YT YouTube2Text

YouTube2Text → Transcripts → Claude Fable 5.2: How I’d Make Money With It

Claude Fable 5.2: How I’d Make Money With It — Transcript

by Greg Isenberg · 3,428 words · 500 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:00Fable 5.2 is here and I wasn't going to
0:02make this video, but it is so insane and
0:05there's so many ways that you can make
0:07money with it, be more productive,
0:09figure out ways to create hardware. And
0:11today I'm going to go through a bunch of
0:13the prompts that you just need to know
0:15about because it's going to connect
0:16things in your brain, it's going to give
0:18you ideas. And I haven't seen anyone
0:20talk about it. So
0:22by the end of this episode, you are
0:24going to learn how to make money, how to
0:25be productive, how to think about
0:27building things like hardware with Fable
0:285.2. I'm going to give away all the
0:30prompts. [music] Um and if you stick to
0:32the end, I think you're going to
0:33understand how to use Fable 5.2 better
0:36than 99% of the people. So like, comment
0:39and subscribe. And if this episode gets
0:405,000 likes, I'll include in the pinned
0:43comment a link to all the prompts and
0:45some extra bonus ones. So go and pause,
0:48go and do it and then I'll see you at
0:50the end of the episode. Enjoy.
0:52>> [music]
0:52>> The startup ideas podcast.
0:56It's sipping time, baby.
0:58>> [music]
1:00>> When you're going to open up uh Claude,
1:02you know, if you're on I think the max
1:04plan, it's going to say like go and, you
1:07know, bring complex ideas to life with
1:09animations and 3D worlds and describe
1:12the result you want. So you're going to
1:13go ahead and click try Fable 5.2. Um
1:16let's start with the first prompt. Um so
1:19one really cool thing about Fable 5.2 is
1:23you can have it uh just like literally
1:25audit your entire business. So what's
1:27really cool is it can work 10 to 12
1:29hours on its own until the job is done.
1:32So you can give it a job that would
1:33take, you know, a high-end consultant a
1:36week and just have it cook. So you can
1:39say something like this. We're going to
1:40go ahead and try it out here.
1:43Audit my entire business, the website,
1:45pricing, onboarding, emails and top five
1:46competitors, work until it's actually
1:48finished, keep a running log, give me a
1:50ranked list of the 10 changes most
1:52likely to make money with the evidence
1:55for each. So we're going to go ahead and
1:57do that. I'm going it my business
2:00late checkout.agency
2:03design agency.
2:05Um so it's going to go ahead and do
2:06that. Um and while we do that, let's go
2:09uh let's go and and look at another
2:11making money prompt. So you can actually
2:14do a freelance job end-to-end with 5.2.
2:19So if you look at the you know the blog
2:21post from Anthropic, they say in their
2:24in their official blog post
2:26that you know the I think the yeah the
2:28real freelance score went from 21% to
2:3041%. So I think there's a huge
2:33opportunity to actually sell uh results
2:36as a productized service. What do I mean
2:38by that? We can go ahead and give 5.2.
2:42Let's go ahead and open a new
2:44chat here.
2:47Um we can go and find a job posting. We
2:49can say do the whole job as if you were
2:51a freelancer, then tell me which parts
2:53you're confident in, which parts human
2:55should check, and what you charge for it
2:56as a fixed price service.
2:58So I don't have a real job posting off
3:01the top of my head. So I'm actually
3:02going to go and have 5.2 find it.
3:06Um actually you go find something you
3:10think you can do that um can make me
3:15money.
3:16And you know we we can go ahead and do
3:18that.
3:19Um
3:21uh what's really cool is there's so many
3:25jobs that uh are freelance jobs that if
3:28you use 5.2 for and you create a great
3:32loop for it and you're you know you
3:34could sell it and and not go hourly,
3:37right? You can go and sell it as a job
3:38to be done. I think there's going to be
3:40thousands
3:41uh of people who come and go and do
3:44this. So I think this is a huge
3:45opportunity. Number three, 100 niches at
3:48once. So one of the cool things about
3:505.2 and I know a lot of you are going to
3:52be happy about this is it's much better
3:56at running a lot of sub agents. So you
3:57can actually say something like this,
3:59spin a spin up a sub agent for each one
4:02of these 100 niches. Each one finds the
4:05top five complaints people have, what
4:07they pay for today, and how hard it is
4:09to reach. Give me one rank table and the
4:11five I should build build for. So I
4:13don't know if this will be able to do
4:14this
4:17uh it'll finish but by the time this is
4:19done but we're going to give it a try
4:20that's for sure.
4:21Um
4:24underserved
4:26and we're going to go check in uh later
4:28about that
4:30so you can see
4:32um so you can go see it.
