WEBVTT 00:00:00.000 --> 00:00:02.680 It used to be that if you had an idea, you know, 00:00:02.680 --> 00:00:05.299 for a software business, there was this huge 00:00:05.299 --> 00:00:07.919 toll booth right in your way. A very expensive 00:00:07.919 --> 00:00:10.759 one. Exactly. You either had to learn to code 00:00:10.759 --> 00:00:13.279 yourself, which is a six -month commitment. Easy. 00:00:13.660 --> 00:00:17.320 Minimum. Or you had to hire a dev team. And a 00:00:17.320 --> 00:00:19.879 decent agency, let's be real, starts at, what, 00:00:19.920 --> 00:00:21.920 two grand a month? Oh, and goes up fast. That 00:00:21.920 --> 00:00:24.359 was the barrier to entry. That financial wall 00:00:24.359 --> 00:00:27.480 acted as a filter. Mm -hmm. I mean, it killed 00:00:27.480 --> 00:00:29.679 a lot of bad ideas, sure, but it also must have 00:00:29.679 --> 00:00:31.719 killed so many brilliant ones that just didn't 00:00:31.719 --> 00:00:33.560 have the cash to get off the ground. But now? 00:00:34.060 --> 00:00:36.640 Here we are. It's February 2026. And reading 00:00:36.640 --> 00:00:40.219 this guide from Max Anne at AI Fire, it feels 00:00:40.219 --> 00:00:42.320 like that whole landscape hasn't just shifted 00:00:42.320 --> 00:00:44.299 a little bit. It's totally inverted. It really 00:00:44.299 --> 00:00:46.380 has. We're actually looking at a world where 00:00:46.380 --> 00:00:49.200 you can build software, automate these crazy 00:00:49.200 --> 00:00:52.500 workflows, and even clone your own presence for 00:00:52.500 --> 00:00:55.299 video. For zero dollars. The zero dollar tech 00:00:55.299 --> 00:00:58.560 stack. And honestly, the price isn't even the 00:00:58.560 --> 00:01:01.979 most shocking part. It's the speed. But there 00:01:01.979 --> 00:01:05.099 is a catch. Because the tools are free, the amount 00:01:05.099 --> 00:01:07.700 of noise is just deafening. You see it every 00:01:07.700 --> 00:01:11.579 day, 50 new game -changing AI tools launching. 00:01:11.680 --> 00:01:13.620 Which is why we're doing this. We're not just 00:01:13.620 --> 00:01:16.719 reading some huge list of shiny toys. We're going 00:01:16.719 --> 00:01:19.200 to filter for the nine tools that actually pass 00:01:19.200 --> 00:01:22.040 what I call the real money test. Can you actually 00:01:22.040 --> 00:01:23.659 build a business with them? So we're going to 00:01:23.659 --> 00:01:26.159 cover the foundation, then the builder, the Polish, 00:01:26.379 --> 00:01:29.319 the clone, and finally the engine. And we really 00:01:29.319 --> 00:01:31.780 need to spend some time on the clone, partly. 00:01:31.780 --> 00:01:34.140 I was thinking the same thing. Because that idea 00:01:34.140 --> 00:01:37.879 that you can digitally duplicate yourself to 00:01:37.879 --> 00:01:40.340 work while you're asleep, it's the biggest productivity 00:01:40.340 --> 00:01:43.840 unlock, but it's also, frankly, the most unsettling 00:01:43.840 --> 00:01:45.739 part of this whole thing. It feels like we've 00:01:45.739 --> 00:01:48.079 fully crossed the uncanny valley. But before 00:01:48.079 --> 00:01:50.200 we get to the sci -fi stuff, we have to start... 00:01:50.379 --> 00:01:52.260 At the beginning. The brain of the operation. 00:01:52.579 --> 00:01:54.879 A foundation. Come on, right? The chatbots. OK, 00:01:54.959 --> 00:01:56.739 so most people listening probably have a chat 00:01:56.739 --> 00:01:59.579 GPT tab open right now. Yeah. It's the default. 