers. Oh, 00:11:37.990 --> 00:11:39.889 absolutely. The whole agency model is shifting. 00:11:40.110 --> 00:11:42.710 They're becoming personal AI assistant integrators. 00:11:42.850 --> 00:11:44.909 It's less about building code and more about 00:11:44.909 --> 00:11:47.350 being like a personal trainer for a company's 00:11:47.350 --> 00:11:50.870 entire workforce. It's a cool analogy. Imagine 00:11:50.870 --> 00:11:53.389 their service offering in three phases. First, 00:11:53.450 --> 00:11:55.429 like a fitness assessment. That's basically deep 00:11:55.429 --> 00:11:57.649 workflow auditing, finding all those repetitive, 00:11:57.769 --> 00:12:02.000 time -sucking tasks. Then phase two, the custom 00:12:02.000 --> 00:12:04.460 workout plan. This is where they customize and 00:12:04.460 --> 00:12:06.899 fine tune agents, role specific prompts, private 00:12:06.899 --> 00:12:09.279 company data, setting permissions just right. 00:12:09.440 --> 00:12:11.899 And finally, phase three, the training sessions, 00:12:12.120 --> 00:12:14.700 actually training the human staff on how to collaborate 00:12:14.700 --> 00:12:17.240 effectively with these AIs, setting up feedback 00:12:17.240 --> 00:12:19.340 loops so the agents keep getting better. It's 00:12:19.340 --> 00:12:21.840 very hands on, very continuous. The value proposition 00:12:21.840 --> 00:12:24.529 there seems incredibly strong. Take away the 00:12:24.529 --> 00:12:26.970 routine grind, right? Let people focus on the 00:12:26.970 --> 00:12:29.409 high impact, strategic, creative stuff that humans 00:12:29.409 --> 00:12:31.970 do best. And this market shift is significant, 00:12:32.149 --> 00:12:34.929 too. It's moving away from coders who were like 00:12:34.929 --> 00:12:38.129 mechanics building custom API plumbing for specific 00:12:38.129 --> 00:12:41.289 projects to coaches like driving instructors 00:12:41.289 --> 00:12:44.470 for this new digital workforce focused on optimizing 00:12:44.470 --> 00:12:47.450 workflows, human AI teamwork, which probably 00:12:47.450 --> 00:12:50.090 leads to more ongoing recurring revenue models, 00:12:50.210 --> 00:12:52.509 too. Which then naturally leads us to. Maybe 00:12:52.509 --> 00:12:54.429 the biggest question hovering over all of this, 00:12:54.509 --> 00:12:56.929 the future of work itself. Is it ultimately replacement 00:12:56.929 --> 00:12:59.190 or is it augmentation? You've got the Terminator 00:12:59.190 --> 00:13:01.789 view, arguing agents will take over entire jobs, 00:13:01.850 --> 00:13:04.009 causing major economic disruption. Then there's 00:13:04.009 --> 00:13:06.169 the Iron Man view, suggesting AI will mostly 00:13:06.169 --> 00:13:08.850 augment us, boost our abilities, and shift our 00:13:08.850 --> 00:13:11.269 focus to creativity, strategy, human connection. 00:13:11.690 --> 00:13:13.350 And then there's that interesting third way, 00:13:13.450 --> 00:13:16.429 the unlocking potential thesis, which Aaron Levy 00:13:16.429 --> 00:13:19.970 from Box talks about. He suggests AI might actually 00:13:19.970 --> 00:13:23.029 create more work overall by making projects that 00:13:23.029 --> 00:13:25.409 were just too expensive or complex before suddenly 00:13:25.409 --> 00:13:28.309 feasible. It's about expanding the total pie 00:13:28.309 --> 00:13:31.029 of what's possible, opening up new kinds of human 00:13:31.029 --> 00:13:32.970 AI collaborative jobs we haven't even really 00:13:32.970 --> 00:13:36.549 imagined yet. It's a more optimistic growth vision. 