Building a live streaming platform solo: what I underestimated

I've been building Circuit (gocircuit.tv), a live streaming and creator monetization platform, as a solo developer for the past couple of years. Coming from a 20-year background in broadcast production (multi-cam live events, ESPN+, that whole world), I figured I knew what I was getting into on the streaming side.

Things I underestimated:

▸ Streaming infrastructure is a full-time job by itself. Running Ant Media Server, tuning WebRTC vs HLS tradeoffs, managing bandwidth, handling failover - none of this is glamorous and all of it can break a launch.

▸ The "platform" isn't the streaming. It's the dashboard, the analytics, the payout flow, the support tooling. The video is maybe 20% of the actual product surface.

▸ AI-assisted dev is the only reason this is possible solo. Cursor and Claude Code aren't a productivity boost -- they're the difference between a one-person team and needing four people.

▸ Creators don't care about your stack. They care about three things: does it work, do I get paid, and do I look good doing it.

Question for other solo builders here: what's the part of your product you didn't realize would eat 30% of your time?