Aytekin Tank has been writing software since before vibe coding was a concept. He founded Jotform 20 years ago after getting fed up building web forms by hand. When someone with that background says AI-assisted coding is worth taking seriously, it is worth reading why.
Where vibe coding earns its place
Tank argues the biggest unlock is for founders who are not developers. A plant care app idea, a custom reservations tool for a restaurant, a client-facing dashboard for a consultant: all of these used to require finding a technical co-founder or hiring out. Now they do not. The idea-to-prototype gap has compressed significantly.
For internal tooling, the argument is even cleaner. Tank used Claude Code to build a personal fitness and nutrition app for himself. He has no plans to monetize it. It just saves him time. That framing matters: vibe-coded tools do not have to scale or be perfect. They need only to make one part of life easier.
The security number you should not skim past
Tank cites a Daily.dev security scan of 1,645 web apps built with Lovable. The finding: around 45% of AI-generated code contains security weaknesses, and 120 of the scanned apps exposed users’ personal data through critical vulnerabilities. Authentication, payment processing, and encryption still require human expertise that AI models cannot reliably provide, in Tank’s view.
The operator takeaway
Tank’s position is not that vibe coding replaces technical skill. It is that vibe coding works best when someone who understands the underlying logic is directing it, not just accepting whatever the AI produces. Use it to build an MVP fast and automate repetitive internal tasks. Treat anything touching security or money with the same scrutiny you would apply to hand-written code.
