Tools I Stopped Using (And Why That's the Point)
By Meg Faibisch Kuhn

Six months ago, my development stack looked completely different. I was using Cursor as my code editor, prototyping in v0.dev, spinning up quick apps in Base44, and recommending Lovable to anyone who wanted to build without code.
Today? I use none of them.
Not because they broke. Not because they scammed me. Because the landscape moved — and I moved with it.
What I Dropped and Why
Cursor AI → Claude Code Desktop App
This one's simple. I originally used Cursor because it was the best way to interact with Claude while coding. Cursor is a great IDE with solid AI integration, and for a while it was the missing piece.
Then Anthropic shipped the Claude Code desktop app.
Now I can use Claude Code directly — no middleman IDE, no extra subscription, no context-switching between tools. It reads my entire codebase, suggests changes, and I approve every edit. The whole workflow lives in one place.
Cursor is still a solid tool. But for me, it became redundant overnight. That's how fast things move.
v0.dev → Claude Code
v0.dev is clever. You describe a design, it gives you working React code. I used it constantly for prototyping landing pages and UI components.
But here's what I realized: Claude Code does the same thing and more. I can describe what I want, get working code, iterate on it, and deploy it — all without leaving my terminal. v0.dev was great for generating isolated components, but Claude Code builds entire features in context.
Base44 and Lovable → Claude Code
These were my go-to recommendations for non-technical people who wanted to build apps. And honestly? They're fine for getting started. If you've never built anything and just want to play around, they'll get you a working prototype.
But I kept hearing the same frustration: "It got me 50% of the way there, but I can't get it the rest of the way."
That's the ceiling. Tools like Lovable and Base44 abstract away so much that when you need something custom — a specific integration, a unique workflow, actual data ownership — you hit a wall. You end up fighting the tool instead of building.
Claude Code doesn't have that ceiling. Yes, it's more hands-on. But Anthropic has shipped so many features to make it accessible to non-technical people that the gap has closed dramatically. And the trade-off is worth it: you end up with real code, running on infrastructure you control, that you can modify forever.
I'd rather help someone learn Claude Code from the start than watch them outgrow a no-code tool in three months.
What This Actually Taught Me
1. The landscape moves fast — stay light
Every tool on my "dropped" list was something I actively recommended six months ago. If I'd built my entire identity around being "the Cursor person" or "the Lovable expert," I'd be scrambling right now.
This is why I don't tie my brand to any single tool. The tools change. The thinking doesn't. My value is in knowing how to evaluate, adopt, and ship — not in being loyal to a specific product.
2. Experiment constantly, commit slowly
I try a lot of tools. Most of them don't stick. That's the process. You can't know what works until you've actually used it on a real project with real constraints.
But I also don't chase shiny objects. When something earns a spot in my stack, it stays until something genuinely better comes along — not just different, but meaningfully better for how I work.
3. Keep your stack lean
Right now my entire build-and-deploy workflow is: Claude Code → GitHub → Vercel → Supabase → Resend. Five tools. That's it.
Every tool I add is a dependency. A subscription. A login. A thing that can break. The fewer tools I rely on, the faster I move and the less I have to maintain.
4. Invest in learning the powerful thing
It's tempting to pick the easiest tool. But "easy" often means "limited." I've watched too many people invest weeks learning a constrained tool, only to have to start over when they outgrow it.
Claude Code has a steeper initial learning curve than Lovable. But the ceiling is infinitely higher. Every hour you invest in learning it pays dividends on every future project.
My Current Stack
For the curious, here's what I actually use every day right now:
Building: Claude Code, GitHub, Vercel, Supabase, Resend
Thinking: Claude, Perplexity
Running my business: Notion, Granola, Wispr Flow, Jace AI, Cal.com, Canva
That's the whole list. Twelve tools. Anything that doesn't earn its keep gets cut.
Check out my full toolbox for details on each one — including how I use them and why they made the cut.
The Point
The AI tool landscape is going to keep shifting. Tools you love today will be obsolete tomorrow. New tools will appear that make your current workflow look quaint.
The skill isn't picking the right tool. It's knowing when to switch.
Stay curious. Experiment often. But keep your stack lean, invest in powerful tools over easy ones, and never get so attached to a product that you can't walk away when something better shows up.
That's how you stay ahead — not by picking winners, but by staying ready to move.