The myth of the developer bottleneck
AI Code Generation is missing the point
One of the main areas that Generative AI is “excelling” in, apparently, is helping Developer Productivity. It can generate code at a rapid speed. This is seen to be great, after all that’s where the bottleneck is right?
Wrong.
Development is almost never the bottleneck.
It’s often the place where the problem manifests - but it’s not the cause. Mostly this is caused by terrible information being passed from upstream and terrible process downstream. This leads to massive delays, rework and manifests in the “development” stage.
After working nearly 30 years in the industry the biggest thing that causes the most problems is lack of communication, lack of vision, lack of process and Generative AI isn’t solving any of that. These are all very much Human things. Unfortunately many people still don’t want to admit that Software Product Development is a Human affair.
What we will see is more code being churned out, less time to verify it and people moving quickly in the wrong direction, breaking things quicker and sinking deeper into the quagmire of technical debt and rework.
Typing (code generation) has never been the issue.
As Ackoff says:
If we have a system of improvement that’s directed at improving the parts taken separately you can be absolutely sure that the performance of the whole will not be improved.
Adopting Generative AI and only doing that isn’t going to solve the problems people think. It’s likely to make the real problems more evident and much bigger.