WEBVTT 00:00:00.000 --> 00:00:02.399 You know, for a long time, if you were doing 00:00:02.399 --> 00:00:05.120 serious AI coding, especially for React apps, 00:00:05.400 --> 00:00:08.599 it felt like, well, a clothes shop, you had these 00:00:08.599 --> 00:00:11.980 big, expensive tools, bolt, lovable, kind of 00:00:11.980 --> 00:00:14.500 the Goliaths dominating things. Absolutely. It 00:00:14.500 --> 00:00:16.780 was basically if you wanted that top tier AI 00:00:16.780 --> 00:00:19.019 help, you had to pay up. Simple as that. Yeah. 00:00:19.120 --> 00:00:21.960 But the sources we looked into today, they show 00:00:21.960 --> 00:00:24.199 things are shifting pretty fast, actually. Yeah. 00:00:24.280 --> 00:00:27.050 We're looking at this combination. quests which 00:00:27.050 --> 00:00:30.949 is a free open source desktop app builder paired 00:00:30.949 --> 00:00:34.369 up with the deep seek v3 .1 model and here's 00:00:34.369 --> 00:00:35.869 the twist maybe the surprise this completely 00:00:35.869 --> 00:00:38.030 free setup it seems to be giving a developer 00:00:38.030 --> 00:00:40.509 experience that might actually be better than 00:00:40.509 --> 00:00:42.890 what people are paying hundreds a month for. 00:00:43.030 --> 00:00:45.229 Welcome to the deep dive. This really feels like 00:00:45.229 --> 00:00:47.490 a classic disruption story playing out. And our 00:00:47.490 --> 00:00:50.090 sources give us a really good, detailed look 00:00:50.090 --> 00:00:52.990 at this new open source challenger. So our mission 00:00:52.990 --> 00:00:55.250 today is to really unpack how this works. We 00:00:55.250 --> 00:00:56.909 need to get into why this combo is effective, 00:00:57.170 --> 00:00:59.310 look at the design thinking, this kind of cheat 00:00:59.310 --> 00:01:01.729 code for getting high performance for free, and 00:01:01.729 --> 00:01:05.039 the actual day -to -day workflow. Right. And 00:01:05.039 --> 00:01:07.200 importantly, we also got to look at the limitations. 00:01:07.400 --> 00:01:10.040 It's early access. So we need to be real about 00:01:10.040 --> 00:01:12.340 what it can and can't do right now. OK, so to 00:01:12.340 --> 00:01:14.480 really get this David versus Goliath thing, we 00:01:14.480 --> 00:01:15.920 should probably start with the old guard, right? 00:01:16.019 --> 00:01:19.040 The closed systems quests is up against. Exactly. 00:01:19.299 --> 00:01:22.700 For years, if you were a pro reactive using AI, 00:01:22.959 --> 00:01:26.299 you basically bought into those closed, costly 00:01:26.299 --> 00:01:30.480 platforms. Bolt, lovable. They set up that kind 00:01:30.480 --> 00:01:33.670 of high barrier to entry. And Quests is the app, 00:01:33.750 --> 00:01:36.530 this desktop builder, that seems designed really 00:01:36.530 --> 00:01:39.489 to just flip that script, to show that open source 00:01:39.489 --> 00:01:41.790 isn't just for playing around anymore. It's ready 00:01:41.790 --> 00:01:44.109 for professional work. And the real clever bit 00:01:44.109 --> 00:01:46.109 here isn't just the Quests app itself, though 00:01:46.109 --> 00:01:48.489 it's neat. It's how they handle the AI model 00:01:48.489 --> 00:01:50.989 access. They get this powerhouse performance 00:01:50.989 --> 00:01:54.530 for zero cost by plugging into DeepSeek v3 .1. 00:01:54.670 --> 00:01:56.950 Which, crucially, they access for free using 00:01:56.950 --> 00:01:59.670 OpenRouter. Now, for anyone not familiar, OpenRouter 00:01:59.670 --> 00:02:02.049 is like a switchboard for AI models. It lets 00:02:02.049 --> 00:02:04.170 tools like Quest tap into really powerful models 00:02:04.170 --> 00:02:06.569 like DeepSeek without being locked into one vendor 00:02:06.569 --> 00:02:10.610 or paying those direct API fees. So this mix 00:02:10.610 --> 00:02:14.689 -free app, free pro -level model, suddenly gives 00:02:14.689 --> 00:02:16.289 you a dev experience that's right up there with 00:02:16.289 --> 00:02:20.090 the paid guys. But, okay, if it's free, what's 00:02:20.090 --> 00:02:22.729 the design philosophy? What makes it tick? I 00:02:22.729 --> 00:02:24.990 think it's really about subtraction, minimalism. 