Yep. Pretty much all my work on Bubble are rescue missions, where a founder has started scaling a Bubble app and realised itâs really inefficient, really buggy and stressful to operate. Thatâs when I get called in and mostly every time I see a business with a good idea that is a complete mess and needs tons of refactoring.
I actually like doing it. I like the challenge. I like understanding the business model and improving the app with in that. Itâs one dopamine hit after another.
@Newed, great post, and probably the most honest summary of this Iâve seen. I agree with almost all of it. The one place Iâd push back is the framing that Bubble was mostly just trying to help founders find good developers. Bubble has more direct responsibility for the âexpertâ problem than that, and I think itâs worth saying plainly.
For years Bubble built an entire apparatus of credibility: agency tiers, Agency of the Year, certification, marketplace profiles, partner badges. A non-technical founder cannot evaluate architecture, database design, security, or maintainability. That is the entire reason theyâre hiring help. So they do the only thing they can: they rely on what Bubble appears to vouch for. The badge becomes the diligence.
The problem is how little sits behind some of those badges. As you noted yourself, certification largely validates platform familiarity, not the ability to architect something scalable and maintainable. So what a founder reads as âsafeâ and what the credential actually measures are two very different things.
The 2025 Agency of the Year thread is the clean public example of where that leads. Publicly awarded Gold agency. A founder reportedly spent roughly $80k over two years and walked away with nothing usable, and the app was rebuilt from scratch. That is not a âsoftware is hardâ outcome.
And I want to be direct about this part, because the politeness around it is the actual problem: this is NOT one bad project. Iâve personally dealt with other cases like it, and Iâve heard about more, in detail and from people whose judgment I trust, involving agencies this community would unhesitatingly call big or reputable. We all know the stories. We hear them privately, we nod, and then we say nothing, because naming an agency out loud is treated as the unprofessional move. That silence protects the agencies. It does nothing for the next non-technical founder about to hand one of them $80k on the strength of a badge.
The incentive runs one direction. If the client canât tell a foundation from a pile of screens, an agency can bill progress long after it should be clear thereâs no product underneath. When it collapses, the blame lands on the client: unclear scope, not enough ownership, difficult process. But preventing exactly that was the job they were hired for.
Hereâs where AI changes the math in a way thatâs uncomfortable for everyone, me included. If a founder was going to be sold the appearance of progress anyway, they can now buy that appearance for $800 instead of $80k. The vibe-coded app obviously wonât work either. But it looks polished and costs almost nothing, so the founder gets the same illusion at 1% of the price.
One caveat I havenât seen mentioned: The cleanup, the rescues, the rebuilds were a real chunk of senior work, and that work only existed because the bad builds did. AI didnât just take the bad agenciesâ lead flow. It dried up the rescue economy that good devs depended on. Thatâs why the crunch feels structural rather than like demand simply moving to a new tool. The whole arrangement is unwinding at once.
So yes, AI shifted demand and yes, vibe coding is overhyped. But Bubble canât fully separate itself from the reputational damage. When its own endorsements and awards lead to a zero-value outcome, the founder doesnât walk away doubting the agency. They walk away doubting Bubble.