07:22The Next New Thing · Interview
We ran Fable through 600+ tests. It's amazing
Zapier's Automation Bench ran Claude Fable 5.0 against hundreds of realistic business workflows — here's what the numbers actually mean.
June 9thThe argument in one line.
Claude Fable 5 collapses the gap between imagining an app and having one so completely that the new bottleneck is knowing what you want to build, not knowing how to build it.
Read if. Skip if.
READ IF YOU ARE…
- A non-technical founder or operator who wants to replace SaaS subscriptions with tools they own and needs proof the moment has arrived.
- A regular AI builder who wants a curated scan of what is genuinely new with Fable 5 versus what is just incrementally better.
- Anyone running long autonomous agent tasks who wants concrete patterns for thought-partner prompting, clarifying questions, and verification criteria.
SKIP IF…
- A software engineer looking for deep technical implementation details — the hosts are evaluators reacting to demos, not engineers explaining architecture.
- You have already watched the primary source clips (Dan Shipper, Alex Finn, Nate Herk) — this show curates rather than originates.
The full version, fast.
Claude Fable 5 lands as a genuine step change: single prompts — sometimes one sentence — produce apps indistinguishable from funded SaaS on first look. The hosts walk through 12 community builds before landing on three prompting patterns that let Fable run deep without back-and-forth: use it as a devil's-advocate thought partner before building, have it ask clarifying questions so requirements are extracted rather than guessed, and give it hard verification criteria (screenshots, tests, specific behaviors) so it checks its own work. Andrew closes by demoing his own screenshot markup tool built and deployed to Cloudflare in a single session — proof the pattern works for anyone who can own their workflow decisions.
Free for members
Chat with this breakdown — free.
Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.
Create a free account →Where the time goes.
00:00 – 00:48
01 · Intro + promise
Three example builds rattled off as proof of concept, Zapier sponsor, format established.
00:48 – 02:10
02 · Lovable clone in 2 prompts
Riley Brown clones the Lovable mobile app — Prompt 1 builds it, Prompt 2 redesigns via screenshot. Side-by-side with the real app.
02:10 – 02:40
03 · Synthesizer / DAW (Brian build)
Brian built a full digital audio workstation with mixer and sequencer from one sentence in 15 minutes.
02:40 – 03:35
04 · Infinite 3D Library of Babel
Dan Shipper's single-prompt 3D browser game from a Borges story. Ran 3-4 hours autonomously, self-looping.
03:35 – 06:55
05 · Productivity app: Kanban + calendar + Pomodoro
Alex Finn's full productivity suite. Key pattern: asked Fable to ask clarifying questions before building.
06:55 – 07:50
06 · Custom markdown editor
Kieran Klaassen's publicly deployed markdown editor. Hosts reflect on the new baseline complexity.
07:50 – 08:55
07 · 3D virtual drum kit
Bijan Bowen's keyboard-playable 3D drum simulator with wood and brass effects on a circular carpet stage.
08:55 – 09:25
08 · Studio lighting simulator
Josh Daws built a tool for planning office/studio lighting layouts from one small prompt.
09:25 – 10:20
09 · McKinsey-style AI research report
Riley Brown uploaded a McKinsey report and had Fable clone the style for a forward-looking AI presentation with graphics.
10:20 – 11:20
10 · AI Monopoly with multiplayer
Justine Moore's AI-themed Monopoly. Fable added multiplayer link-sharing without being asked.
11:20 – 12:05
11 · Skiing physics game
Alex Ermolov's skiing game. Hosts debate how impressed to be on launch day.
12:05 – 13:00
12 · Turbo Kart (Mario Kart-style)
Single shot, 15 minutes, including game music in the flavor of Mario Kart. Andrew ties it to a grammar quiz game for his kids.
13:00 – 17:55
13 · Token costs + three prompting tips
Warning: one friend burned through the $200/month plan in one request. Tips: lower reasoning tiers, thought partner mode, clarifying questions, hard verification criteria.
