What’s the Future for Software Developers?

Paige Bailey grew up in a small town in rural Texas and started coding on a cast-off Apple computer at about age 9. Today, she’s among the people perched at technology’s cutting edge, as GenAI developer experience lead at Google.

“I have a whole bunch of nieces and nephews who I love desperately, and who are amazing, many of them very, very technical,” she told Chris Pirillo, host of this episode of The New Stack Makers. “Historically, I’d always been kind of like nudging them: when you go to college, you should get a computer science degree.”

But now that generative AI has upended the tech world, she said, “I don’t know necessarily that that’s what we should be telling kids to do.”

Make no mistake: she still thinks understanding how to write software is a useful skill. But learning how to think about problems and ask the right questions might be better preparation for a tech career in the future, Bailey suggested: “It’s becoming much, much more a scenario where this magic is kind of happening in the margins between computer science and some other disciplines.”

And that’s changing the job of developer, in ways that will likely have an impact on current and future devs. Historically, she said, programming “was something that a person had to dedicate multiple years of their life in order to do effectively. Now it’s becoming the more of a tool that everybody should have in their toolbox because generative AI is just making it easier and easier to write code or to automate processes.”

Turning Ideas Into Code

Bailey, a veteran of roles at Microsoft and GitHub as well as Google, worked on the PaLM2 large language model team before landing in her current role.

At Google, developers have been applying AI to the software creation process for some time. She can see how it’s changing the role of developers more generally, based on her own experience.

“Over the course of the last year and a half or so, I found myself, whenever I write code, to be more of this kind of reviewer, the kind of overseer, as opposed to the person that’s trying to do the work myself, or having to hunt through documentation to find the right answers.”

She sees the future of GenAI as not so much replacing developers as facilitating and accelerating their creativity. “There’s always been this struggle for creative folks to have an idea in your brain and then to have this really, really arduous process of doing lots and lots of work and running into issues and brick walls trying to get that idea out into the world such that other people can see it.”

Bailey compares what she believes will be the future of GenAI-assisted software development to the best parts of the open source community — the experience of collaborating with other people to build things.

“Right now, the number of actual developers — like people who are engineers, spending their time building software in the world — is a very, very small percentage of the overall global population,” she said. “And that’s really unfortunate, because, you know, being able to build things share brings a lot of joy.

Rather than being the province of a relatively small number of developers “locked in an ivory tower,” as she put it, “these models have really started democratizing the capability of building things like clouds, artifacts, building software like with V0.dev from Vercel. Being able to automate processes. They’ve really democratized that in a way we’ve never seen before.”

A Boon for Tackling the Backlog

But Bailey doesn’t believe that democratization of software development will result in fewer jobs for professional devs. Instead, “It only feels like there’s more and more potential for new applications, for new work, because of generative AI, because so many new potential, potential projects are unlocked.”

For instance, “There are always way too many things on the backlog than there are days to get to them … if I think of all the bug and [pull request] issues that we have for some of the features that we’d like to add or some of the documentation that we’d like to include, so many of them have been categorized as P2s or P3s, which means we’ll never be able to get to them. Whereas now we have a running shot at getting to them.

She summarized, “I think the mission-critical piece is just making sure that everyone feels like they know how to apply generative AI to their work, to sort of accelerate it and to spot the places where generative AI might not be sufficient just yet and they need to oversee it.

Check out the full episode to follow Bailey and Pirillo’s conversation, including a glimpse of what Google is doing on the GenAI front.