00:01:59.760 --> 00:02:03.319 It's like the Kleenex of AI. But the source argues 00:02:03.319 --> 00:02:05.760 that just defaulting to chat GPT for everything 00:02:05.760 --> 00:02:10.280 is actually. A strategic mistake. It's a huge 00:02:10.280 --> 00:02:13.139 one. You have to stop thinking of AI as one thing. 00:02:13.199 --> 00:02:15.300 Think of these models as like different employees 00:02:15.300 --> 00:02:18.000 with very specific PhDs. I like that. If you 00:02:18.000 --> 00:02:20.860 hire a creative writer to do your forensic accounting, 00:02:21.060 --> 00:02:22.759 you're going to have a bad time. You're going 00:02:22.759 --> 00:02:25.259 to go to jail. Probably. So the source breaks 00:02:25.259 --> 00:02:28.740 down five major players and they are not interchangeable. 00:02:28.919 --> 00:02:30.439 Okay, so let's walk through these employees. 00:02:30.780 --> 00:02:34.460 First up, ChatGPT from OpenAI. That's your general 00:02:34.460 --> 00:02:37.319 manager, the all -rounder. great at reasoning 00:02:37.319 --> 00:02:39.900 pretty good at writing you know it's reliable 00:02:39.900 --> 00:02:41.740 if you don't know who to ask you start there 00:02:41.740 --> 00:02:44.060 but it's a generalist exactly then you've got 00:02:44.060 --> 00:02:46.900 gemini from google and gemini is the the research 00:02:46.900 --> 00:02:49.719 analyst its superpower is just this massive context 00:02:49.719 --> 00:02:53.060 window and multimodal analysis you can throw 00:02:53.060 --> 00:02:56.960 a 50 -page pdf a messy spreadsheet and a youtube 00:02:56.960 --> 00:02:59.919 link at it and just say find the connection Wow. 00:03:00.159 --> 00:03:02.479 It sees and hears things in a way the others 00:03:02.479 --> 00:03:04.860 just can't yet. Okay. Then there's Claude from 00:03:04.860 --> 00:03:07.699 Anthropic. That's the logician. If you're writing 00:03:07.699 --> 00:03:09.800 code or you're trying to build a really complex 00:03:09.800 --> 00:03:13.319 legal argument, Claude tends to edge the others 00:03:13.319 --> 00:03:15.680 out. It hallucinates less on like logic puzzles. 00:03:15.979 --> 00:03:19.199 And then Llama and Grok. Llama from Meta is just 00:03:19.199 --> 00:03:22.210 the sprinter. It's all about speed. And Grok, 00:03:22.229 --> 00:03:24.310 the one that's built into X. Right. Think of 00:03:24.310 --> 00:03:27.110 Grok as your investigator who has real time access 00:03:27.110 --> 00:03:29.610 to the public square. If you need to know what 00:03:29.610 --> 00:03:31.409 people are feeling right now about some breaking 00:03:31.409 --> 00:03:35.229 news, Grok is tapped into that live fire hose 00:03:35.229 --> 00:03:38.189 of X data. So the others are working with a lag. 00:03:38.370 --> 00:03:40.490 Yeah, they're working off training data. Grok 00:03:40.490 --> 00:03:43.550 is the unfiltered, messy present. But here's 00:03:43.550 --> 00:03:45.969 the insight that really struck me. It's not just 00:03:45.969 --> 00:03:48.330 about picking the tool. It's how you use it. 00:03:48.560 --> 00:03:51.400 Max Anne outlines this strategy called the Reddit 00:03:51.400 --> 00:03:54.199 search workflow. Oh, I love this because it flips 00:03:54.199 --> 00:03:57.439 the whole AI workflow upside down. How so? Well, 00:03:57.439 --> 00:03:59.039 usually people just stare at the chat bot and 00:03:59.039 --> 00:04:01.400 go, give me the business idea. And it spits out 00:04:01.400 --> 00:04:04.419 something generic like start a coffee shop. Start 00:04:04.419 --> 00:04:07.060 a drop shipping business. Exactly. It's totally 00:04:07.060 --> 00:04:09.419 useless. The Reddit search strategy is different. 