00:02:24.990 --> 00:02:27.830 If you want speed and efficiency, you cut out 00:02:27.830 --> 00:02:29.750 everything else. Like you said, it's the focused, 00:02:29.969 --> 00:02:32.930 streamlined, linear, not the massive do -everything 00:02:32.930 --> 00:02:35.310 JIRA of app building. Yeah, that comparison clicks. 00:02:35.949 --> 00:02:38.810 Bolt often feels like it's trying to be the JIRA. 00:02:39.550 --> 00:02:41.449 you know, every possible feature crammed in. 00:02:41.750 --> 00:02:44.590 Quests feels way more focused, just the fastest 00:02:44.590 --> 00:02:46.949 route to getting that UI built and working. And 00:02:46.949 --> 00:02:48.810 you see that focus in the interface itself. It 00:02:48.810 --> 00:02:51.330 uses these browser -like tabs, which means you 00:02:51.330 --> 00:02:53.669 can spin up several different AI agents building 00:02:53.669 --> 00:02:55.509 completely different things and just tab between 00:02:55.509 --> 00:02:57.870 them, like instantly. It's super clean, yeah. 00:02:58.439 --> 00:03:00.860 Minimalist design avoids all that feature clutter. 00:03:00.960 --> 00:03:03.659 Plus, because it's a native desktop app not running 00:03:03.659 --> 00:03:06.080 in a browser tab, it's really light on resources. 00:03:06.259 --> 00:03:09.219 We're talking like 200 megs of RAM maybe compared 00:03:09.219 --> 00:03:11.379 to gigs for some of the browser -based ones. 00:03:11.539 --> 00:03:13.620 They also made some smart choices on the tech 00:03:13.620 --> 00:03:16.490 stack underneath it all. to keep it fast, using 00:03:16.490 --> 00:03:18.909 Vite for the front -end, not the heavier Next 00:03:18.909 --> 00:03:22.090 .js, and Anno for back -end stuff. Right. Those 00:03:22.090 --> 00:03:24.150 choices make sense for an app focused on building 00:03:24.150 --> 00:03:27.150 UIs quickly. Using heavy Next .js can feel like 00:03:27.150 --> 00:03:29.050 using a semi -truck to, I don't know, drive across 00:03:29.050 --> 00:03:31.750 town. Overkill. Vite is more like that nimble 00:03:31.750 --> 00:03:33.870 sports car you need. And they use something called 00:03:33.870 --> 00:03:37.110 ORPC to tie it together, basically a way to make 00:03:37.110 --> 00:03:39.090 sure different parts talk reliably and securely. 00:03:39.599 --> 00:03:41.659 keeps it fast and consistent. So that lightweight 00:03:41.659 --> 00:03:43.520 approach, why does that really matter for the 00:03:43.520 --> 00:03:45.719 developer using it day to day? It just means 00:03:45.719 --> 00:03:48.759 less friction, faster setup, the app feels snappier, 00:03:48.840 --> 00:03:51.379 and you can multitask much more smoothly. Okay, 00:03:51.439 --> 00:03:53.939 let's zoom in on that DeepSeek v3 .1 part because 00:03:53.939 --> 00:03:56.319 that's the core of this free performance thing. 00:03:56.439 --> 00:03:58.680 It's kind of the cheat code giving you unlimited 00:03:58.680 --> 00:04:02.039 professional level AI for coding for free. And 00:04:02.039 --> 00:04:03.539 we should be really clear here. So this isn't 00:04:03.539 --> 00:04:06.210 some, you know, budget model they found. DeepSeq 00:04:06.210 --> 00:04:09.189 v3 .1 genuinely holds its own against models 00:04:09.189 --> 00:04:12.289 like Claude, even GPT -4 for actual coding tasks. 