YT YouTube2Text

YouTube2Text → Transcripts → Starting An AI Side Hustle Is EASY (Copy This System)

Starting An AI Side Hustle Is EASY (Copy This System) — Transcript

by Sandeep Swadia · 2,590 words · 387 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:00If you're thinking about starting an AI
0:02side hustle or building an AI business,
0:06watch this [music] before you start
0:07because the easiest way to make money
0:10with AI may be the worst business model
0:13to build. I learned this as a CEO and
0:16board [music] member at tech companies
0:17that created billions of dollars in
0:19value. And they all say the same thing.
0:21Speed matters in the world of AI, but
0:24only when it helps you build something
0:26differentiated
0:27and defensible and figure out how to
0:30build a [music] strong business that can
0:32make money for years. 40 years ago, if
0:35you wanted to do a television show in
0:37America, there were basically three
0:40doors you could knock on. ABC, NBC, and
0:44CBS. Now, if all three of them said no,
0:47your show basically didn't exist. If you
0:50have a phone and a Wi-Fi connection, you
0:52can launch your show on YouTube today.
0:55You don't need anyone's permission. But
0:56at the [music] same time, by the time
0:58you wake up tomorrow, another 20 million
1:01videos will be uploaded on YouTube.
1:03[music]
1:04So, YouTube makes creation and
1:07distribution easier, but what becomes
1:09harder is your ability to make something
1:12valuable. And AI is doing the same thing
1:16to intelligence and execution. This
1:18makes building a side business easier.
1:21But AI itself is not your edge because
1:25your competitors have the same tools.
1:27Low startup costs, great. Low
1:31differentiation, fatal. That's the
1:34artificial illusion. Easy to build
1:36doesn't mean [music] easy to win. The
1:38real question is what should you build
1:41that's harder to copy? What's that place
1:45where your AI side business can succeed
1:49and thrive? To find that place, you have
1:52to pass through four gates. I call them
1:55deal [music] gates. We will pass through
1:57each of those gates and bring each one
2:00to life with a different AI business
2:02example. Let's go to the first gate,
2:05demand. How do you know you have a good
2:07business idea when somebody starts
2:10paying for it? In 2006, Elon Musk was
2:13trying to do something that sounded
2:15insane. His company wanted to build
2:18millions of electric cars, but he did
2:21not know if there would be any [music]
2:23demand for even one of them. Almost
2:26nobody believed that electric cars could
2:28compete with Toyota or BMW. So Tesla
2:32unveiled the Roadster and offered the
2:35first 100 cars as a special signature
2:39series. Every buyer had to put down a
2:42$100,000 deposit and they didn't get a
2:45car. All hundred were sold out in 3
2:48weeks, but not a single one was
2:51manufactured yet. So by August of 2007,
2:55Tesla had pre-sold 570 Roadsters. Months
3:00before manufacturing even began, he had
3:03an answer to a question most founders
3:06don't even bother to ask. If I build it,
3:09who would pay and how much? Your
3:13customers have a vote. They have a veto.
3:16And it comes down to the same thing.
3:18Will they cut you a check? If nobody
3:22pays, you should be willing to change
3:24the offer, change the customer base, or
3:28even change the idea. And that's why a
3:30lot of us build AI side businesses
3:32backwards. Someone shows you that you
3:35can build um a website with nothing but
3:38a prompt and it just takes 10 minutes.
3:41So you think, "Oh man, I I can build an
3:44AI website business." But all you've
3:46done here is proven that AI can make a
3:50website. You have not proven that
3:52anybody wants to pay you to make it for
3:55them. Why wouldn't they build it
3:57themselves? So AI can supply solutions
4:00for you but it cannot create demand for
4:03you. That you have to do yourself. Now
4:06how do you do that? How do you verify
4:08demand? You can do three things. First
4:12pick narrowly the niche of the niche.
4:15Don't choose a solution for small
4:17businesses. I mean that doesn't exist.
4:20Choose a solution for dentists with less
4:23than three locations in three towns
4:26around you. or pick independent
4:28recruiters or Shopify brands that are
4:31doing under a million dollars in
4:33revenue, something like that. Because
4:35the narrower the customer, the easier it
4:38is for you to understand what they
4:40actually need versus what they think
4:43they want. Second, talk to buyers. Don't
4:46interview your friends. I used to do
4:47that. They will tell you what makes you
4:49feel better. Your actual buyers will
4:52tell you whether there is a pain and
4:55where is that pain and what gets in
4:57their way and if that problem is painful
5:00enough for them to pay for it to
5:04disappear. And third now build the
5:07smallest test. In the tech world they
5:10call it the MVP, minimum viable product
5:14or service. Use AI to build the smallest
5:18working version and put it in the hands
5:20of your future buyers. Now you can tweak
5:23it, change it, test again, keep going
5:26until someone figures out that this is
5:28valuable enough and they'll pay for it.
5:30Now let's take a real world example like
5:33AI bookkeeping. Now here's the
5:35competitive dynamic that you will see
5:38with almost every idea that you'll come
5:41up with in today's world. You're never
5:43going to be able to enter a completely
5:45empty market unless you're super lucky.
5:48For example, in bookkeeping, QuickBooks
5:51already owns a large portion of the
5:54market, at least in the US, and all