00:04:09.520 --> 00:04:11.780 You open up your chat bot and you say, search 00:04:11.780 --> 00:04:14.259 Reddit for the most common complaints first time 00:04:14.259 --> 00:04:16.500 business founders have. You're not asking for 00:04:16.500 --> 00:04:19.449 ideas. You're asking for pain points. You're 00:04:19.449 --> 00:04:22.449 asking for pain. Because Reddit is where people 00:04:22.449 --> 00:04:25.149 complain, honestly. You find a pattern, like 00:04:25.149 --> 00:04:27.750 founders really struggle to validate their ideas 00:04:27.750 --> 00:04:30.430 before they build. And that's a verifiable market 00:04:30.430 --> 00:04:32.990 gap. Precisely. You haven't written a single 00:04:32.990 --> 00:04:35.509 line of code, but you already know you're solving 00:04:35.509 --> 00:04:38.290 a real problem. So why do you think most people 00:04:38.290 --> 00:04:41.350 fail at this very first, most critical step? 00:04:41.639 --> 00:04:43.639 Because they treat AI like it's entertainment 00:04:43.639 --> 00:04:46.699 or a magic wand, not a high -powered research 00:04:46.699 --> 00:04:48.920 assistant to solve problems. They treat it as 00:04:48.920 --> 00:04:51.399 a toy, not a tool. That's it, exactly. Okay. 00:04:51.459 --> 00:04:53.800 So we found our problem. We know what pain we 00:04:53.800 --> 00:04:55.720 want to solve. Now we have to actually build 00:04:55.720 --> 00:04:58.240 the thing. Which brings us to tool number two, 00:04:58.480 --> 00:05:02.319 lovable. And the promise here is, well, it's 00:05:02.319 --> 00:05:05.339 huge, building functional web apps just using 00:05:05.339 --> 00:05:09.230 plain English. No code. This is a massive shift. 00:05:09.430 --> 00:05:12.189 I mean, not long ago, your options were learn 00:05:12.189 --> 00:05:14.509 to code for half a year or pay a developer a 00:05:14.509 --> 00:05:17.750 fortune. Right. Lovable is part of this new wave 00:05:17.750 --> 00:05:20.430 where you just describe the software and it just 00:05:20.430 --> 00:05:24.529 builds it. The example in the source is a business 00:05:24.529 --> 00:05:27.509 idea validator. I saw that. You literally just 00:05:27.509 --> 00:05:30.050 type a prompt. You type, create a tool where 00:05:30.050 --> 00:05:32.350 I can input a business concept, have it evaluated, 00:05:32.550 --> 00:05:36.300 and then give it a grade from A to F. And lovable 00:05:36.300 --> 00:05:39.300 doesn't just write text back. It generates the 00:05:39.300 --> 00:05:42.000 actual forms, the logic, the scoring rules, the 00:05:42.000 --> 00:05:45.180 entire user interface. You are watching a product 00:05:45.180 --> 00:05:47.660 appear in front of you. I have to admit something 00:05:47.660 --> 00:05:50.579 here. Go on. I've tried tools like this, and 00:05:50.579 --> 00:05:52.839 I still get so frustrated. I fall right into 00:05:52.839 --> 00:05:55.360 that garbage in, garbage out trap. I'll type 00:05:55.360 --> 00:05:57.720 something vague like, make it better. And then 00:05:57.720 --> 00:05:59.639 I get annoyed when the AI can't. you know, read 00:05:59.639 --> 00:06:02.000 my mind. That is the universal struggle right 00:06:02.000 --> 00:06:03.740 now. I mean, the source calls it the reality 00:06:03.740 --> 00:06:06.360 check. One prompt is never enough. You have to 00:06:06.360 --> 00:06:08.379 iterate. But a bad prompt is make it better. 