00:04:12.650 --> 00:04:14.789 This feels like a moment where open source performance 00:04:14.789 --> 00:04:17.870 really caught up. That's huge because it means 00:04:17.870 --> 00:04:20.069 it's good at the hard stuff, generating complex 00:04:20.069 --> 00:04:23.589 code, helping debug tricky issues, even brainstorming 00:04:23.589 --> 00:04:25.910 architectural approaches, stuff that usually 00:04:25.910 --> 00:04:28.500 takes a lot of manual effort. So for a solo dev 00:04:28.500 --> 00:04:30.899 or maybe a small startup team, the value proposition 00:04:30.899 --> 00:04:34.639 is just stark. You get pro quality code, fast 00:04:34.639 --> 00:04:37.579 responses, almost no usage limits, and it costs 00:04:37.579 --> 00:04:40.500 nothing. That fundamentally changes the economics 00:04:40.500 --> 00:04:43.399 of building software. It almost sounds too good. 00:04:43.519 --> 00:04:46.470 Is there like a hidden catch with... using deep 00:04:46.470 --> 00:04:49.129 seek via open router like do they throttle you 00:04:49.129 --> 00:04:51.629 eventually not that we've seen reported the benefit 00:04:51.629 --> 00:04:54.370 seems to be genuine access they've managed to 00:04:54.370 --> 00:04:56.910 separate the model's power from the usual licensing 00:04:56.910 --> 00:04:59.829 cost structure now that multi -tab workflow you 00:04:59.829 --> 00:05:01.629 mentioned earlier that's what really takes quest 00:05:01.629 --> 00:05:03.709 beyond just being a code generator it makes it 00:05:03.709 --> 00:05:06.129 feel like yeah a supercharged web browser but 00:05:06.129 --> 00:05:08.629 specifically for building apps right like imagine 00:05:08.629 --> 00:05:10.569 you're juggling three different client projects 00:05:10.569 --> 00:05:13.610 maybe a movie tracker app a minesweeper game 00:05:13.610 --> 00:05:16.930 clone and some complex UI components. Normally, 00:05:17.050 --> 00:05:19.310 switching between those contextually, it's a 00:05:19.310 --> 00:05:22.329 nightmare. You lose so much time. But with Quests, 00:05:22.449 --> 00:05:24.649 you just pop open a new tab for each. Each tab 00:05:24.649 --> 00:05:27.410 has its own dedicated AI agent working on that 00:05:27.410 --> 00:05:29.910 project. Switching is just like changing browser 00:05:29.910 --> 00:05:32.850 tabs. Minimal friction keeps your head clear. 00:05:33.029 --> 00:05:35.500 And the instant preview. That's kind of magical 00:05:35.500 --> 00:05:37.139 when you see it. As soon as the AI generates 00:05:37.139 --> 00:05:39.339 the app, boom, there's a fully interactive preview 00:05:39.339 --> 00:05:41.139 right there in the panel. Yeah, you get the visual 00:05:41.139 --> 00:05:44.019 immediately. Plus, it gives you a live local 00:05:44.019 --> 00:05:46.699 host URL, so you can instantly test it on your 00:05:46.699 --> 00:05:49.939 phone, tablet, any device on your network. And 00:05:49.939 --> 00:05:52.180 there are one -click buttons to open the generated 00:05:52.180 --> 00:05:55.600 code in VS Code or Cursor, your editor. How does 00:05:55.600 --> 00:05:58.279 that live local host feature... change things, 00:05:58.439 --> 00:06:01.160 that instant feedback loop. It just cuts out 00:06:01.160 --> 00:06:03.920 so much waiting and friction, instant testing, 00:06:04.040 --> 00:06:06.199 easy sharing with clients or colleagues for feedback. 00:06:06.860 --> 00:06:09.319 Okay, let's do the reality check, though. This 00:06:09.319 --> 00:06:11.379 thing is labeled early access. It's important. 00:06:11.500 --> 00:06:13.819 It's not a fully polished, finished product yet. 00:06:14.000 --> 00:06:16.579 And the biggest single issue right now, the thing 00:06:16.579 --> 00:06:18.740 stopping you from easily taking a prototype into 00:06:18.740 --> 00:06:21.699 full production, it's the lack of support for 00:06:21.699 --> 00:06:24.829 environment variables. Ugh, yeah. I still wrestle 00:06:24.829 --> 00:06:27.089 with prompt drift myself sometimes, so I get 00:06:27.089 --> 00:06:29.850 wanting perfection. But that environment variable 00:06:29.850 --> 00:06:32.810 thing, it's a hard stop for many real -world 00:06:32.810 --> 00:06:35.910 apps right now. If you need API keys for external 00:06:35.910 --> 00:06:38.689 services, databases, anything like that, currently 00:06:38.689 --> 00:06:40.829 you can't really integrate them easily. So yeah, 00:06:40.930 --> 00:06:43.110 there are other missing nice -to -haves you can't 00:06:43.110 --> 00:06:45.810 attach images to prompts, no fine -grained element 00:06:45.810 --> 00:06:48.050 editing yet, can't import an existing project. 