Had a heated debate with my co-founder Kalpan Shah. I was hyped about tools like Lovable, bolt.new, and Firebase Development Company (Firebase Studio AI). “These are the future,” I told him. “We can build anything with them.” But he pushed back — hard. “Everyone's getting seduced by slick demos,” he said. “Just because a tool can turn text into code doesn’t mean it’s ready for real dev work.” And... he had a point. I’ve been playing with these tools. Built a few basic apps — a to-do list, a note-taker. (The stuff YouTube influencers love to showcase.) So Kalpan challenged me: Build a 10-feature MVP for an ATS (Applicant Tracking System). Easy, right? Except... it wasn’t. Hours burned. Bugs everywhere. Weird auth issues. Bizarre logic bugs. And the biggest one? No clear way to fix them unless you really understand software engineering. Kalpan was right. These tools generate beautiful UIs in seconds. But when the real world hits — edge cases, integrations, backend weirdness — the magic fades fast. He’s still exploring and learning a ton from AI tools. But here’s our shared insight: The gap between what these tools promise and what they deliver — that’s where the churn will happen. It’s not that they’re bad. They’re just... not the shortcut people think they are. Curious to hear from real devs here: Are we being too harsh? Or is Kalpan spot on — and most influencers are overselling the dream? Let’s talk in the comments. P.S. If you’ve tried building something real with these tools — drop your experience below.
Download the medial app to read full posts, comements and news.