00:06:08.500 --> 00:06:11.420 Right. A good prompt is change the rating so 00:06:11.420 --> 00:06:14.300 A is worst and F is best. Add a confidence score 00:06:14.300 --> 00:06:16.740 and make the primary color blue. You have to 00:06:16.740 --> 00:06:19.100 be the manager. Give specific directions. So 00:06:19.100 --> 00:06:23.199 is coding dead or is it just different? It's 00:06:23.199 --> 00:06:25.420 not dead, but it's becoming mechanical. You're 00:06:25.420 --> 00:06:28.540 no longer managing the syntax, you know, the 00:06:28.540 --> 00:06:30.959 semicolons and brackets. You're managing the 00:06:30.959 --> 00:06:33.560 logic and the outcome. You manage the logic, 00:06:33.720 --> 00:06:36.120 not the syntax. Yeah. I really like that. It 00:06:36.120 --> 00:06:37.639 just frees you up to think about the product. 00:06:37.939 --> 00:06:40.540 So let's move on to speed and polish. Because 00:06:40.540 --> 00:06:42.579 even if the software works, if it looks like 00:06:42.579 --> 00:06:45.279 some generic AI template, nobody's going to trust 00:06:45.279 --> 00:06:48.680 it. The source points to two tools here, WhisperFlow 00:06:48.680 --> 00:06:52.089 and 21st .dev. Let's start with Whisper. This 00:06:52.089 --> 00:06:54.829 is about pure velocity. The math is really simple. 00:06:55.029 --> 00:06:57.209 Humans type at, what, 50 or 60 words a minute? 00:06:57.490 --> 00:06:59.730 But we speak at around 250 words per minute. 00:06:59.829 --> 00:07:02.250 So it's a 4 or 5x speed increase. Theoretically, 00:07:02.250 --> 00:07:03.870 yes, if you can talk to your computer instead 00:07:03.870 --> 00:07:06.480 of typing. But voice -to -text is usually awful. 00:07:06.579 --> 00:07:09.720 It catches all the ums and eras and the weird 00:07:09.720 --> 00:07:12.139 pauses. And that's where WhisperFlow is different. 00:07:12.279 --> 00:07:14.800 It doesn't just transcribe. It edits as it goes. 00:07:14.920 --> 00:07:17.699 It strips out the filler words. It adds punctuation. 00:07:17.699 --> 00:07:19.759 It makes paragraphs. So you can just ramble. 00:07:19.759 --> 00:07:22.519 And what comes out is a clean prompt for lovable. 00:07:22.860 --> 00:07:26.360 That is a huge workflow unlock. Okay, so what 00:07:26.360 --> 00:07:30.180 about 21st .def? This takes on the generic AI 00:07:30.180 --> 00:07:32.120 look. You know what I mean, right? Those websites 00:07:32.120 --> 00:07:33.759 that all just kind of look the same because they 00:07:33.759 --> 00:07:36.819 use the same default template. 21st .dev is a 00:07:36.819 --> 00:07:39.959 library of high quality design components. So 00:07:39.959 --> 00:07:42.319 if you need an animated robot or a cool pricing 00:07:42.319 --> 00:07:45.379 table, you find it there, copy it, and paste 00:07:45.379 --> 00:07:47.920 it into lovable. You can assemble a custom look 00:07:47.920 --> 00:07:49.939 without being a designer. But the source does 00:07:49.939 --> 00:07:51.699 make a good point here. It says for your core 00:07:51.699 --> 00:07:55.680 brand identity, the logo, the vibe, you should 00:07:55.680 --> 00:07:58.310 probably still hire a human. Yes, a hybrid approach. 00:07:58.550 --> 00:08:01.470 Use AI for the heavy lifting, the code, the layout, 00:08:01.689 --> 00:08:04.089 but then go to Fiverr, spend 50 bucks and get 00:08:04.089 --> 00:08:06.230 a real designer to do the logo. Because if your 00:08:06.230 --> 00:08:08.470 logo looks cheap, your whole product feels cheap. 