00:06:48.329 --> 00:06:51.329 But that environment variable issue, that's the 00:06:51.329 --> 00:06:53.529 main gatekeeper preventing full production use 00:06:53.529 --> 00:06:55.629 today. But then you zoom out again. This is brand 00:06:55.629 --> 00:06:57.569 new. It's being actively built. It's open source. 00:06:57.829 --> 00:07:00.449 And it already delivers, what, 80 % of what you 00:07:00.449 --> 00:07:02.269 need for really fast prototyping and learning 00:07:02.269 --> 00:07:04.769 for free. So the tradeoff right now is pretty 00:07:04.769 --> 00:07:08.009 clear, isn't it? You get amazing speed for UI 00:07:08.009 --> 00:07:10.990 prototyping at zero cost. But backend integration 00:07:10.990 --> 00:07:13.509 is basically off the table for now. Yeah. So 00:07:13.509 --> 00:07:16.519 if I can't use API keys yet. What should I focus 00:07:16.519 --> 00:07:18.560 on building with it? Stick to UI -heavy stuff. 00:07:18.759 --> 00:07:21.399 Building component libraries, simple single -page 00:07:21.399 --> 00:07:23.819 apps, learning React patterns, anywhere that 00:07:23.819 --> 00:07:25.939 Instant Preview gives the biggest boost. Okay, 00:07:25.980 --> 00:07:27.540 let's do the head -to -head, the tail of the 00:07:27.540 --> 00:07:30.560 tape, right? Quests versus the big guys like 00:07:30.560 --> 00:07:33.220 Bolt and Lovable. Speed and responsiveness. It 00:07:33.220 --> 00:07:35.579 feels like Quests gets a pretty clean knockout 00:07:35.579 --> 00:07:38.220 there. Being a native desktop app just makes 00:07:38.220 --> 00:07:41.980 it feel snappier. Lower memory use, faster startup 00:07:41.980 --> 00:07:44.100 than the browser competitors. Workflow, though. 00:07:44.970 --> 00:07:47.230 That's maybe more of a split decision. Quests 00:07:47.230 --> 00:07:50.069 is definitely the champ for pure, rapid prototyping. 00:07:50.129 --> 00:07:52.689 Getting a clean React code base for a UI idea, 00:07:52.870 --> 00:07:56.129 super fast, minimal fuss, nothing seems faster 00:07:56.129 --> 00:07:58.689 for that. But yeah, the premium tools still justify 00:07:58.689 --> 00:08:01.430 their cost with those endgame features. More 00:08:01.430 --> 00:08:03.670 advanced integrations like built -in database 00:08:03.670 --> 00:08:05.949 connections, authentication handling, better 00:08:05.949 --> 00:08:07.730 features for teams working together. And the 00:08:07.730 --> 00:08:09.569 enterprise stuff, right. Security guarantees, 00:08:09.790 --> 00:08:12.290 dedicated support contracts. That's still the 00:08:12.290 --> 00:08:15.310 stronghold for the paid tools. So if you look 00:08:15.310 --> 00:08:17.470 past the specific features, what's the biggest 00:08:17.470 --> 00:08:20.689 advantage the paid tools still hold on to? Probably 00:08:20.689 --> 00:08:23.110 those advanced integrations, collaboration tools, 00:08:23.230 --> 00:08:26.050 and that robust enterprise level support structure. 00:08:26.550 --> 00:08:29.250 Midroll sponsor read, to be inserted here by 00:08:29.250 --> 00:08:31.449 the writer. Okay, let's pivot a bit and think 00:08:31.449 --> 00:08:33.789 about the bigger picture here. The strategic 00:08:33.789 --> 00:08:36.289 angle of this whole thing, this combination free, 00:08:36.529 --> 00:08:39.750 plus open source, plus genuinely high quality 00:08:39.750 --> 00:08:42.870 performance. That mix really does change the 00:08:42.870 --> 00:08:45.210 game. Yeah. For an individual developer, it's 00:08:45.210 --> 00:08:48.529 just liberating. Zero cost to learn and experiment. 