Why Don't Most Organizations Make Better Software?

David Heinemeier Hansson says most organizations do not know what they want or how to make it better. They are not bottlenecked on implementation. They are bottlenecked on ideas, vision, and taste. Implementation is only a small segment of the work. Extra coding capacity does not help unless ideas, vision, and taste already exceed the ability to build. You can make a lot of shitty ideas come true and still have nothing worth shipping.

Microsoft has had tens of thousands of programmers and endless resources, endless programming capacity for decades. Just being able to write a lot of code does not produce great, compelling software. He loves Microsoft some of the time and picks on them because that long record already showed the limit of raw programming capacity.

The capacity has been around for about six months. Six months is not very long in a human life cycle for internalizing what is going on. No other form of progress has moved as fast as AI. The critique asks why AI is not going faster, as if the last three months should already have rewritten the whole world into a utopia of software goodness.