00:08:08.689 --> 00:08:11.029 So why does design matter so much if the code 00:08:11.029 --> 00:08:14.189 underneath is solid? Because users judge credibility 00:08:14.189 --> 00:08:17.290 in a split second based on visual polish. If 00:08:17.290 --> 00:08:20.089 it looks broken, they assume it is broken. It's 00:08:20.089 --> 00:08:23.050 the halo effect. The visuals signal competence. 00:08:23.310 --> 00:08:25.569 That's it. Speaking of visuals, let's talk about 00:08:25.569 --> 00:08:29.360 tool number five. Nanobanana. Great name. Which, 00:08:29.480 --> 00:08:31.839 aside from the name, is an image generator that's 00:08:31.839 --> 00:08:34.159 powered by Gemini. Right. And the theme here 00:08:34.159 --> 00:08:37.120 is the same iteration. The source uses creating 00:08:37.120 --> 00:08:39.940 a logo as the example. The first prompt, it gives 00:08:39.940 --> 00:08:42.879 you something. Okay. A little busy. Yeah. But 00:08:42.879 --> 00:08:44.539 then you refine it. You say, make it simplistic, 00:08:44.980 --> 00:08:48.299 clean, Apple style. And then suddenly the second 00:08:48.299 --> 00:08:50.919 or third version is professional grade. It seems 00:08:50.919 --> 00:08:52.860 like people expect the AI to be a mind reader. 00:08:53.210 --> 00:08:55.789 They do. And that's just the wrong mental model. 00:08:55.870 --> 00:08:58.009 You're not ordering a pizza here. You're sculpting. 00:08:58.009 --> 00:08:59.990 You have to chip away at the prompt until the 00:08:59.990 --> 00:09:02.129 image appears. So what's the wrong mental model 00:09:02.129 --> 00:09:04.590 people have here? Expecting perfection on the 00:09:04.590 --> 00:09:06.870 first click. It's an iterative process. Simple 00:09:06.870 --> 00:09:08.990 as that. Okay, before we get to the cloning part, 00:09:09.110 --> 00:09:10.470 which I know is what everyone's waiting for, 00:09:10.570 --> 00:09:13.690 we need to take a quick break. All right, we're 00:09:13.690 --> 00:09:16.309 back. And this is where things get really interesting. 00:09:16.909 --> 00:09:20.210 Segment five, the clone. We're talking about 00:09:20.210 --> 00:09:23.269 Heijin and Eleven Labs. This is the kind of stuff 00:09:23.269 --> 00:09:24.809 that just feels like it shouldn't be possible 00:09:24.809 --> 00:09:29.409 for zero dollars. It is genuinely wild. So the 00:09:29.409 --> 00:09:32.230 process is basically two steps. First, you use 00:09:32.230 --> 00:09:34.450 Hagen. You record about two minutes of video 00:09:34.450 --> 00:09:36.710 of yourself just talking to the camera. Hagen 00:09:36.710 --> 00:09:39.110 analyzes that and creates a digital avatar that 00:09:39.110 --> 00:09:41.860 moves like you, blinks like you. Looks exactly 00:09:41.860 --> 00:09:44.600 like you. And step two is the voice. That's 11 00:09:44.600 --> 00:09:46.820 Labs. You record a few minutes of your audio 00:09:46.820 --> 00:09:49.039 and it clones your voice. Then you combine the 00:09:49.039 --> 00:09:51.960 two. And suddenly you can type any text script 00:09:51.960 --> 00:09:54.360 and your digital clone will speak it on video 00:09:54.360 --> 00:09:57.240 with your voice, your face, perfectly lip synced. 