How does software development change with AI tools?

AI software development tools are fundamentally changing the roles played by software developers in several ways.

Software developers are going to be writing less-- or, often, no-- code moving forward. The tools, whether in an IDE or a terminal window, handle writing standard code better and faster than any developer.

In this mode, the developer's role is to translate application requirements for the AI agents and make sure that the AI does it's job. Effectively, the AI is a team made of junior developers and the software developer becomes a combined architect and manager of the team.

The best way to ensure working software is to have the AI agent self correct by using test cases that it can run itself to detect and diagnose problems.

The other role for software developers is to do novel development.

AI excels at writing code that it has seen 100 examples of. Pull information from a database, format it, display it in an editable fashion, grab changes, push them back into a database? AI tools have enough examples of how to do that from existing applications that they can implement a best-in-class approach.

But asking an AI agent to do something it hasn't seen before isn't guaranteed to produce good output. If the software needs a new algorithm implemented, or a better way of optimizing performance, or a different approach? An experienced, creative software developer will need to write the code. At least, the first time?

AI agents can still be involved. They can work in the IDE and provide assistance-- they can make sure that the correct syntax is used. They can write test cases. They can comment and augment the code. But until they've seen the newer-faster-better approach, they can't write it-- because their strength is in sorting through code that has been written in a similar fashion multiple times and putting out the best version.