00:13:36.769 --> 00:13:38.990 So thinking of that agency shift again, how does 00:13:38.990 --> 00:13:41.879 this personal trainer approach fund? change the 00:13:41.879 --> 00:13:45.059 value they actually provide. Well, agencies seem 00:13:45.059 --> 00:13:47.960 to be moving beyond just building tools. They're 00:13:47.960 --> 00:13:50.340 becoming experts in optimizing that collaborative 00:13:50.340 --> 00:13:53.399 energy, that synergy between humans and AI systems. 00:13:53.860 --> 00:13:57.350 Mid -roll sponsor read. OK, so trying to bring 00:13:57.350 --> 00:13:59.330 this all together. We've really traced a fascinating 00:13:59.330 --> 00:14:01.490 path today. We started with those specialized 00:14:01.490 --> 00:14:04.090 digital tools, good at one thing. And now we're 00:14:04.090 --> 00:14:06.610 seeing the rise of truly general purpose digital 00:14:06.610 --> 00:14:09.610 workers and the speed and adaptability of this 00:14:09.610 --> 00:14:12.269 vision based AI, the way it just interacts with 00:14:12.269 --> 00:14:14.529 the interfaces we already have. It's here now. 00:14:14.669 --> 00:14:17.409 It's turning business capacity from that slow 00:14:17.409 --> 00:14:20.190 fixed staircase into something dynamic like that 00:14:20.190 --> 00:14:22.830 spiking crypto chart analogy. Yeah, it really 00:14:22.830 --> 00:14:24.830 boils down to a strategic choice, doesn't it? 00:14:25.129 --> 00:14:27.549 Knowing precisely when to use a specialized power 00:14:27.549 --> 00:14:30.809 tool for those focused heavy lifting tasks versus 00:14:30.809 --> 00:14:33.970 when to deploy a versatile multi -tool for complex 00:14:33.970 --> 00:14:36.169 adaptable challenges across different systems, 00:14:36.309 --> 00:14:38.450 that's becoming critical. And for businesses, 00:14:38.509 --> 00:14:40.629 for agencies, the future isn't just building 00:14:40.629 --> 00:14:42.629 tech anymore. It's about becoming these coaches, 00:14:42.769 --> 00:14:45.769 guiding that human AI partnership, and crucially, 00:14:45.789 --> 00:14:48.789 unlocking this huge backlog of innovative projects 00:14:48.789 --> 00:14:50.610 that just weren't economically possible before. 00:14:50.970 --> 00:14:53.250 This deep dive really does highlight a fundamental 00:14:53.250 --> 00:14:56.289 shift, and it's happening right now. We genuinely 00:14:56.289 --> 00:14:59.350 encourage you, listening, to start looking at 00:14:59.350 --> 00:15:01.970 your own workflows, identify those repetitive 00:15:01.970 --> 00:15:05.309 time -draining tasks, and begin thinking about 00:15:05.309 --> 00:15:08.409 preparing for a future that's really defined 00:15:08.409 --> 00:15:11.450 by this kind of intelligent, impactful human 00:15:11.450 --> 00:15:14.370 -AI collaboration. Yeah, the introduction of 00:15:14.370 --> 00:15:16.570 general -purpose AI agents. This isn't some sci 00:15:16.570 --> 00:15:18.710 -fi concept for the distant future. We can just, 00:15:18.769 --> 00:15:21.600 you know... theorize about. It's happening now. 00:15:21.860 --> 00:15:24.759 It's actively reshaping industries as we speak. 00:15:24.940 --> 00:15:26.960 So the real question isn't if it's going to transform 00:15:26.960 --> 00:15:29.059 your field, but whether you are going to choose 00:15:29.059 --> 00:15:32.159 to lead that transformation, help shave it, or 00:15:32.159 --> 00:15:34.620 just get swept along by the current, it really 00:15:34.620 --> 00:15:36.960 is about finding that sweet spot, blending uniquely 00:15:36.960 --> 00:15:39.940 human creativity and strategic insight with this 00:15:39.940 --> 00:15:42.539 incredible new potential for AI -powered execution. 00:15:42.980 --> 00:15:44.600 Absolutely. Couldn't agree more. Thank you for 00:15:44.600 --> 00:15:45.960 joining us on the Steep Dive. Let us know what 00:15:45.960 --> 00:15:47.340 you think, and we'll catch you on the next one.