17:55 – 21:45
14 · Andrew's Markup — live demo
Andrew demos his custom screenshot markup tool built with VS Code + Claude, deployed to Cloudflare. Voice-dictated requirements, Fable asked clarifying questions, deploy worked first try.
Lines worth screenshotting.
- A single sentence to Claude Fable 5 produced a fully playable 3D browser game based on a Borges short story after 3-4 hours of autonomous looping with no human intervention.
- Fable added multiplayer link-sharing to an AI Monopoly game without being asked — it inferred that you would want to play with others.
- Setting reasoning to medium or low for simpler questions can prevent the token burn that wiped out one developer's entire $200-per-month plan in a single request.
- Build your own markdown editor is the new build your own to-do list — the baseline complexity of what a single AI session produces has moved up a full tier.
- Having Fable ask you clarifying questions before it builds extracts requirements you did not know you had — one host learned he wanted exactly 8 color options only because Fable asked.
- Long-running autonomous tasks only work when you give hard verification criteria — not check your work but you are not done until you can confirm with screenshots and passing tests.
- The $10B Lovable app was two-shot cloned by a non-engineer using Claude Fable 5 and was indistinguishable from the original on first glance.
- Playing devil's advocate is a distinct prompting mode: ask Fable to find holes in your plan before you build anything, then argue back — this is different from asking it to execute.
- Dictating requirements as a voice note and having Fable organize them into a spec is a legitimate fast path from idea to deployed product.
- The gap between imagining an app and having an app is now so narrow that the only real constraint is knowing what you actually want.
Three patterns that make Fable actually finish the job.
WHAT TO LEARN
Fable 5 can run for hours and build things that would have taken a team a week — but only if you give it the right scaffolding before it starts.
- Use the model as a devil's-advocate thought partner before issuing any build command — make it argue against your plan, not just execute it.
- Prompt it to ask YOU clarifying questions before it writes a single line; this extracts requirements you did not know you had and would have discovered painfully later.
- Hard verification criteria are not optional on long autonomous tasks — tell it exactly what done looks like (screenshots, passing tests, specific UI states) or it will decide for itself.
- Fable has adjustable reasoning tiers; using medium or low for simple sub-tasks inside a complex build can prevent the kind of token burn that drains a $200/month plan in one request.
- The new skill is knowing what you want, not knowing how to build it — the model will handle the how, but you are still the only one who knows the what.
- Building your own tools (markdown editor, screenshot markup, notes app) is now faster than learning a SaaS product well enough to customize it to your workflow.
Things they pointed at.
18:20toolMarkup Hero
Lines you could clip.
05:45
“It's the lack of back and forth that makes it feel magical. It just gets it.”
13:48
“A friend burned through all his tokens in the $200 a month plan in one request.”
20:38
“You don't have to learn how to code, but you do still need to make key decisions.”
Word for word.
Read-along
Don't just watch it. Burn it in.
See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.
00:00Cloud Fable is so good. One guy said create a productivity app for me, and he got something on the level of Notion. Another person said create Mario Kart for me, and he got a fully playable game.
00:09Another person said I want you to create Lovable for me. Now Lovable is a $10,000,000,000 company that creates an app that you can build other apps with.
00:18He just two shot at it, told Claude to make it, and he got it. We're gonna show you all these different things that are created. We'll give you prompts that will let you create it, and I'll even show you at the end what I built myself.
00:28Oh, and we'll give you tips for how to not burn through all your tokens, but actually use it to get the best value at the lowest cost. Let's get right into it. Presented by Zapier, the AI automation company.
00:41Claude Fable or Mythos
00:44is absolutely insane. On the left here, we have the actual level lovable mobile app.
00:50And here, we have a lovable mobile app that I built in two prompts with Claude Fable five. This is the chat thread. Here's prompt one, and then I just said run it on simulator two and then go.
01:03So that just counts as one prompt. And then it did some stuff, did some stuff, and then I just screenshotted all the screens on Lovable. I just need you to redesign it to look exactly like this.
01:15Brian, this is incredible. We have seen so many creations with Fable.
01:19We're gonna go through a bunch of them today. And what this guy did, I wanted it first because he didn't just create a new app. He created an app that lets you build just about any other app, and he actually demoed it, and it works really well.