00:08:48.669 --> 00:08:50.610 You're not burning expensive credits. You get 00:08:50.610 --> 00:08:52.490 full control. You can actually modify the tool 00:08:52.490 --> 00:08:54.889 if you want. And absolutely no vendor lock -in. 00:08:54.990 --> 00:08:57.370 And for small teams, startups. This could be 00:08:57.370 --> 00:09:00.210 huge operationally. Wipes out those potentially 00:09:00.210 --> 00:09:03.350 hefty licensing fees right at the start. Plus, 00:09:03.409 --> 00:09:05.570 the lightweight stack might mean less technical 00:09:05.570 --> 00:09:07.750 debt down the line. Keeps development moving 00:09:07.750 --> 00:09:10.610 fast without breaking the bank. It's also just... 00:09:11.399 --> 00:09:13.519 A statement, isn't it? To the whole software 00:09:13.519 --> 00:09:15.600 world, it proves the open source community can 00:09:15.600 --> 00:09:17.919 build a truly credible professional alternative 00:09:17.919 --> 00:09:21.080 to these closed, expensive systems. It kind of 00:09:21.080 --> 00:09:23.639 levels the playing field. How does that democratization, 00:09:23.639 --> 00:09:26.340 that leveling, specifically help maybe newer 00:09:26.340 --> 00:09:28.559 developers or students just getting started? 00:09:28.879 --> 00:09:31.279 It just lowers the barrier completely, makes 00:09:31.279 --> 00:09:33.340 these powerful AI tools accessible everywhere, 00:09:33.580 --> 00:09:35.740 speeding up learning and innovation globally. 00:09:36.019 --> 00:09:37.919 Yeah. All right, so if someone's listening and 00:09:37.919 --> 00:09:39.080 thinking, okay, I want to try this. What's the 00:09:39.080 --> 00:09:41.700 process? The pre -flight check, the setup. It's 00:09:41.700 --> 00:09:44.139 actually dead simple, like five minutes. Download 00:09:44.139 --> 00:09:47.559 Quests, install it, go to OpenRouter, set it 00:09:47.559 --> 00:09:51.659 up to use DeepSeek V3 .1. That part's free. Point 00:09:51.659 --> 00:09:54.480 Quests to it, create your first project. Done. 00:09:54.909 --> 00:09:57.169 Okay. And then for flight school, actually using 00:09:57.169 --> 00:09:59.129 it effectively, some best practices. First, yeah, 00:09:59.210 --> 00:10:01.409 start simple. Don't try to build your magnum 00:10:01.409 --> 00:10:03.950 opus on day one. Focus on those single page UI 00:10:03.950 --> 00:10:06.450 heavy things where the AI really shines right 00:10:06.450 --> 00:10:09.750 now. Second, get good at prompting. Be really 00:10:09.750 --> 00:10:12.850 specific. Tell it exactly what UI, what UX you 00:10:12.850 --> 00:10:15.509 want. Mention responsive design needs. Even specify, 00:10:15.690 --> 00:10:17.889 you know, use modern React hooks or whatever 00:10:17.889 --> 00:10:19.830 patterns you prefer. And when you're optimizing 00:10:19.830 --> 00:10:21.690 those prompts, what's maybe one detail people 00:10:21.690 --> 00:10:24.120 forget to include? Maybe specifying accessibility 00:10:24.120 --> 00:10:27.440 requirements. Or naming a specific component 00:10:27.440 --> 00:10:30.299 library you like. Maybe Tailwind or Material 00:10:30.299 --> 00:10:34.360 UI. Those details help. And third, really lean 00:10:34.360 --> 00:10:36.960 into that multi -tab workflow. It's powerful. 00:10:37.259 --> 00:10:39.460 Build variations of a component in different 00:10:39.460 --> 00:10:42.019 tabs to compare. Or build out a whole library 00:10:42.019 --> 00:10:45.259 of reusable components, one per tab. It multiplies 00:10:45.259 --> 00:10:47.840 your output. Looking ahead, the roadmap for quests 00:10:47.840 --> 00:10:50.159 is where things get really interesting. Because 00:10:50.159 --> 00:10:52.240 this is being built out in the open by the community. 