The impact of AI on software engineering jobs

“AI will reduce the demand for software engineers.”

That’s what another founder told me over breakfast this weekend.

He’s dead wrong. And in this post, I’ll explain why AI tools like ChatGPT and GitHub Copilot will actually *increase* the demand for developers.

Let’s start off where my friend and I agree: We both believe there we’re at an inflection point with AI and that developers who use these tools will become massively more productive.

My friend argued that this will allow companies to achieve the same amount of work with fewer engineers and that ultimately companies would build smaller tech teams.

This seems reasonable, but the argument is flawed. It implicitly assumes that the amount of work is fixed when that’s not actually the case.

Many problems could be solved with tech that companies haven’t been able to justify building solutions for yet because it’s been too expensive. If building tech were cheaper, then companies would be willing to build more.

Now, with developers getting an AI productivity boost, the cost of their output decreases. As a result, the number of problems worth solving with tech increases. Developers will be able to generate more value for companies, and companies will hire more of them.

And yes, developer salaries will increase. At least for the ones that learn how to use these tools to their full potential.

At this point, you might be thinking that this is all theory mambo jumbo. But there are several historical instances where technological inventions resulted in this phenomenon.

One of the most memorable (and shocking) instances of this was an invention by Eli Whitney, who was born in America in 1765. When he saw how much work went into processing cotton, he embarked on a project to build a device that would reduce the manual labour required. His hope was that this device would reduce the need for manual labour to the point where it would end slavery. Unfortunately, Eli did not understand economics very well.

His invention, the “cotton gin,” rapidly separated cotton fibers from their seeds. The machine worked very well, and farmers across America adopted it. It marked a turning point in the slave trade but, much to his horror, in the opposite direction of what he had hoped.

By reducing the labour required, he made it much cheaper to grow cotton. Cotton farming become incredibly profitable and farmers expanded their operations to meet European demand. Larger farms required more labour, and the slave trade boomed for another 70 years.

A depressing story to be sure. But it's an important part of American history and illustrates the economic phenomenon quite clearly.

Fortunately, that's as far as the comparison goes. Unlike the cotton gin, I firmly believe that AI will have a profoundly positive impact on the world and enable developers to make life better for billions of people.