01:31I'll show it in a minute. These are exciting. Recreated Lovable.
01:34Recreated an app builder. It it this is like a new form of content. I'm scrolling,
01:38like, Fable first reactions of people, like, one shotting incredible apps. And the thing is that they're inspiring me. I was able to build, and I'd like to show everyone what I built because I saw what everyone else was building.
01:47And the cool thing about this is he built an app that builds other apps, and he told it, now build Notion. And so you can see on the left, it built Notion for him.
01:57This is Lovable. We know that. This is a a what is it?
02:00A $10,000,000,000 company is what Lovable is? On the right is the thing that he just created Look. Also told it to create Notion.
02:05Uh, I can literally hello.
02:08There.
02:09This was his clone of Lovable to then build Notion. Right? Like
02:15Alright. Here's you. You saw it and you said you created a a MIDI.
02:21What is this?
02:22I was just, like, you know, playing around with it. This morning, my first shot with it was a, like, a synthesizer, a music production, like a like a DAW, digital audio. Workstation's got a mixer, got a got a little sequencer in there, and that, you know, one shot, like, fifteen minutes, you know, one sentence.
02:39Okay. And yeah. Let's see.
02:42Crazy. Alright. This one I really love from Dan Shipper.
02:45This is the Infinite Library of Babel from the Borges story. It contains all of the books in the universe because BookBurg is strange. If you look, you can even go into bookmarks and click one of my articles after automation, and it finds it in the library.
02:58It's truly infinite. Look, I could go up the stairs. I can, like, look down.
03:04And he's gonna show the prompt that built This seems like it took a long time to make, right? Wrong. I made this entire thing in a single prompt with Fable five, the new model from Anthropic.
03:17Like like, literally, let me show you. So this is a prompt. It's from four days ago or so.
03:21I got this model a little bit ahead of time. Read Jorge Luis Borges' The Library About k. Let me zoom in on what he's doing.
03:29And then plan and execute end to end a browser playable three d game in which the player has dropped in blah blah blah blah blah. Loop until it's done.
03:37Just wrote that, press enter, and it just went off. It read the story, and then it just ran and ran and ran. You can see it's, like, loop looping itself, it's checking its work.
03:45After three or four hours or so, done.
03:48We're gonna go and see something else. I want you to see another example of what Alex Finn did, and what I like is when people show the prompts. Here's his prompt.
03:56Paste it. Feel free to pause right here and grab it so you can do this with me. I wanna build out a personal productivity tool.
04:02I want it to include to do list, Kanban board calendar, comprehensive note taking app, Pomodoro timer, and even a little game I can play to take a break. Before building anything, please ask me any questions you have on this idea so we can build a fully fleshed out app. I'm gonna hit enter.
04:18And
04:19And we're gonna see what he did. And the thing that I like is that he did ask, please ask me questions ahead of time, and we're gonna go over what,
04:27uh, some of the best practices for using this. Here's the result. UI is a little ugly, but to be honest, I also answered the question to make it colorful and playful.
04:35So I guess it is very colorful and playful. Let's see what we got. Looks like it has absolutely everything in it.
04:41Let's do a quick task. Edit the Fable five video. Hit enter on that.
04:46Boom. It's on your plate. That's very, very nice looking.
04:49Let's check out the calendar, full calendar. Let's give it a a an event right now. Check video performance.
04:56We'll give it tomorrow. We'll give it a date for Thursday. We'll give it a color at the event.
05:01Boom there. This is a fully working calendar.
05:04What he's showing there is something that, like, could easily be done with, you know, Claude Opus or even, you know, earlier versions of it. I think what is really interesting now with with with Fable is the level of detail and the amount of how do you say it?
05:21Like like, correctness in the very first prompt. So we're it it seems like my first impression of it is, like, we're stripping away the need to have these ex extensive prompts and specs.
05:34I mean, like, the the synthesizer I made, some of these, like, crazy, like, three d rendering stuff. It's just going deep
05:40with without needing to go back and forth and and revise things. It's true. It's the lack of back and forth that makes it feel magical.