5:56these AI native startups are coming
5:59after it. So, your opportunity is not AI
6:02bookkeeping. It's a narrow customer with
6:06a problem that the existing tools can't
6:10solve. I'll give you an example. After I
6:12retired or graduated
6:16for full-time work, I was advising
6:19private equity firms and serving on
6:22boards. And none of the offtheshelf
6:26bookkeeping tools quite fit what I
6:30wanted to do, what I needed. So, I ended
6:33up hiring an independent bookkeeper who
6:35then built a custom process around my
6:39workflow and used AI to help build it,
6:42help run it. That's the kind of hidden
6:45demand that you're looking for. So, I
6:48would narrow cast a lot. maybe
6:50independent consultants like me or maybe
6:52laundromats or something very small and
6:55talk to 10 of them and find out what
6:57QuickBooks or the AI startups or their
7:01current bookkeeper doesn't solve. Then
7:04build the smallest version around that
7:07particular pain. The cost of running
7:10those kind of tests has collapsed. You
7:13can research faster, build a rough
7:15version in days. So AI changes the
7:19economics of entrepreneurship
7:21completely. But choosing the customer,
7:23earning their trust, reading reactions,
7:26reading the room, deciding what matters
7:28to them, to you, all of that, that's the
7:31hard part. And that's still yours. Now,
7:34let's enter the second gate, the edge.
7:37Your edge comes from many places. Let's
7:40cover three. Imagination, focus, and
7:42context. Let's take imagination first.
7:45Many years ago, there was a street
7:47performer in Canada. His name was Guy.
7:50So, he would breathe fire, played music,
7:53and he had this one interesting dream.
7:56He wanted to start a circus. That was a
8:00startup idea. The problem was circus was
8:03a fading industry. Kids were going to
8:06theme parks instead of watching animals
8:08do tricks. Traditional circuses were
8:11losing money or shutting down. But Guy
8:15imagined something completely [music]
8:17different. He combined circus with the
8:21idea of a Broadway show. So he would
8:23bring incredible gymnasts, dramatic
8:26lighting, original music, storytelling,
8:29no animals. And that became Cir to Sole.
8:32Today, CIRC generates roughly a billion
8:36dollars in revenue every year. Sometimes
8:38[music] your edge comes from
8:40imagination. Sometimes it comes from
8:44obsessive focus. I'll share my own
8:46story. When I became CEO of a company,
8:49our company had three possible paths for
8:54growth. And two of those markets were
8:57much bigger than the one we were already
9:00in. So it was very seductive. It was
9:03very tempting to build something for
9:05those markets. But instead of spreading
9:08ourselves across three opportunities, we
9:11decided to double down even further. One
9:14market, one problem, one product, one
9:19type of customer. And you know that
9:21focus paid off within a few years. The
9:24company had a great exit. And finally,
9:27sometimes your edge will come from
9:29context. Andrew Ang is one of the expert
9:32voices in AI and he calls this
9:36>> context advantage over AI.
9:38>> So you [music] have knowledge our
9:39relationships our conversations our
9:42mistakes our workflow our nuances and
9:45that is something [music] AI will not
9:48have. Okay but now what if you are 26
9:52and you don't have tens of years of
9:54industry experience which means that you
9:56don't have that context. Well that's
9:58okay. you can still build context fairly
10:01quickly. And it's the same trick. You go
10:04super narrow, get close to people who
10:06live the problem, people who have the
10:08context, and learn obsessively. Let's
10:11say you're passionate about photography.
10:14That's great. You can go in so many
10:17directions from there. You could build
10:19an AI service for e-commerce brands or
10:23you can build an AI workflow for real
10:26estate. Or you can have an AI solution
10:29for wedding photographers. All three of
10:32them are three very different problems
10:34that you can solve. But where is your
10:36edge? Now if you take wedding
10:38photographers for example, then you have
10:40to talk to a lot of them and learn where
10:43their workflow breaks, which moments
10:46matter most to them. Where do they lose
10:48money? Where do they lose time? What do
10:51their clients care about? And what
10:53happens during the wedding season? and
10:56what do these guys do when the wedding
10:58season is over? All these questions when
11:01you go deep into them, that's what
11:03creates an advantage that others don't
11:05have because that's the context that
11:07tells you what to build. AI can research
11:10the niche. It can analyze workflows and
11:14build you prototypes literally in one
11:16day. And your job is to create that edge
11:20through your imagination, your focus,
11:23and the context that you can create. So
11:25you can decide what good looks like. Now
11:29let's pass through the third gate
11:31access. Every product needs to match its
11:35distribution channel. Dollar Shave Club
11:37is such an interesting story about
11:39distribution. They were selling razors
11:42and [music] blades to men. But Gillette
11:45was the 800 lb gorilla in that category.
11:48[music]
11:49So fighting them in the supermarkets or
11:52in drugstores, I mean that made no
11:55sense. So Dollar Shave Club found
11:57another route to [music] distribute
11:59their product direct to consumer. They
12:02came up with the idea of
12:03subscription-based service. [music] So
12:05you pay every month and then every few
12:07weeks or something you get your razor