00:09:57.419 --> 00:10:01.039 Whoa. I mean, it's just imagine scaling that. 00:10:01.120 --> 00:10:03.539 You could have a billion views on an ad campaign 00:10:03.539 --> 00:10:06.419 running 24 -7 delivered by you. While you are 00:10:06.419 --> 00:10:08.600 literally asleep in bed. And nobody would know 00:10:08.600 --> 00:10:11.250 it's not actually you. That's the power. The 00:10:11.250 --> 00:10:13.929 use cases are just massive. You can test 50 ad 00:10:13.929 --> 00:10:15.750 scripts in 10 minutes without ever turning on 00:10:15.750 --> 00:10:17.909 your camera. You can keep posting content while 00:10:17.909 --> 00:10:20.509 you take a month off. It decouples your physical 00:10:20.509 --> 00:10:23.230 self from your digital output. So what is the 00:10:23.230 --> 00:10:25.809 ultimate utility of cloning yourself? It's exactly 00:10:25.809 --> 00:10:28.490 that. Decoupling your physical time from your 00:10:28.490 --> 00:10:31.629 content output. You become scalable. You become 00:10:31.629 --> 00:10:34.850 scalable. That's a powerful thought. Okay, let's 00:10:34.850 --> 00:10:36.610 move to the machinery behind it all. Segment 00:10:36.610 --> 00:10:41.820 six, the engine and the scribe. We have NANN 00:10:41.820 --> 00:10:44.919 and Granola. And NANN is probably the most technical 00:10:44.919 --> 00:10:47.139 tool on this list. It seems like it. Think of 00:10:47.139 --> 00:10:50.700 NANN as the glue, the automation engine. So say 00:10:50.700 --> 00:10:53.500 a customer asks a question on your website chatbot. 00:10:53.580 --> 00:10:55.899 You can set up NANN to take that question, run 00:10:55.899 --> 00:10:58.480 it through an AI, have the AI check the weather 00:10:58.480 --> 00:11:00.940 or your inventory, format a reply, and send it 00:11:00.940 --> 00:11:03.320 back. It automates the entire logic flow. Correct. 00:11:03.559 --> 00:11:05.240 And the source makes a big deal about it being 00:11:05.240 --> 00:11:07.720 open source and self -hostable. Right. Why does 00:11:07.720 --> 00:11:10.120 that matter so much? It's all about privacy. 00:11:10.360 --> 00:11:13.100 A lot of automation tools are just black boxes. 00:11:13.179 --> 00:11:15.480 You have no idea where your data is going. With 00:11:15.480 --> 00:11:18.159 NENA, you can run it on your own server. Your 00:11:18.159 --> 00:11:21.000 data stays yours. Which is a perfect lead -in 00:11:21.000 --> 00:11:24.360 to Granola, the meeting notes tool. We've all 00:11:24.360 --> 00:11:27.019 been in those calls where a bot joins and announces 00:11:27.019 --> 00:11:29.820 this meeting is being recorded. It kills the 00:11:29.820 --> 00:11:32.360 vibe instantly. It does. Grinnell is different 00:11:32.360 --> 00:11:34.879 because it runs locally on your machine. It doesn't 00:11:34.879 --> 00:11:37.240 join the call as a bot. It just listens to your 00:11:37.240 --> 00:11:40.139 system audio, transcribes it, and summarizes 00:11:40.139 --> 00:11:42.080 it using AI that's running on your computer. 00:11:42.399 --> 00:11:44.940 So no data ever gets sent to the cloud. Exactly. 