05:46It just gets it. Alright. Before I show the next one, let me ask you this, Now that you see this, do you think that this is gonna replace
05:52Notion for people? Are they gonna start building their own Notion? Are you gonna do that?
05:56I mean, that is the real interesting question. I have already started building my own internal tools. I don't pay for a project management tool anymore.
06:03Like, I don't I don't even pay for a Notes app. I built my own Notes app. You know, like, I'm doing that all the time, but I'm a builder.
06:09I've been a builder for many years. Right? And so and I love building with AI, of course.
06:14The real question now, as we're literally seeing people clone Notion.
06:20I I saw a guy clone Slack. I mean, and it's like an identical it you cannot tell the difference on first glance, you know.
06:28What does that mean for the business who would otherwise pay for a Slack or for a Notion, you know. Like, just for the sake of, like, killing your your SaaS subscription, probably not a good a good use case.
06:40But if if Notion is just not
06:43tailored to the way that your operation works or your pipeline or your all the nuances, then, like, yeah. So, like, spin up your own. You know?
06:51Alright. I'll show you what I built that is like that. But let's take a look at one more.
06:54I like this one a lot. This, I think, is a sign of things to come. What he did was he said, I wanna build my own markdown editor.
07:02In fact, he I love that he even included the link for anyone to try it. Let's take a look at it.
07:09Oh, this is not his markdown editor. We're gonna open it up right here. I'll here's his markdown editor.
07:17New document, markdown, create markdown, and here it is. It just works.
07:22This is me typing on your editor.
07:26And he called becoming the thing that that, like, all the builders are building their own markdown editors. I I've got mine that I've been using for a couple weeks, and, like, now that's becoming, like it used to be build your own to do list.
07:37Now now it's like just build your own markdown editor. You know? I could comment even.
07:41This is a Oh, I love it. Comment,
07:44and it goes over there. Nice. Incredible.
07:46I like this. Okay. Next.
07:51This guy, Bijan,
07:54he decided to create some fun. Course, is the virtual drum kit sim, which I have already run-in the web chat. I don't really like, this is always just fun to run as, like, the final test, I guess.
08:06So let's peek at it, and I do have the speaker on click anywhere to power on the audio engine. That's very, very good. Spatially, this may be one of the best arranged ones.
08:19More often than not, they're not necessarily properly arranged good. We can move around the scene. We have some wood effects.
08:24We have some brass effects, and then it's in a nice little circular carpet stage thing. We even have a stool.
08:31So let's check the important part. Does it slap? Alright.
08:35It does. Let me turn the speaker up here.
08:40Like that he can use keys to do it. This is just kinda like watching an elephant dance. It's not really useful, but it's interesting because it can do it.
08:49Alright. Studio lighting. He
08:54said, you know, you set up your studio. You wanna figure out, and he's another one who included a link. A cool one.
08:59Tell me back. Because you could well, you know, you could see how, like, professional filmmakers would actually use something like this. Maybe they already do use software that does this, but this guy just built it with with Fable, you know.
09:12Yeah. I can see people setting this up just for their own office. You know, you take some measurements of your office.
09:18You put it in, and you wanna know where the desk should go, have it adjusted.
09:21It does seem like a lot of these, you know, demos that that we're seeing float around today, uh, again, like, just, like, one shot, one one small prompt can get you to that level of detail.
09:32That seems pretty impressive to me. Yeah.
09:34Here's a guy who basically replaced McKinsey with a prompt. Let's let's look at it. He said, look.
09:40Here is a McKinsey report, and I'm pausing so I can read it. Create a presentation like this except on AI and for the second half of twenty twenty six. Make high quality research predictions and create graphics.
09:51Make it the same style and everything. I wanted to feel like a professional doc that would pass as a McKinsey presentation like the one I uploaded.
10:02And here you go.
10:07The level of research I think I'd wanna take a look at and see, is it smart? Is it useful? Did it do it based on what I want?
