Pieter Levels launched a new project and made $10,000 in the first day.
It's been really cool to watch Danny Postma and Pieter Levels race to build and launch their profile pic generators built on top of Dreambooth.
Pieter launched and made $10,000 IN THE FIRST DAY. Danny is at $2,000+ in 12 hours.
These are the first, but they won't be the last.
I think it's important to note, as cool as AI is, that "AI" is the only new part here. The rules of building a good business remain the same.
To dig a deep moat:
• obsess over customers' needs
• iterate faster than the competition
• capitalize on network effects
• develop a strong/unique brand
That said...
The hype wave is huge and now is the time to ride it.
The technology behind these projects is called Dreambooth and it's darn easy to run.
Research projects are basically consumer-ready apps at this point that can be run in the browser.
Slap a nice UX on that puppy and you're off to the races.
Here's a tutorial I found: https://techpp.com/2022/10/10/how-to-train-stable-diffusion-ai-dreambooth/
Ideas for your next $10,000/day business💡
• Virtual clothes try on
• Preview hairstyles
• Insert yourself/friends into memes
• Next-gen school portraits/pet photography
What else can we build? Who else is building a new business on top of stable diffusion?
EDIT:
Found another tutorial that looks good: https://biyo.notion.site/biyo/How-to-create-your-AI-generated-avatars-for-free-no-coding-needed-84c98ce95405478cb8dda4fb8b9d4f29
Protect your momentum like your life depends on It
We are live on Product Hunt
Don't be a Jerk. Use this Tip Calculator.
Opsgenie vs. Splunk: Choosing the Right Incident Management Solution
It's interesting to me how many people here are whining that @levelsio has a big audience and saying that his products aren't "truly tested" because of it.
For one, indie hacking is not supposed to be some sort of fair competition with equal starting lines. In fact, it's not any kind of competition at all. It's the lawless wild west. You shouldn't be comparing your resources with someone else's. Instead, you should be using every resource at your disposal to find success, because the rules of the game are that there are no rules. Anything goes. If you have a rich uncle who can fund you, go for it, even though most people don't have that. If you were born in a first-world country with a first-world education, leverage that, even though many people don't have that. If you were born with above-average intelligence, use it, even though 50% of people cannot.
When you think in terms of what's "fair," you handicap yourself. You limit yourself to the "standard" channels because you are a "standard" indie hacker. But it's better to get creative and ambitious and break free of these shackles. When I launched Indie Hackers itself, I had zero audience. But I still leveraged Levels' huge audience, by interviewing him and doing a good job and getting him to tweet about IH several times. Just because somebody else has something you don't, doesn't mean there's no way for you to make use of it.
Secondly, the idea that, "I couldn't do that because I don't have Levels' audience," is missing the entire point. It's not like he was just handed his audience. He built it! He didn't make $10k in one day, because this didn't happen in one day. It took years and years of hard work, including the authority and attention and skills he built while working on InteriorAI in recent months, and including his constant building and tweeting over the past 5-10 years. Calling this a "one-day success" is like calling Usain Bolt's 100m world record a "10-second" success. It's insulting because it ignores the tireless years of work that were 99.999999% responsible for his success and only looks at the last 10 seconds.
If you're saying you couldn't have done this because you don't have Levels' name/audience, you're wrong. The reality is you can do this, because you absolutely can build your own name and audience up the way he has. You just have to put in the same time and effort he has. Simple, but not easy.
This. 1000% this.
Everyone complaining that Pieter Levels is only successful because of his current audience is missing the entire point that EVERY ONE OF US could be Pieter Levels (or @csallen) by simply following the steps he literally and publicly lays out for us: continued, consistent effort + humble efforts to please customers + humble promotion + quickly building & launching just-barely working versions of an unproven idea over time will 1. build the 'unfair' audience, 2. weed out garbage ideas, and 3. produce long-term sustainable income that dwarfs what most people earn from a 9-5.
The big difference between Pieter Levels and many of us, is the CONSISTENCY and LAUNCHING parts. Many of us don't consistently work on our projects long enough to validate or launch, then get depressed or angry at ourselves or at others who put the time in.
A hundred times this.
Leverage the heck out of every unfair advantage you can. There are people hustling just trying to get to wherever you already are. Don't waste the opportunities you already have.
Not knocking it in anyway. But the true test of the product(s) would be if a first time founder, unknown to most, launched the product(s). And done the rounds, and launched it well.
Thus if it didn't have Pieter Levels name attached to it would it have made $10,000 in the first day?
Yeah. Pieter is an inspiration and has absolutely earned his success but I'm pretty sure at this point he could make a blank screen with a stripe checkout and get 10k MRR .
Couldn't have said it better
I suppose so. But Pieter started out like any of us, with zero audience and he still managed to earn $$ from JustF*ckingDoIt, FlipBook, and then larger slow & steady growth from NomadList and Remote OK and Make Book.
Sure $10k in a day is going to be due to a larger pool of potential customers from his audience, but I'd be willing to bet he'd still have made a sizable profit - albeit a bit slower - if he didn't have the audience.