12:09and cartridge delivered [music] to you.
12:11And 5 years after launching, Uni Liver
12:14bought that company for a billion
12:16dollars. That's the power of finding the
12:19right path to your customer. And
12:22sometimes your customer decides what the
12:25distribution channel would look like. If
12:26you're selling software to laundromats,
12:29well, the channel might be Google maps
12:32and cold calls and literally walking
12:34into these stores. The question is where
12:37do your product and your customer
12:40naturally meet? All right, let's walk
12:42through a real world example. Let's say
12:44your AI side business idea is an AI
12:48interview simulator. So you could build
12:51an AI interview agent that studies the
12:55resume and job description. It looks at
12:58the company, the culture and becomes the
13:02interviewer. And it would change the
13:04questions based on the answers received.
13:06But here's the hard question. Where do
13:09you find the customers? You can't go
13:11hunting for job hunters one by one. I
13:14mean, they're scattered everywhere and
13:16you don't know where they are. But maybe
13:19you'll sell through a university career
13:21center or coding boot camps or
13:25recruiters. Maybe figure out what to do
13:27with LinkedIn outplacement firms.
13:29There's so many channels now. Now you
13:31started thinking about how to distribute
13:33your product. And you'll also notice
13:35something interesting that the person
13:38using your product may not be the person
13:40who pays for it. The job seeker may be
13:43the user, but a university or a
13:46recruiter or an employer might be the
13:49buyer. But try to find that one
13:51relationship that can get you to
13:54hundreds of these users who already need
13:56what you're trying to build. That's
13:59access. Now AI can identify these
14:02potential channels for you. It can
14:03research partners and even build
14:06prospect lists. But your job is to pick
14:08the most effective distribution channel
14:11and build that partnership that opens
14:13many many doors for you down the line.
14:15So if you see that path clearly, then
14:18you have access and if you can't well
14:21you may have a very good product but you
14:24don't have a business yet. All right. So
14:26now let's go to the fourth and final
14:28gate loop. Every Formula 1 driver does
14:32countless practice runs. The driver goes
14:35out, runs a few laps, comes back in.
14:37[music]
14:38The engineering team will pull data from
14:40hundreds of sensors. They will compare
14:42[music] it with what the driver felt,
14:46make a lot of adjustments, small
14:47adjustments, and then send the car back
14:50out. Then they measure again. Their goal
14:53is simple. Every lap should make the
14:56next lap smarter. Your AI side hustle
15:00should work the same way. Every customer
15:02interaction creates data, right? Every
15:05sales call, every support ticket,
15:07customer reviews, customer interviews,
15:10there's lots of gorgeous insights
15:11sitting right inside your business. Just
15:15take all of that and feed it into claude
15:18or chat GPT or Gemini and AI will find
15:22patterns that you would never be able to
15:24find and it will build these learning
15:27loops from that data for you. Your loop
15:30must make your business more efficient
15:33every time it runs. And that learning
15:35loop is the most critical thing you can
15:38build because [music] the world is
15:41already beginning to look like a
15:43high-speed car chase to me at least.
15:46Things are changing so fast that your
15:48advantage in this lab may not last in
15:52the next one. So in that world, we all
15:55have to start working at the speed of
15:57AI. You can outsource all the grunt work
16:00to [music] AI but you cannot outsource
16:03learning and thinking that stays with
16:06you. So those are the four gates demand
16:10edge access and loops. You know in the
16:14world that we're heading towards
16:17builders are going to win. So you should
16:20always ask how do I cultivate that
16:23mindset of a builder? The judgment that
16:26you need to build that you want to build
16:30will be built by constantly tinkering
16:33and struggling through your AI side
16:37business itself. And that benefit will
16:40not show up on your balance sheet.
16:42Here's one way to think about it. If I
16:44gave you a piece of wood and asked you
16:47to make a violin, you would have no idea
16:51where to [music] begin. Neither would I.
16:53But there is a city called Cremona in
16:57Italy where violin making is a sacred
17:02art. The makers typically build only
17:05three to six violins per year. They
17:09select they shape each piece of wood
17:13according [music] to its individual
17:15acoustic response. It can take months to
17:18make a single violin. They can pick up a
17:22piece of spruce or maple and hear
17:25[music] the possibilities in that wood
17:28that the rest of us just can't.
17:32That comes after years of struggle,
17:36building, listening, refining,
17:40honing their craft.
17:43That's judgment.
17:45It's your struggles that teach you what
17:49good looks like.
17:51Thank you and I love you.

About this transcript

This page contains the full transcript of Starting An AI Side Hustle Is EASY (Copy This System) by Sandeep Swadia, generated from the public captions YouTube serves with the video. The transcript has 2,590 words across 387 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 translate the transcript into 100+ languages.

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, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.