10:14And that's where I think it would, uh, that's where I think I would have to decide whether this is better than what we had with Opus or if it's just another version of AI slop. Justine said, I want a Monopoly game.
10:28And instead of having all these, uh, New Jersey style places,
10:33I actually want it to be AI focused. I was looking into this one. And again, like, just looking at it, it's a Monopoly board.
10:40You're you're seeing the typical Monopoly interface, if if you will, if if you were to make it a video game. So, again, that's something that you probably could create with Opus, you know, a couple weeks ago or even a couple months ago.
10:51No problem. But then you then I started reading into it, and she was talking about how like you know, and it also took it upon itself to build in the multiplayer interface.
11:01So you can, like, share a link with someone and actually play Monopoly multiplayer. And that and that again, it's just like a level above capability just right out of the box or right through the first prompt that's like, woah.
11:14You know? That's
11:16that that was fast. And because it comes now with our subscription for another few days, I wanna see what I could build right now before it's too late. This one, I don't love when people do this.
11:27He's basically said, look. I one shot it. No interventions.
11:29The writing physics are actually fun. First time, a game crossed from somehow works interesting to play.
11:35So he made a game for himself. This is kind of interesting. I like it.
11:39Don't know I don't know that it really makes much of a difference for me.
11:42Yeah. I don't know. Again, it's just, like, kinda, like, visually stunning.
11:45You can see the the movement mechanics. So there there's definitely a lot of, like, processing under the hood that's happening here.
11:53I I don't know. Like, it's it's hard it's hard to know, like, how impressed I should be with this stuff on on the very first day, you know?
11:59Alright. Final one. This is Turbo Kart.
12:04It's basically him recreating Mario Brothers.
12:09Single shot, about fifteen minutes of work. And then and then I wanna go into some of the issues that people have had with this, and then I'll show mine. I will say, like, I've been I've been making games like this with my kids and and giving them a taste of, like, vibe coding, like, a, first time game with with with Claude.
12:24This is, you know, months ago. And it would do things like that and and try to create some some music, but it was just, like, very rudimentary, like, beats and stuff.
12:34That is, like, in the flavor of a of a Mario Kart game sounding music and the and the level of detail and the graphics. Again, like, this stuff has always has has been possible, but it's it's it's the details without
12:48needing to prompt for every little thing. Okay. Here's what I'm gonna do with this, actually.
12:51My kids have an issue where they they're getting the my friend and I versus me and my friend use wrong, and so I created a game for them where they're asked the question and they have to answer it.
13:04I think what I'll do is if they get five of them right, they get to play Mario Kart like this or if I love it. Right, if they get 10 of them right. Okay.
13:11Let's go on and talk about a couple of issues that people have found with it and how to get around them. This one's a big one. It's come up with a lot of people.
13:18I don't think it's that good for that. I mean, it's very slow. It's very expensive.
13:22It's extremely token hungry. One thing that's a sort of a pro tip that's not intuitive for this model is you can set it to lower reasoning levels. Like, you know, instead of max or extra high, you can put it to medium or or low for more basic questions.
13:35That's not something that was very intuitive to me, but that's how people inside of I didn't notice that either. And you know what? And I had a friend, Josh, who I'd interviewed before here who, um, asked for one request, and it burned through all his tokens in the $200 a month plan.
13:49So that's actually helpful to be able to shift over into into the easier, uh, version of it. And then here's another set of tips,
13:58and this is something that I did do. Some usage tips here. So the first one is to treat this thing as a thought partner.
14:04This thing is really smart, and involving it in your planning and your ideation is a really good idea, but you wanna take that with a grain of salt. Don't just use it as a thought partner and say, okay. How would I do this?
14:16It gives you an answer, you say, okay. Go. Let's do it.
14:18Use it as a thought partner, meaning brainstorm with it and make it play devil's advocate, and you play devil's advocate. That I didn't do. I did say ask me questions.
14:26I didn't ask the devil's advocate to sharpen my thinking on it. Yeah. Actually, that that's another area where I will start to pay attention to, like, how much better is this?
14:36Because the thought partner thing is something that I've been using AI and Claude specifically for for a really long time, like, over a year.
