What the comments here fail to understand is that even if 80% of these is vibecoded slop, 20% of them isn't (classic Pareto), and that's still a ~17% increase of non-slop. Quite a lot.

An LLM being used to generate the code doesn't say anything inherent about the quality of the app. It just changes the distribution. It's like how knowing a person is a man makes it a lot more like they're into watching team sports, but it doesn't mean they actually are. Or that a woman isn't. As of April 2026, an app's code being primarily LLM written it's a lot more likely to be slop, but there's plenty that aren't slop. Similarly, there were millions of slop apps out there before anyone out there was using LLMs to code.

More and more people will start agreeing with this take and admitting it over time as the Overton Window shifts and it becomes more acceptable. But this has been true for 2 years now. The distribution has shifted, ironically towards a higher % of slop, because 2 years ago harnesses and other tooling weren't good enough yet to allow people with zero coding experience to build an app. It still took more effort, higher barrier. But people were already generating useful code with LLMs, and yes, even creating useful, non-slop, thought out apps whose code was majority LLM-generated.

"Show them! Where are they!?". Look at these comment sections, then take a guess why we aren't linking them everywhere and shouting it from the rooftops. Because it hasn't reached mainstream acceptance, so it's bad for business. People immediately associate it with slop, and they're not necessarily wrong in doing so because as said, 80% is slop. So for us 20%, who didn't vibecode something in a day but worked on it for months to produce something actually valuable, all it does is invite negativity with business impact. In offline, informal settings you'll find way more people who are willing to say it and indeed show.