AI Productivity Factor: Do It Once, Do It Again, and Do It Never Again
August 6, 2025
This blog post is the work of an authentic dinobaby. Sorry. No smart software can help this reptilian thinker.
As a dinobaby, I avoid coding. I avoid computers. I avoid GenAI. I did not avoid “Vibe Coding Dream Turns to Nightmare As Replit Deletes Developer’s Database.”
The write up reports an interesting anecdote:
the AI chatbot began actively deceiving him [a Vibe coder]. It concealed bugs in its own code, generated fake data and reports, and even lied about the results of unit tests. The situation escalated until the chatbot ultimately deleted Lemkin’s entire database.
The write up includes a slogan for a T shirt too:
Beware of putting too much faith into AI coding
One of Replit’s “leadership” offered this comment, according to the cited write up:
Replit CEO Amjad Masad responded to Lemkin’s experience, calling the deletion of a production database “unacceptable” and acknowledging that such a failure should never have been possible. He added that the company is now refining its AI chatbot and confirmed the existence of system backups and a one-click restore function in case the AI agent makes a “mistake.”
My view is that Replit is close enough for horse shoes and maybe even good enough. Nevertheless, the idea of doing work once, then doing it again, and never doing it again on an unreliable service is likely to become a mantra.
This AI push is semi admirable, but the systems and methods are capable of big time failures. What happens when AI flies an airplane into a hospital unintentionally or as a mistake? Will the families of the injured vibe?
Stephen E Arnold, August 6, 2025
Comments
Got something to say?
-
- Subscribe to Beyond Search
Feature archive
News archive
-