14:43On a regular basis, almost a daily basis, I am having a voice note conversation, strategic planning, strategic coaching conversation with Claude.
14:54And how I would evaluate it, as as he's saying, is, um, how does it how does it help me change my thinking
15:03on on a decision or on a strategy? Right. You know?
15:06Um, rather than just kind of regurgitate back to me what I'm saying. But in general, because it can build so much and it does take a while to work, it is better to have it think with you before you build it so that Yes. You know, it's going in the right direction and not just going in some magical direction.
15:20The the ask clarifying questions, that's another really, really strong pattern.
15:25I've been using it. I think a lot of builders have been using it for a while. You know, when you're creating a a PRD or a spec, I'll give it, like, the raw requirements or my raw goals.
15:35Ask me like, literally, I type in ask me clarifying questions on almost everything. Or or or or another one I like is, like,
15:44let me know whatever I may have missed here or what I'm overlooking or what the gaps are. You know? I told you guys you can get in there.
15:52And the grill me skill was originally from Matt Pocock, so shout out to Matt. But I changed it up a little bit so that it has, like, brainstorm docs. So if I come in here and I go to my brainstorm folder, which is right here, which will automatically get created if you use my skill, I've got different sessions where it's grilled me relentlessly for I'm talking, like, 15 questions, 25 questions, 30 questions, and it's extracted so much more knowledge out of my head into my AIOS.
16:18So that's a really great way to start off too. You could literally set up your AIOS and say, use the grill me skill to figure out
16:25I find that that's incredibly helpful that I want you to make me think about it so that I I think of things that I didn't imagine that I might want in this Relationship map, Claude. The most important is to to verify its own work. You'll notice if I go into
16:39Claude and I go to this session where I prompted it to build this, you know, this relationship map thing, at the end of my prompt here, I said I gave it some context. I said, this is a demonstration for YouTube, so don't feel the need to make this production ready, but it should be easy to understand, so don't make it confusing.
16:54That's not even what I meant to show you guys. I meant to show you guys this. And then once you have built that, use a dynamic workflow to verify that everything is accurate and works as expected.
17:03I I wonder if I need to do that. I think I can test.
17:07I don't know that it's important for all of the things that I do, but I guess for what he's doing is he's basically building a second brain using, um,
17:14using Fable, and so he would want it to be tested a few times. Yeah. The verification thing is coming up a lot, especially this week, but it's it's been a topic this year.
17:23Uh, if you're gonna have these models, which are much more capable now of these long running deep tasks, like go build something big, and I'm gonna walk away for three hours and expect it to be done. The way to make that work is giving it really strong verification criteria, saying, like, it's done when these things are true.
17:44Oh, I see. And these are these are hard checks that, like, you're not done until you can absolutely confirm with screenshots, with, you know, tests written, with whatever it might be.
17:56Okay. You know what? Let me show what I built here.
17:59After I saw all these, I got inspired. I love to take screenshots and mark them up.
18:04I use so many different tools, but but none has felt great. So here, watch. I'm gonna take a screenshot of this, and I'm gonna go into an app that I do like.
18:11It's called new markup, uh, or markup hero. I'm gonna hit command v. I paste it in there, and then this is someone else's site.
18:19It's nice. And then I could, you know, mark it up. I could add arrows and so on.
18:24But it doesn't have everything that I like in it, so I built my own. Oh, wow. Super easy.
18:30Look. Immediately when I click mine, it automatically paste it. I don't even like the waste of time to hit command v.
18:36I'm coming in here and hitting the bookmark, and so I wanted immediately to come in, and then I have only the tools that I care about. I can draw on the screen circle.
18:44I can have an arrow go up, I can actually So picking up the arrow like this. And then I could either hit this button to copy or I could hit command shift c to copy and I do this so many times. It's one of my bookmarks on every single device and it only has now the features that I care about.
19:01It's named Andrew's Markup. And here, let me show you how I did it. Because I kinda did some of what Nate was talking about.
19:06Here, I use Versus Code to access Claude.