He's proven from the beginning that he knows how to find and ride a wave of a new trend and build things for it to profit from - with or without his current audience.
I think we agree overall. It's not a 1 day success and shouldn't be branded as one. He no doubt busted his ass for years to be able to do this.
The push back here is because the claim is that anyone can get in on the hype, but uses two prominent examples (one reasonably famous, and one with a previous 7 figure exit in the niche) as case studies.
That's all the main comment of this thread said; it implied bias in the outcome due to another significant variable.
These inherent advantages should absolutely be embraced by those who have them, but they should also be pointed out in these success stories to balance out the survivorship bias. Push back on these headlines, as inspirational as they are, is healthy IMO. Lest people start to think they too can hit a few thousand MRR day one as an unknown, launch, hear crickets, and give up too early because their expectations were out of whack.
With a big audience it's always easier.
I'm sure the idea would have generated revenue anyways, but much slower and with an ad campaign or something like that
(copied from another response)
It's a perfect storm of hype wave + natural virality + pre-existing audience.
The same project launched a month from now or without the pre-existing audience would not have the same results.
If you've got the ability to execute quickly, rn is the time to do it, but you're probably going to find some other way to kickstart the viral mechanism:
That's a framework I'm sure we all want.
But the most important part is... its Levels.
Danny made his bones with Headlime, probably not as known as Pieter. And that'll reflects in the sales of each.
Although not related to AI, I can't see many people buying a perfume called Burnt Hair, if a certain Chief Twit wasn't selling it.
I bought Burnt Hair :) And I was going to make the same exact comment with the same reference. I would not have bought Burnt Hair if it wasn't made by Mr. Musk.
The product here is pretty cool, and the tech is neat. But the only reason it made $10k in a day is because of the person/audience. I think there's a lot to be learned from that, like building an audience, etc.
I've noticed in general articles/posts here tend to have clickbait titles and get lots of responses because there is some amount of misleading disconnect between the title and content. Anyone can build that audience, but that aspect of it will take years. I think the title actually minimizes the years of work put in to create the audience.
There is one way around this, which is to partner with someone who already has an audience and split the proceeds, such as with a YouTuber or someone with a related blog. But the OP didn't cover any of that, which is why I think the reactions are warranted.
I hear you.
What does it smell like? If you say Burnt Hair then this just bolsters the debate :) PS tell me you bought the Boring Company Flamethrower. I'd buy that even if Elon wasn't involved.
cold outreach, content marketing & paid ads is Pieters framework?
For sure not every product works without an audience. I don't think this falls into that bucket.
The product is sellable regardless of who's selling in this case.
Even Burnt Hair works as a gag-gift. Could see it doing some numbers around Halloween, Christmas, and April Fools with the right Facebook Ad Campaign.
I'll have to agree to disagree on that.
Sure the product may be sellable . But the zeitgiest surrounding AI can also be attributed to that. (See Charles McKay's 'Extraordinary Popular Delusions and the Madness of Crowds').
But the main denominator that caused it sell $10,000 in the first day (which is the hook here as its been used x 3) is not that it involves AI (which again may be attributable, certainly piques interest). Its because Levels is selling it - and the audience (hes rightly built up with great businesses/projects in the past) and even the blind faith that goes in hand in hand with them.
Totally agree and wish more people would point things like this out before preaching it in a way that almost "every" average founder could do this. Maybe after the second or third launch of a new SaaS but most first time founders dont have this audience.
Very true.
Tying in with the conversation on Twitter
I imagine that most people familiar with Pieter Levels immediately translate this headline and article to 'Person with a gigantic audience and reach unsurprisingly sells a lot of a trendy thing quickly', but I think all the haters and cynics should also take note of the power of the audience he has built for himself - it is incredible, should be commended, and it would be crazy not to be relentlessly capitalising on it :)
Pieter is amazing, remember he launched 70+ businesses/projects before he started his ascension to success...
"Overnight success 10 years in the making"
Wow, I never think he has done that many projects, amazing!
Wrote about Dreambooth here - https://mythicalai.substack.com/p/what-is-dreambooth-and-how-to-use
Super interesting how a fork of a fork of Stable Diffusion spread so fast and how it solves the repeatable issue generative art AI has.
You could monetize anything with that audience.
Correct, feel free to launch 70+ products over 5-10 years.
My thoughts exactly!
And that's what consistent hard work gets you over 10 years.
"Virtual clothes try on"
Flash back to 1983. Jan Derksen, my favorite CS professor, proposed the same idea, and took it further. Find clothes you like. Upload your honest measurements (a potential weak link and a worthy UX design challenge).
BOOM! Robots create cost-effective, tailored clothes just for you.
Sorry, no refunds. ;)
Did he actually provide this service? Or was it concept-only?
Dana Todd's business sounds like a modern equivalent. https://twitter.com/danatodd
Concept only, and way ahead of its time. (I miss Jan. He was the most broadly intelligent person I've ever met. Taking 4 courses from him was a blessing.)
