From Physics PhD to $100K/month AI Music Video Platform
TL;DR: Nico studied physics and completed a PhD despite not being passionate about it. In 2018, he discovered he loved programming while running scientific simulations. After joining a deep tech startup in 2020 where he learned AI and computer vision, he started Neural Frames in late 2022 as his first full-stack project. The product combines his two interests: computer vision and music. Initially positioning as 'text to video for everyone', he struggled until a friend advised him to niche down specifically to musicians. This single decision became a huge unlock. He launched on Hacker News within a week of going live, reaching #6 on a Sunday which brought first internet money, backlinks, and 350 concurrent users. He built in public aggressively, putting his photo everywhere, recording YouTube tutorials, and adding footer text like 'no VC money, just a tiny company in love with text to videos'. This personal branding created a flywheel effect. Now a team of 5, Neural Frames serves 1,500 paying customers ($100K MRR), has 100K monthly active users, and has generated 1.5 million AI videos. Major costs include $45K/month for GPUs and text-to-video APIs (almost half of revenue). His advice: don't be scared to switch careers, AI tools make building easier than ever, and customers pay for problems solved, not the tech.
Key Insights
- Niching down from 'text to video for everyone' to 'AI music videos for musicians' was the biggest unlock - it made the product feel purpose-built instead of generic
- Launched rough MVP on Hacker News within one week, reached #6, which brought first revenue, backlinks, and valuable feedback from technical audience
- Playing the indie hacker card aggressively (founder photo everywhere, 'no VC money' in footer, personal YouTube tutorials) created connection and trust that drove word-of-mouth
- Combined two existing skills (computer vision from PhD work + music hobby) to find unique product opportunity where interests intersect
- Bootstrap advantage: can target smaller markets that VCs wouldn't fund, giving freedom to build sustainable business without billion-dollar pressure
Actionable Takeaways
- If you have a broad AI tool, niche it to one specific use case - users should instantly see themselves in your product, not have to imagine how they'd use it
- Launch your MVP on Hacker News within the first week, even if it looks terrible - the technical crowd appreciates scrappy builds and early launches
- Use SEO tools (like Ahrefs) to validate ideas: look for keywords with 1,000+ monthly searches and low difficulty - people searching + few solutions = opportunity
- Build in public by making your founder story visible everywhere - put your photo on the site, record video tutorials yourself, share your indie journey in the footer
- Follow the 'fuck around find out' principle: do many random things across different domains, then watch for magical intersections where your skills combine into unique opportunities