From a developer's perspective, this is a double-edged sword. On one hand, the speed of deployment is insane. You can move from a concept to a deployed build in a few hours using a tight AI workflow. On the other hand, "vibecoded" apps often lack architectural rigor. They work until they don't, and then the "developer" (who doesn't actually know the codebase) is stuck.
For Apple, this is a volume play. More apps mean more developers paying the annual fee and more potential for niche utilities that wouldn't have existed otherwise. However, the risk is a diluted user experience. If the store gets flooded with fragile, AI-generated clones that crash or leak memory because the prompts weren't specific enough about state management, the overall quality of the ecosystem dips.
The real winners here are those using a hybrid approach. Instead of pure vibecoding, using tools like Claude Code or Cursor to handle the boilerplate while maintaining a strict eye on the logic is where the actual productivity gains are. It's not about removing the coder, but about upgrading the coder into an architect who manages an AI workforce.
Next Open Source Web Scraping: A Deep Dive →
All Replies （4） #
@Max75Fair point, but this feels different since now people with zero dev skills can flood the store in seconds.
source & further reading
promptcube3.com — original article
DeepSQL: Self-Hostable DBA Agent for Postgres & MySQL
Gary Marcus's Critique of LLM Logic
Coding Agents vs. The Craft: Why I'm Slowing Down