And thanks for the link to Ms. Todd's site. Very cool!
Whoa, I found this in an Email newsletter ~10 seconds after that last comment. Talk about serendipity... https://beta.unspun.io/
For all those commenters saying "yeah but Pieter Levels has an AUDIENCE!! and it's that's so UNFAIR!!":
from @levelsio on Twitter:
"It took building 4x AI projects over 2 years to get a hit again:
$0 http://ideasAI.com (startup ideas by AI) in 2022
$0 http://thishousedoesnotexist.org (house exteriors by AI)
$$$ http://interiorai.com (interior designing by AI)
$$$$$ http://avatarai.me"
So go spend 10 years building an audience, and 2 years building within a niche space, and launch your projects.
(couldn't paste the screenshot :/)
Just shows how building products is about practicing the right mindset and iterating...iterating...iterating...iterating
by his own admission " >95% of everything I ever did failed"
Nice discussion on the comments, more of this please!
This is so cool! I've been trying to figure out how to automate this process.
Is there any way I could use Google colab via API?
Or any other rentable GPU and run it on a server?
Super exciting times, and I've so many ideas I want to try!
Happy Building! Cheers!
Yea that first tutorial does everything in colab
Super new to Google Colab; how can I run that via API, though? It seems like Google Colab does not have that option. Or maybe, I am not aware of it :D
Any suggestion/direction is appreciated! Cheers!
Hey Anthony, really nice post. Brings lots of value! Thanks for the links and tutorial.
I just subscribed to your newsletter ;)
Thanks Florian, appreciate you :)
Very well written, Anthony, I think it's super interesting.
Within AI Copywriting it was also the first ones who came with a proper product and marketed, they won.
Interesting times ahead 🚀
Thanks Phillip, it's exciting times to be indie hacking
I am personally running InvokeAI locally if I need any AI generate images, it works great and also has a web interface, it's based on Stable Diffusion: https://github.com/invoke-ai/InvokeAI
The difference in results for s a reflection on f who is building the product, it matters a lot.
Also, I got the second one to work yesterday, super fun.
No need to pay at all...
Audience really matter alot in terms of building a product. That one of the advantages Pietier levels has.
If that's the case, How did he gain success with his early projects when he didn't have an audience?
What do they actually offer to the people? Just the generated pictures? or do they store the model and allow people to generate whatever pictures of themselves? Storing each user's models and generating the pictures on demand is not easy and not scalable.
Why not? If it costs less than $39 per user then they will be profitable.
I'm not sure how the user retrieves the images, but I think I read on Pieter Levels' twitter that he deletes images after 24 hours per his Terms of Service. Could be wrong of course, plus I think it has something to due with privacy.
It's the wild west in generative AI web app building and there are so many untapped markets. Currently most of the generative AI apps are general purpose tools. That means there are so many specialized tools to yet to be built!
Yea, it's a huge opportunity. Recently started a newsletter just to give myself some focus and a reason to dig into the trends. https://boteatbrain.com/
Writing about what I learn while I hunt for ideas to build. 20% jokes, 80% informative.
Subscribed :)
Welcome aboard 🫡
Let me know what you think of the next issue
Their stories a definitely inspiring to the audience here. I'd guess that building a similar project "out of nothing" is the dream to many here.
It's a perfect storm of hype wave + natural virality + pre-existing audience.
The same project launched a month from now or without the pre-existing audience would not have the same results.
If you've got the ability to execute quickly, rn is the time to do it, but you're probably going to find some other way to kickstart the viral mechanism:
Wow! mind-blowing... Thanks for sharing!
Very inspiring to watch it all unfold. I keep wondering how he identified the need for this? I might have missed some of his initial tweets, but did he test in on himself and then share (which drove people to want to try it out as well)? Very curious to know....
He was working on a similar project before hand https://interiorai.com/
Not sure what exactly led to the pivot.
I'm writing https://boteatbrain.com/ now in order to try and spot these trends early and help indie hackers get ahead of the curve.
Thanks for responding. I subscribed to boteatbrain last week. I look forward to learning more.
Sweet :D
Should be publishing the next issue soon
SOLID post.
I have a dev agency criov.com and Thats why i am trying hard to have big audience cuz once you build a community the rest is easy
Good for them
wooo
pietel has been my inspirational since 2020
So I utilized one of the instructional exercises yesterday and it worked perfectly. This might be idiotic however is there a method for utilizing that again today without preparing everything everywhere? Like is there something saved in my drive I can open up without going through the whole cycle (or if nothing else preparing part) briefly time? I saved the Dreambooth colab from yesterday however don't see the connection any longer...
So I used one of the tutorials yesterday and it worked great. This may be dumb but is there a way to use that again today without having to train everything all over? Like is there something saved in my drive I can open up without having to go through the entire process (or at least training part) for a second time? I saved the Dreambooth colab from yesterday but don't see the link anymore...
This comment was deleted 2 years ago.