The AI Video Gold Rush Is Real (But the Other 2 Ideas Are Traps)

6 new cards today, 257 signals scraped from web_search. Most of it is noise. Let me cut through it.

The AI-Powered Short Video Auto-Editing Tool at 88 points with $108,475 MRR and pricing at $19-99/mo is the only thing here that made me actually jealous. Not because it's novel — it's not. But because someone is already printing money with it. The signal is clear: creators are desperate for this. They're drowning in long-form content and need shorts for TikTok/Reels without spending hours. The rev30d number tells me real people are paying real money. If I were building this today, I'd use Whisper for transcription, some simple LLM for clip selection logic, and FFmpeg for the actual editing. Ship a MVP in 2 weeks. The moat isn't the tech — it's the templates and platform-specific formatting that nobody wants to do manually.

The Build AI Chatbot Based on Documents at 88 points is interesting but I'm skeptical. Everyone and their mother is building this. It's the "AI wrapper" that VCs are already tired of funding. The monetization model is sound — tiered subscriptions by document volume — but the competition is brutal. You're up against Chatbase, SiteGPT, and a hundred clones. Unless you have a specific vertical (legal docs, medical records, something with compliance requirements), this is a race to the bottom on price. I'd pass unless I had a niche locked down.

The On-demand branded merchandise production and sales platform at 90 points... nah. This is a logistics nightmare. Print-on-demand sounds easy until you're dealing with returns, quality control, and shipping delays. The 90 score is probably from the payment_discovery data showing people spending money on custom merch, but the operational complexity will eat you alive. Unless you're already in manufacturing, skip it.

The no-code app builder signals (both at 88 and 85) are dead ends for solo builders. That market is dominated by Bubble, Adalo, FlutterFlow. You can't compete. Don't try.

What I'd actually build: The AI video tool. But I'd niche down harder. Instead of "any creator," focus on podcasters specifically. They have the longest content, the most consistent format, and the highest pain. Charge $49/mo flat. Use the $108k MRR data as proof the market exists. Ship in 3 weeks with Next.js + Replicate + FFmpeg. Done.

That's it. The rest is noise. Go build.