There was a golden moment, when AI vibe coding tools made it feel like anyone with an idea could build a MVP or heck, even a sellable end-to-end product, from scratch, solo, all by themselves. We thought we could compete with well funded teams. Tools like Cursor, Lovable, Claude Code, etc. helped ship ideas at the speed of light. 😅

But now that dream of building and running a one-person startup seems to be getting foggy. Not because of technical limitations, but because of creeping monetization, and enterprise-first approach of these AI tools.

The Rug Pull

First it was Cursor, that replaced its generous 500 requests/month with an opaque restrictive rate limit. Many Pro users, myself included, bumped into “rate limit exceeded”. No warnings. No transparency. Although Cursor team acknowledged the confusion, apologized and even issued refunds to some users. But it has now established this trend. AI tools are moving away from enabling solo builders (at least that’s what it looked like), to a focus on serving enterprise and monetization strategies.

Now Claude Code has also announced a similar “rate limit” that they claim will affect about only 5% of users. But I am pretty sure eventually it will be increased to limit further usage by retail users. Basically, make users join the party, build dependency, then pull the rug. 😅

Anti-Commodifying

I was hoping with improving capabilities of AI models, vibe coding tools will be commoditized. But instead it seems to be going in the opposite direction. Instead of these tools becoming interchangeable or cheaper, they're becoming less accessible, more opaque, and increasingly designed to maximize revenue at the expense of creative flow.

What was once a playground for solo experimentation is now a minefield. I find myself constantly second-guessing: can I afford to ask the AI for this refactor? Will this multi-step prompt burn too much of my daily/weekly quota?

Pro-Tip to Solo Builders

Never take annual Pro subscriptions of any of the AI tools. Stick to the monthly plan. You never know when they'll change their rate limits and throw you under the bus.

Going forward this will be my strategy:

• Diversify AI tool dependencies

• Budget for higher costs than initially planned

• Stay flexible with monthly subscriptions

• Build skills that don't rely entirely on AI assistance

• Support tools and platforms that maintain transparent, builder-friendly policies

The tools are evolving. Let’s just hope they don’t forget who believed in them first.