TheValueProvider•This week
10 Side Projects in 10 Years: Lessons from Failures and a $700 Exit
Hey folks, I'm sharing my journey so far in case it can help others. Entrepreneurship can sometimes be demotivating. In my case, I've always been involved in side projects and what I've realized is that every time you crash a project, the next one makes it a bit further. So this is a long-term game and consistency ends up paying off
The $1 Android Game (2015, age 18)
What Happened:
500 downloads, 1€ in ad revenue
Ugly UI, performance issues
Key Lessons:
Don’t be afraid of launching. Delaying for “perfection” is often a sign that you fear being ignored. I was trying to perfect every aspect of the game. In reality, I was delaying the launch because I feared no one would download the app.
Commit to the project or kill it. At some point, this project was no longer fun (it was just about fixing device responsiveness). Most importantly, I wasn't learning anything new so I moved to smth else.
The Forex Bot Regret (2016, age 19)
What Happened:
Lost months identifying inexistent chart patterns
Created a Trading bot that was never profitable
Key Lessons:
Day trading’s real winners are usually brokers. There are plenty of guys selling a bot or systems that are not making money trading, why would they sell a “money-printing machine” otherwise...
Develop an unfair advantage. With these projects, I developed a strong coding foundation that gave me an edge when dealing with non-technical business people. Invest countless hours to create a skills gap between you and others, one that becomes increasingly difficult for them to close (coding, public speaking, networking, etc.)
The $700 Instagram Exit (2018, age 21)
What Happened:
Grew a motivational account to 60k followers
Sold it for $700
90% of followers were in low-income countries (hard to monetize)
Key Lessons:
Follower quality > quantity. I focused on growth and ended up with an audience I couldn’t truly define. If brands don’t see value, you won’t generate revenue. Also, if you do not know who you are creating content for, you'll end up demotivated and stop posting.
Great 3rd party product + domain authority = Affiliate marketing works. In this case, I could easily promote an IG growing service because my 50k+ followers conveyed trust. Most importantly, the service I was promoting worked amazingly.
The Illegal Amazon Review Marketplace (2020, age 23)
What Happened:
Sellers were reimbursing buyers for positive reviews
Built a WordPress marketplace to facilitate “free products for reviews”
Realized it violated Amazon’s terms
Key Lessons:
Check for “red flags” when doing idea assessment. There will always be red and orange flags. It’s about learning to differentiate between them (e.g. illegality, 100% dependence on a platform, etc.)
If there’s competition, it’s good, if they are making money it’s even better. I was thrilled when I saw no competition for my “unique idea”. Later, I discovered the obvious reason.
Copying a “Proven” Business Model (2020, age 23)
What Happened:
Tried recreating an Instagram “comment for comment” growth tool
Instagram changed the algorithm and killed the growth strategy that the product used.
Key Lessons:
Do not build a business that depends 100% on another business, it is too risky. Mr. Musk can increase Twitter on API pricing to $42,000 monthly without notice and Tik Tok can be banned in the US. Due to the IG algorithm change, we had built a product that was not useful, and worse, now we had no idea how to grow an IG account.
Consider future project synergies before selling. I regret having sold the 60k follower IG account since it could have saved me a lot of time when convincing users to try the service.
NFT Marathon Medals (2021, age 24)
What Happened:
Created NFT race medals
Sold 20 for 5€ each, but spent 95% of meetings explaining “what is an NFT?”
Key Lessons:
Market timing is crucial. As with every new technology, it is only useful as long as society is ready to adopt it. No matter how promising the tech is in the eyes of SV, society will end up dictating its success (blockchain, AI, etc). In this case, the runner community was not ready to adopt blockchain (it is not even prepared today). Race organizers did not know what they were selling, and runners did not know what they were buying.
The 30-day rule in Fanatical Prospecting. Do not stop prospecting. I did prospecting and closed deals 3 months after the outbound efforts. Then I was busy executing the projects and had no clients once the projects were finished.
AI Portal & Co-Founder Misalignment (2023, age 26)
What Happened:
Built a portal for SMEs to find AI use cases
Co-founders disagreed on vision and execution
Platform still gets \~1 new user/day
Key Lessons:
Define roles and equity clearly. Our biggest strength ended up killing us. Both founders had strong strategic skills and we were constantly arguing about decisions.
NextJS + Vercel + Supabase: Great stack to create a SaaS MVP. (but do not use AI with frameworks unless you know how they work conceptually)
SEO is king. One of our users creates a use case on “Changing Song Lyrics with AI.” Not being our target use case, it brings 90% of our traffic.
Building an AI Tool & Getting Ghosted (2024, age 27)
What Happened:
SEO agency wanted to automate rewriting product descriptions
Built it in 3 weeks, but the client vanished
Key Lessons:
Validate manually first. Don’t code a full-blown solution for a problem you haven’t tested in real-world workflows. I kept rewriting code only to throw it away. Jumping straight into building a solution ended up costing more time than it saved.
Use templates, no-code, and open-source for prototyping. In my case, using a Next.js template saved me about four weeks of development only to hit the same dead end, but much faster.
Fall in love with your ICP or walk away. I realized I didn’t enjoy working with SEO agencies. Looking back, I should have been honest with myself and admitted that I wasn’t motivated enough by this type of customer.
Ignoring Code Perfection Doubled Traffic (2025, age 28)
What Happened:
Partnered with an ex-colleague to build an AI agents directory
Focused on content & marketing, not endless bug fixes
Traffic soared organically
Key Lessons:
Measure the impact of your actions and double down on what works. We set up an analytics system with PostHog and found wild imbalances (e.g. 1 post about frameworks outperformed 20 promotional posts).
You have to start somewhere. For us, the AI agents directory is much more than just a standalone site, it's a strategic project that will allow us to discover new products, gain domain authority, and boost other projects. It builds the path for bigger opportunities.
Less coding, more traction. Every day I have to fight against myself not to code “indispensable features”. Surprisingly, the directory keeps gaining consistent traffic despite being far from perfect
Quitting My Job & Looking Ahead (2025, age 28)
What Happened:
Left full-time work to go all-in
Plan to build vertical AI agents that handle entire business workflows (support, marketing, sales)
Key Lessons:
Bet on yourself. The opportunity cost of staying in my full-time job outweighed the benefits. It might be your case too
I hope this post helps anyone struggling with their project and inspires those considering quitting their full-time job to take the leap with confidence.
![[D] Elon Musk has a complex relationship with the A.I. community](https://jfbmhhfxbbrxcmwilqxt.supabase.co/storage/v1/object/public/resource-images/MachineLearning_AI_for_business_20250328_183449_processed_image.jpg)
![Looking for an accountability partner as a solo founder. [I will not promote]](https://jfbmhhfxbbrxcmwilqxt.supabase.co/storage/v1/object/public/resource-images/startups_AI_for_business_automation_20250328_183753_processed_image.jpg)
![[D] Next big thing in IT is ML/AI. It used to be the Internet. Do you think this would make it harder for amateurs to start a business in IT or would be much the same?](https://jfbmhhfxbbrxcmwilqxt.supabase.co/storage/v1/object/public/resource-images/MachineLearning_AI_for_business_20250328_183423_processed_image.jpg)