00:10:52.519 --> 00:10:54.639 And that can mean development moves incredibly 00:10:54.639 --> 00:10:57.139 fast. Yeah, the top priorities seem clear. Those 00:10:57.139 --> 00:10:58.980 foundational things to make it production ready. 00:10:59.179 --> 00:11:01.120 Getting environment variables working that's 00:11:01.120 --> 00:11:04.059 critical for API keys. Adding database integrations. 00:11:04.059 --> 00:11:07.059 Maybe Supabase first. And solid project import 00:11:07.059 --> 00:11:10.370 -export. Whoa. Just imagine scaling that community 00:11:10.370 --> 00:11:12.830 speed, right? Adding those key features maybe 00:11:12.830 --> 00:11:14.950 faster than a big corporate team could even plan 00:11:14.950 --> 00:11:17.389 them? That potential for rapid, sustained improvement? 00:11:17.669 --> 00:11:19.470 That's the real long -term threat to the established 00:11:19.470 --> 00:11:21.769 players. And then looking a bit further out, 00:11:21.870 --> 00:11:24.590 medium -term stuff people are asking for. Things 00:11:24.590 --> 00:11:26.610 like collaborative editing so teams can work 00:11:26.610 --> 00:11:28.789 in it together. More deployment options pushing 00:11:28.789 --> 00:11:31.590 straight to Vercel or Netlify. Maybe a library 00:11:31.590 --> 00:11:33.730 of shared app templates. And it's worth remembering, 00:11:33.929 --> 00:11:36.649 if you use it, you can influence this. Go star 00:11:36.649 --> 00:11:40.289 the project on GitHub. submit really clear, detailed 00:11:40.289 --> 00:11:43.269 feature requests, explain the why, and if you 00:11:43.269 --> 00:11:45.970 code, maybe even contribute directly to speed 00:11:45.970 --> 00:11:49.269 things up. So, boiling it down, what's the single 00:11:49.269 --> 00:11:51.950 most urgent thing on that roadmap to get Quests 00:11:51.950 --> 00:11:54.190 ready for real production apps? Definitely support 00:11:54.190 --> 00:11:56.610 for environment variables. That unlocks all the 00:11:56.610 --> 00:11:58.990 external API integrations. It's the essential 00:11:58.990 --> 00:12:01.309 infrastructure piece. So let's recap the big 00:12:01.309 --> 00:12:05.610 idea here. Quests paired with DeepSeq v3 .1 feels 00:12:05.610 --> 00:12:08.129 like... A pivotal moment. Yeah. It's offering 00:12:08.129 --> 00:12:11.370 a genuine free open source path for serious development 00:12:11.370 --> 00:12:13.769 without really compromising on the quality or 00:12:13.769 --> 00:12:15.429 the performance you get. Yeah, the direction 00:12:15.429 --> 00:12:18.009 seems pretty clear. Open source AI tools are 00:12:18.009 --> 00:12:20.710 catching up fast, achieving parity with the premium 00:12:20.710 --> 00:12:23.289 stuff. This is about giving developers real choice, 00:12:23.409 --> 00:12:25.649 making these tools accessible to everyone, everywhere. 00:12:26.009 --> 00:12:28.009 So the simple recommendation we took from the 00:12:28.009 --> 00:12:30.269 sources, if you're currently paying for Bolt 00:12:30.269 --> 00:12:33.330 or maybe Lovable, just spend a weekend with Quests. 00:12:33.960 --> 00:12:36.620 Seriously, kick the tires on that rapid prototyping 00:12:36.620 --> 00:12:39.120 workflow. See how it feels. The time you save 00:12:39.120 --> 00:12:42.019 might genuinely surprise you. It feels like the 00:12:42.019 --> 00:12:44.820 open source revolution in AI development is really 00:12:44.820 --> 00:12:47.379 here. It's started. The only real question left 00:12:47.379 --> 00:12:49.940 for you, the developer listening, is can you 00:12:49.940 --> 00:12:53.120 afford not to be exploring this? Thanks for diving 00:12:53.120 --> 00:12:55.059 deep into these sources with us today. Lots to 00:12:55.059 --> 00:12:57.019 think about. It really is an exciting time to 00:12:57.019 --> 00:12:58.860 be building things. We'll talk to you next time.