19:10So I'll show what using. CleanShot for the for the screenshotting.
19:13I do too, but I do not like having it on my computer because it's just another thing that uses resources, and I just don't like that. So here, I opened up, uh, Claude in here, and I said, what model are you using? It said I'm using Opus four eight.
19:26I said, switch to Fable. It said, I can't switch to Fable, but you can. Just do this.
19:31Command model and select Fable five, so I did. And then I said, okay. Look.
19:36I pasted the link. I said, I want you to copy this site and put my cloud and put it on my Cloudflare account. Interview me about it.
19:42Make sure that you've got the features that are important to me. Here's what I think is important. Obviously, I just talked into it, so I dictated all these features.
19:48And then it started to it started to interview me. The thing about interviewing though is it's mostly, like, uh, what's it called?
19:55It's mostly multiple choice questions, so it's fairly easy. And then once I had it, it said, now go over to Cloudflare, click this link to connect it, and there you got your link, and it freaking worked.
20:06Then I said, look. Again, I I really think that interview me or ask me clarifying questions pattern is so important, especially for the builders out there who are just becoming builders.
20:16Right? Like, if you're if you're just running a business and now you're starting to, like, gain all these kinda superpowers to build any tool that you want, You know, the the way to shortcut the level of experience that a real software developer would have, at this point, you're like, you don't have to learn how to code.
20:34You don't have to learn how to wire up every piece of programming logic, but you do still need to make key decisions. Right?
20:44Like, I think it's easy to say like, alright, I just I just let Fable run all day long and it built me some magical app. Like, that's really cool, but does that app actually solve your problem in your business the way that you need it to be solved?
20:58Like, Fable could make those decisions for you, or you can bake into it what you actually need to build.
21:05And
21:06having it extract that out of you through through questions is is huge. Yeah. A simple example for me was I didn't want, like, infinite color options.
21:13That's another distraction for me. I just wanted it to pick eight for me, and that's fine. And I wouldn't have thought to ask for that, except it asked me the question.
21:20Alright. I've got a link to all these, uh, videos below. And if you're into Fable five, Wade from Zapier, our sponsor, has tested Fable five, run it over 600 tests against he did 600 tests on it, and then he compared how Fable five responded to those tests versus how Opus eight, Opus seven, Gemini, and all these other platforms.
21:42There's a link right here for you to go follow and see it.
The Hook
The bait, then the rug-pull.
One sentence. One prompt. Notion-level productivity app. The hosts open by reeling off three impossible-sounding builds — Lovable clone, Mario Kart, a ten-billion-dollar app builder rebuilt in two prompts — and the video delivers on every one.
Named ideas worth stealing.
14:00concept
Thought partner before builder
Use Fable as a devil's advocate before issuing any build command. Brainstorm, make it argue against your plan, argue back. Only then build.
Steal forAny long autonomous build task — prevents building the wrong thing confidently
15:00concept
Clarifying questions first
- Ask me clarifying questions before building anything
- Let me know what I may have missed
- Grill me on the gaps
Prompt Fable to interview you before it builds. Extracts requirements you did not know you had. Multiple-choice answers make it fast.
Steal forAny new tool or app build — especially useful for new builders
17:20concept
Hard verification criteria
Give Fable explicit done-when conditions: screenshots, passing tests, specific UI behaviors. Not 'check your work' — 'you are not done until X is confirmed.'
Steal forAny task where Fable runs unsupervised for more than 30 minutes
13:40concept
Reasoning tier dial
Fable has adjustable reasoning levels. Use medium or low for simpler questions to prevent token hemorrhage. Most users leave it on max and overspend.
Steal forAny session mixing complex and simple tasks
How they asked for the click.
VERBAL ASK
21:30link
“There's a link right here for you to go follow and see it — Zapier tested Fable five across 600 tests.”
Soft Zapier sponsor close with a genuine test-data value prop. Not a hard sell.
MENTIONED ON CAMERA
FROM THE DESCRIPTION
PRIMARY CTAWhere the creator wants you to go next.
OTHER LINKSAlso linked in the description.
Visual structure at a glance.