4:34Number four, rebuild the ancient
4:36software. So this is a huge opportunity
4:40just in general. I've talked about it on
4:42the podcast before but there's so much
4:45uh tech debt, old software, and there's
4:49opportunities to go and rebuild that for
4:51people. So you can go basically say
4:52something like this. Here's the
4:54spreadsheet my client has run their
4:56business on for 12 years. Figure out
4:58what it actually does, rebuild it as a
5:00simple web app, move the data over, and
5:03write a one page guide for the staff. So
5:05I I anticipate a lot people we talked
5:07about you know pay per service. There's
5:10going to uh instead of hourly. In the
5:11past
5:13uh people would just go and do this
5:14hourly for people, right? $50 an hour,
5:16$100 an hour, whatever it is. There's a
5:18just a an opportunity now to use 5.2 and
5:22go ahead and do that uh per per output.
5:25Um and then the fifth one, you know deal
5:28checker for buying a a boring business.
5:31You know one of the big critiques of AI
5:34in general is it'll go and hallucinate
5:37basically.
5:38And
5:40when you speak to people in legal, when
5:42you speak to people in finance, when you
5:44speak to people in you know different
5:47spaces where you can have like
5:49hallucination, right? Um now it finally
5:52says I don't know. So what's really cool
5:55about that is it's it's it it's it's a a
5:58human being, right? A human being would
6:00say I don't know to a question but up
6:02until this point AI hasn't said I don't
6:06know. So this is like a full-on new
6:09uh era in the AI world, right? The know
6:12era. And this is really cool. I don't
6:15know. I'm not going to say it's AGI, but
6:18you know, uh it if if it was AGI, that
6:22that does feel like kind of
6:25human-like, right? So you can say
6:26something like, here's the seller's P&L
6:28on listing. Find everything that doesn't
6:30add up. Mark each number as verified,
6:32unverified, or suspicious, and write the
6:3415 questions I said ask before making an
6:37offer. So much opportunity to do stuff
6:39like this. This has been out of reach.
6:41Finance, legal, stuff like that has been
6:44out of reach. Um but I I think you're
6:46going to see a lot more professionals
6:49actually going and using 5.2. So that's
6:52exciting for them as well. Let's go
6:54ahead and check uh our prompts and see
6:57what we got so far. So um
7:01when you actually go through this is for
7:03yeah LCA, our our design agency and
7:06design engineering uh agency.
7:09Um it goes and does an incredible job at
7:12ranking all the different calls to uh
7:15all the different feedbacks.
7:17And I'm not going to go through exactly
7:20all of this. Um but it's just
7:23every time I've been I've asked uh
7:26this prompt on on a bunch of different
7:28things that we've been incubating, it
7:30has been just so uh to the point. Like
7:33literally all this is to the point. So
7:35whatever it is you're working on, this
7:36is a no-brainer. Do an audit and then
7:39you have this running log too. Um let's
7:41go to the freelance job. Um
7:46uh yeah, how do you automate a freelance
7:48job? So it's gone. It's gone and picked
7:51a job, the equity analyst to produce
7:53report on TDM X on Upwork. $450 fixed
7:57price, intermediate level was posted two
7:59days ago. Now it's actually gone
8:01and done the work.
8:04And look at this.
8:06This is insane.
8:08This is going and like would someone pay
8:11you know, it's still working on it, but
8:12would someone pay hundreds of dollars
8:14for this? Absolutely.
8:16And that's the crazy part. Like this is
8:20there's so much opportunity to go and
8:23actually do things like research, um
8:27things like I've talked about on this
8:28channel before, creative uh
8:30doing creative output um using some of
8:33these tools. So a lot of a lot of uh
8:35stuff an opportunity here.
8:39And uh still working on this one.
8:43Um but yeah, the sub agent thing is a
8:45really big deal. Um you're getting
8:47better output. It's going to be faster.
8:49There's just like multiple things that
8:51you can do at the same time. So that's
8:52really really cool. Let's go to being
8:55productive um because you know,
8:59I want to be productive. You want to be
9:01more productive. Um so what are some
9:03tips to uh with Fable 5.5? Uh the grill
9:07me prompt. Interview me with 20
9:09questions about how I make decisions in
9:11my business one at a time, then turn my
9:13answers into a skill file you'll follow
9:15from now on. This is a
9:19uh 5.2
9:21uh prompt that everyone, literally
9:24everyone should should be using. This is
9:27going to make you're just going to get
9:28more more output, better output from
9:315.2. It's a no brainer. Number seven,
9:34run the apps that never had an API.
9:37What's really cool about 5.2 is it's
9:39really strong at operating a computer.
9:42So you can say something like open our
9:44old invoicing app, export last month's
9:46invoices, match them against the bank
9:47statement and list anything unpaid, and
9:49then ask me before sending anything.
