The Pitfalls of Inept AI Usage
People are using AI ineptly to produce large batches of untested or test after code. This is inflating output based measures of productivity while actually slowing delivery down and reducing quality.
It creates massive, poorly understood changes riddled with large scale code duplication. This is increasing qa time, failures in production leading to more time in support, troubleshooting and maintenance. Which is then addressed with more poorly understood changes riddled with large scale code duplication.
This is an organizational change problem combined with a lack of understanding and respect for engineering. People made to use rapidly evolving tools pervasively before they have learned how to use them well. Then blaming the failure on the tools and people.
We are trying to replace the plow with the tractor by tying a rope to the steering wheel putting a brick on the accelerator and letting the tractors drive themselves.
