Here we have yet another example of companies paying people to clean up after genAI messes. This time, the messes stem from vibe coding.

(Source: 404media , "The Software Engineers Paid to Fix Vibe Coded Messes")

As this list of AI cleanup grows, keep in mind:

These are just the AI failures that make it to the news.

Just imagine the failures we don't hear about, because the companies in question have managed to hide the problems.

You might think that the takeaway lesson is to not use genAI. Not quite!

The lesson is that genAI works best when you:

• assign it tasks that it is fit to take on, and
• thoroughly check the work before releasing it to the public

If you do that, you're far less likely to hire an external company for a rush cleanup/fix job.