9:53Um so I'm happy to see that uh in 5.2 uh
9:57once you start playing with it, you're
9:58just going to see it's it's really good
10:00at operating a computer. Feels a lot
10:02better. Um so that's been that's a huge
10:05win uh
10:06win here. Number eight, red team my
10:08plan. So you can say something like,
10:11here's my plan for the next quarter,
10:13argue against it like an investor who
10:14wants to pass, then tell me the one
10:16thing that you that would change your
10:18mind. One of the things I I I did here
10:20is I literally gave it my vision deck
10:23for LCA, my design engineering agency,
10:26and I was just like,
10:28you know, I had had some of my goals in
10:30there and I asked it what it would
10:31change, and I was uh floored with just
10:35how good it was. Um I had actually asked
10:38uh a really well-known sort of advisor
10:41to also give me feedback, and it was in
10:44some ways the 5.2 feedback was even
10:46better than that. So kind of crazy. Um
10:50if you if we if we move on to the video
10:53and content section, you know, if you've
10:55been watching this podcast, Startup
10:56Ideas podcast, you know that I've been
10:58talking a lot about marketing
11:00engineering and just using AI agents to
11:03do marketing. 5.2 is an insane deal um
11:08around just like what you've been able
11:11to do. My friend uh
11:13Riley uh
11:15just posted just a a a a 40 a 40 or 30
11:19or 40 second like basically one prompt,
11:22almost like documentary style video um
11:27that he was able to do. It's actually
11:29insane. You can do explainer videos. I
11:32saw my friend uh my friend Hunter create
11:34an explainer video, one prompt, and it's
11:36beautiful. Just absolutely stunning
11:39visuals. So where are the opportunities
11:41there, right? Faceless content on on
11:43explainer videos. Why aren't you
11:44creating a vertical um you know YouTube
11:47channels and Instagram? You know, own
11:50that distribution, right? Distribution
11:51is everything.
11:53Um so if you can start
11:55pumping out stuff using 5.2 and it only
11:58costs you $1 $2 $3 $4 $5 a video, um you
12:02can create a really great business
12:04there. So I I I I I was thinking about
12:07that in in the context of real estate.
12:09Like,
12:10you know, real estate, you know, on one
12:12hand such a great business, um but is so
12:15capital intensive. This is so little
12:17capital intensive and in some ways,
12:19social media is almost like real estate
12:22in that sense, right? You're you're
12:24you're putting out these like mini real
12:26estate videos and stuff like that and it
12:28earns you AdSense and it earns you
12:30followers that you can later sell. So
12:33there's a lot of opportunity here. I'm
12:35going to go through three um
12:38three quick prompts. You can have it be
12:40your short form editor. Um by the way,
12:42I'm going to include all these prompts
12:44in the description. You can go grab it.
12:46Um but obviously like, comment,
12:47subscribe for more. Um here's a raw four
12:50minute talking head video. Cut the
12:52silences and retakes, find the strongest
12:5445 seconds, add captions and simple
12:56motion graphics and export it to video
12:58vertical and then show me three
13:00different hooks. So this is a no-brainer
13:02thing using it as a short form editor.
13:04I'm I'm doing this. I'm and uh the the
13:08the 10th prompt is the explainer videos
13:10for the niche. So that's what I talked
13:11about with Hunter. Um you can say
13:13something like make a 60 second animated
13:16explainer on this topic for this niche.
13:18Write the script, build the animation,
13:20encode and render it, then make five
13:22variations with different first lines.
13:25And then lastly, one episode every
13:27format. So you just say, you know,
13:28here's the transcript, uh make a PDF
13:31playbook, five tweets in my voice, a
13:33newsletter, a free tool someone could
13:34use in two minutes, and then flag any
13:36claim you couldn't verify. So this is
13:38about doing long form content and then
13:40turning that into a tsunami of different
13:43formats. So uh
13:45I think there's a huge opportunity to do
13:47long form and then just sort of
13:49uh have it
13:51pump out all these different things
13:53because now 5.2 is that that good. And
13:57last but not least, um I've been excited
14:00about this for a while. I'm actually
14:01going to do an episode on hard one
14:03person hardware companies and well
14:04hardware companies in general robotics
14:06and AI, but uh 5.2 designed three
14:10working devices from a single prompt in
14:13uh Uh,
14:14I was like I saw that Anthropic had said
14:17that I think in in one of their blog
14:18posts. So I you know, you can just go
14:21and and and give it a prompt like this.
14:23Design a small device that does one job
14:25for one niche. Give me the parts list
14:27with the prices, the circuit design, the
14:29firmware, and an enclosure I can print.
14:31And then you can say tell me what it
14:32costs to make 100. And you can go and
14:35make like a like a full-on Shopify with
14:37it, right? So like
14:39maybe design a small device that um
14:43maybe is
14:45uh
14:46kind of like BlackBerry
14:49the phone
14:52that doesn't have a million
14:56apps and is focused on people who aren't
15:03trying to doom scroll all the time.
