"Stop building useless stuff" — a rant against the AI SaaS rush
The post argues that most SaaS products, directory sites, and chatbots built quickly with like Cursor won't survive more than a few months. It criticizes the common pattern of using an overkill, near-identical (Next.js, Supabase, PostgreSQL, deployed on ) as if that alone guarantees success. Claims like "I hit $10k MRR in 30 days" or "I built a million-dollar SaaS with AI" are called mostly exaggerated or outright false.
The current AI-tools gold rush is compared to past hype cycles around crypto, NFTs, and — different costume, same empty promise. The people actually making consistent money, it argues, aren't the builders but those selling the dream that anyone is "one AI prompt away from financial freedom," and they don't even need real expertise to do it.
Key points
- Most AI-coded SaaS/directory/chatbot projects (built with tools like Cursor) reportedly fold within a few months
- Criticizes the near-universal overkill stack: Next.js + Supabase + PostgreSQL on
- Claims like "$10k MRR in 30 days" are mostly called exaggerated or false
- Argues the real money is made by people selling the dream of easy AI success, not by the builders themselves
- Compares the current AI SaaS hype to past crypto, NFT, and cycles