I am a developer, and from my discussions with other experienced developers, I'm not sure they are in denial about AI for code generation. Many are actually excited about how much more productive they can be when using this technology effectively. We have been discussing our approach to evaluating candidates and conducting interviews with the consideration of this technology.

One challenge we face is determining how to evaluate a candidate if they are using AI for code generation during the interview. Our conclusion so far is not to prohibit it, but to change the way we test. Most of the senior engineers participating in interviews agree that the future involves working together with this technology, as it saves time.

I, for one, eagerly await further advancements in this technology. As it stands today, developers often perform tasks that could be automated and probably should not have been done, as it's like reinventing the bicycle. For instance, across the world, login forms are written multiple times a day by different developers, despite their similarities.

However, it's true that the current technology doesn't yet enable people without development experience to create functional software. I recently encouraged my wife, a teacher, to try programming games for her students, but it didn't work as expected.

I anticipate changes in the way we write code, such as test-driven development combined with AI I experiment with, and tools like GitHub that can fix broken pull requests. In the hands of capable engineers, this technology has the potential to increase software developers' productivity by an order of magnitude. Will developers lose jobs as a result? That remains unclear. My company would love to have ten times more developers but can't afford it. What if we could do ten times more with the developers we have? The supply of code is likely to increase significantly, but it has never met the demand up to this point. The question is, how much demand is there for more software?