15:07So you have like an idea in the shower,
15:09you're like oh my god, wouldn't it be
15:11cool if there existed some phone that
15:15was more minimalist. Um actually a
15:18friend of mine has a business like this
15:20called the minimalist phone. But let's
15:21say you were trying to create something
15:22like this.
15:23Um you can go and just ask it to do this
15:26and it's going to design you the
15:28hardware. Um it's going to give you the
15:30part list. It's going to give you the
15:32circuit design. It's going to give you
15:33the firmware. This
15:35uh was something that would literally
15:38cost thousands and thousands and
15:39thousands of dollars that you can now do
15:41with 5.2 with a high degree certainty
15:44that this is actually not it's not going
15:47to hallucinate. Um
15:49and of you know, of course there's going
15:51to be some hallucination like in in AI
15:54in general you just have to assume there
15:56might be some hallucination or there
15:58might be some it's you know, even when
15:59you work with people like there are some
16:01hallucinations. So I wouldn't 100% trust
16:03it, but um from my experience working
16:06with 5.2 um I've just been really
16:09impressed with what what I've been able
16:11to get out of it.
16:13So look, it's saying I'll design this 4G
16:16and text handheld with a physical thumb
16:17keyboard and a monochrome screen
16:20starting with live part pieces, the
16:21circuit firmware and printable enclosure
16:24enclosure. And it's you know, I really
16:27like just like the UI of of
16:29of Cloud in general, honestly, with the
16:32side like I think you know, this
16:34artifacts over here.
16:36Um it's starting to get like really nice
16:38and and and cool and good. So overall,
16:40I'm excited about that.
16:43So here you can see the their naming the
16:46the device Plain one, a calls and text
16:48handheld with a real keyboard. Plain one
16:50is a pocket 4G phone that does six
16:52things, calls, texts, contacts, notes,
16:54clocked and alarm. It has a 45 key thumb
16:57keyboard, a 2.7 in black and white
16:59screen. You can read in sunlight and no
17:01browser, feed or app store. The screen
17:04is the anti doom scroll feature. It's a
17:07sharp memory LCD. I mean, this is crazy.
17:10Like you know, I've talked about vibe uh
17:13vibe
17:15vibe manufacturing. I've talked about
17:16vibe manufacturing on the pod before. Um
17:19we're with 5.2 we are truly here.
17:24Uh we are truly here. Um
17:26you know, I think when Astra came out,
17:28we just saw
17:30it it felt like okay,
17:32this is where it's going and you're
17:34going to be able to do this.
17:36Um with 5.2 it just takes it even to the
17:38next level.
17:40So it's gone and created the actual
17:43device. And if you go through the
17:45document, the Plain one document, it has
17:48basically everything, all the parts, the
17:50whole circuit design, like look how
17:52crazy this is, the power rails, the pin
17:56map, the modem connections, the display
17:59connectors, the keyboard matrix. If you
18:01wanted to have it designed an operating
18:04system based on uh
18:08an operating system based on Android,
18:10like that would be I think a really cool
18:12idea. Um
18:14it it can you know, could do that. It's
18:16showing you what it could look like.
18:18Um this is bonkers. Like this is
18:20absolutely bonkers. You know, if you
18:22don't look at this and you're and it
18:24doesn' get your mind going 100 miles a
18:27minute,
18:28uh I don't know cuz this is just just
18:31absolutely incredible. Vibe
18:32manufacturing is here. And I think one
18:35of the coolest parts about 5.2 in
18:37general is it's probably the most
18:41thoughtful
18:42model that exists today. And that means
18:45you're going to get better output. It
18:46means it's going to control tools. Tools
18:48like Blender, DaVinci, whatever tools
18:50you're using or giving it, it's going to
18:52just be better at using those tools.
18:55So we now have the most thoughtful model
18:57on the planet. What are you going to do
18:59with it? These are just a few of my
19:02examples and I there's hundreds more,
19:04thousands more. We're going to be
19:05discovering these use cases. Please let
19:07me know in the comment section how
19:09you're using it, what you're seeing
19:10work. Um, this has been a pleasure just
19:13riffing with you, going through some of
19:15my ideas. Hope it's been helpful. Hope
19:18it's gotten that creative juices
19:19flowing. I am rooting for you to go and
19:21build whatever your heart desires. This
19:24is an incredible time to be building,
19:26which fires me up. I love a model
19:28release day. I know you do too. I'll see
19:31you next time and have a creative day.

About this transcript

This page contains the full transcript of Claude Fable 5.2: How I’d Make Money With It by Greg Isenberg, generated from the public captions YouTube serves with the video. The transcript has 3,428 words across 500 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.