00:11:45.059 --> 00:11:47.799 It's up every single time. With Notebook LM, 00:11:47.980 --> 00:11:51.320 you just upload your PDFs, your... Google Docs, 00:11:51.320 --> 00:11:53.080 your website links, you dump it all in once. 00:11:53.259 --> 00:11:56.600 And the AI basically learns your business. It 00:11:56.600 --> 00:11:59.139 creates a grounded knowledge base. So now when 00:11:59.139 --> 00:12:01.820 you say write me a sales email, it already knows 00:12:01.820 --> 00:12:04.340 who you are, what you sell and who you sell to. 00:12:04.539 --> 00:12:06.620 It just connects the dots for you. And I had 00:12:06.620 --> 00:12:08.919 that audio overview feature, which is kind of 00:12:08.919 --> 00:12:10.899 meta because it basically turns your documents 00:12:10.899 --> 00:12:13.220 into a little podcast like this one. It does. 00:12:13.320 --> 00:12:15.320 It's great for listening to research, but we 00:12:15.320 --> 00:12:17.659 should add the warning that. source brings up 00:12:17.659 --> 00:12:20.440 about business model. Right. The old saying, 00:12:20.539 --> 00:12:22.960 if you don't pay, you are the product. That's 00:12:22.960 --> 00:12:25.860 it. Google tools, free tiers. They often use 00:12:25.860 --> 00:12:27.899 your data to train their models. If you're working 00:12:27.899 --> 00:12:29.620 on something top secret, you have to check the 00:12:29.620 --> 00:12:32.039 privacy policy. Paid tools sell subscriptions. 00:12:32.399 --> 00:12:35.600 Free tools often sell your data. So after all 00:12:35.600 --> 00:12:38.919 this, do these tools guarantee a successful business? 00:12:39.579 --> 00:12:41.879 Absolutely not. These tools just lower the barrier 00:12:41.879 --> 00:12:44.700 to entry. But humans, you know, we still have 00:12:44.700 --> 00:12:46.899 to create the actual value. You still have to 00:12:46.899 --> 00:12:49.740 solve a real problem. The tools are the lever, 00:12:49.820 --> 00:12:51.580 but you still have to be the one to push it. 00:12:51.740 --> 00:12:54.659 Well said. Correct. All right. Let's wrap this 00:12:54.659 --> 00:12:56.639 up. We've gone through a lot. We started with 00:12:56.639 --> 00:12:58.600 the foundation, choosing the right chatbot for 00:12:58.600 --> 00:13:01.419 the job, then building with lovable. Getting 00:13:01.419 --> 00:13:04.179 the prompts right. Then speed with whisper, Polish 00:13:04.179 --> 00:13:07.580 with 21st .dev, and visuals with nano banana. 00:13:08.029 --> 00:13:09.850 And then the sci -fi stuff, cloning yourself 00:13:09.850 --> 00:13:12.669 with Hagen and Eleven Labs and all the plumbing 00:13:12.669 --> 00:13:15.470 underneath with Engen and Granola. I think the 00:13:15.470 --> 00:13:18.769 big idea for me is that the excuse of I can't 00:13:18.769 --> 00:13:22.370 afford to build it, that's just gone. The cost 00:13:22.370 --> 00:13:25.250 is zero. The barrier isn't money anymore. It's 00:13:25.250 --> 00:13:28.190 just resourcefulness. That is it. But my final 00:13:28.190 --> 00:13:29.649 thought, my little warning for everyone listening 00:13:29.649 --> 00:13:32.169 is this. Don't just become a collector of these 00:13:32.169 --> 00:13:34.990 tools. It's so easy to just bookmark them and 00:13:34.990 --> 00:13:37.029 feel productive. That little dopamine hit of 00:13:37.029 --> 00:13:39.850 saving for later. Exactly. Don't do that. Pick 00:13:39.850 --> 00:13:42.129 one problem. Maybe use that Reddit search method 00:13:42.129 --> 00:13:44.870 we talked about and just build one thing. Just 00:13:44.870 --> 00:13:48.029 one. Because a mediocre tool in the hands of 00:13:48.029 --> 00:13:50.529 a builder is worth infinitely more than the best 00:13:50.529 --> 00:13:52.590 tool in the hands of a collector. That's the 00:13:52.590 --> 00:13:55.309 challenge. Pick a problem, open the toolkit, 00:13:55.529 --> 00:13:58.110 and build something real. Thanks for listening 00:13:58.110 --> 00:14:00.129 to The Deep Dive. We'll see you next time.