AI Hype, Broken Tools, and What Actually Works
Cutting through the noise — what AI tools actually deliver on their promises and which ones are just riding the hype wave.
Every other day there's someone on my YouTube feed claiming they made $2,000 a month with an app they built in 20 minutes using AI. And I'm just sitting here thinking — really? Based on my actual experience of building apps and software, it doesn't feel real.
The Thumbnail Problem
My wife and I had this conversation yesterday about something as basic as creating a YouTube thumbnail. Everyone says AI does everything now. So she tried it. Told the AI what she wanted. And she still had to edit a whole bunch to get it to look right. So what's the point? Why do they say it's so easy?
That's the same story with Lovable, Replit, Cursor — all the AI coding tools. We use them a ton. They're great. But do they work for everything and for everyone as fast as people are claiming? Not really.
Where It Actually Works
Here's where I've seen AI genuinely deliver value:
Rebuilding From Scratch I had this experience where I spent four hours trying to debug adding Google Analytics to an existing project using Cursor. Couldn't figure it out. Then I said "forget it" and prompted a brand new project from scratch — the whole thing, with Google Analytics baked in — in the same amount of time. That was my aha moment.
Quick Prototyping The kids building million-dollar businesses on nothing — they don't care, they don't wait, they don't plan. They just do it. Try, fail, try, fail, and something pops. We lose that urgency as we grow. There's something to learn from that energy.
The Disposable Mindset When building something only takes 15-30 minutes and saves hundreds of hours, who cares if it's not perfect? You'll rebuild it when the models get better anyway. The app becomes disposable.
What Doesn't Work
Expecting AI to handle complex, nuanced business logic without guidance. Expecting it to "just figure it out" when your requirements are fuzzy. And especially — expecting it to maintain and debug itself when things break in production.
The tools are incredible. But the hype is way ahead of the reality for most people. If you go in with realistic expectations and use AI as a force multiplier rather than a magic wand, you'll get real value out of it.