In this follow-up, I explore "prolific AI psychosis," where users generate massive AI output without real value. Unlike true psychosis, this is a milder disconnect from reality, marked by an inability to judge quality.

For example, a developer might produce thousands of code lines daily, yet the code lacks utility. A productive dev writes less but meaningful code. Similarly, an author writing 100,000 words a day likely sacrifices care compared to 1,000 quality words.

The core issue isn't output volume but perception—the user can't assess their own work's worth. Current AI tools, like chatbots, are simple, but advanced agent harnesses tackle complex tasks, making this psychosis more prevalent.

These agents can build apps or fix bugs autonomously, but they also make poor decisions. The result: impressive feats alongside broken software, highlighting the need for critical oversight.