- AI is poised to change the software development profession
- Opinions are split on whether coding skills will remain integral to the job
- But everyone seems to agree AI will free up developers' time to let them be more creative
There’s no question at this point that artificial intelligence (AI) will change software development and engineering. But what is still up in the air is how exactly things will change.
AWS CEO Matt Garman seems to think AI is poised to take over coding entirely. Garman reportedly told employees that software development in the future won’t necessarily require coding as a skill, but will require developers to think critically about how to innovate.
In two years, he said, “It's possible that most developers are not coding” and instead are focused on “How do I go build something interesting for my end users to use?”
Stuart Munton, Chief for Group Operations & Technology at IT consulting firm AND Digital, had a slightly different take though. He told Fierce that AI is more likely to automate the foundational elements of coding and free up developer’s time to tackle “some of the more gnarly, interesting, creative problem-solving bits.”
“The AI revolution is going to make humans more human. It’s going to mean that we’re going to need to be focused much more on those skills than knowing the syntax of JavaScript, than knowing how a log file is constructed,” Munton said. Those skills, of course, being creativity, innovation and empathy.
Munton’s comments are similar to what Wayfair’s Head of Martech, Data, and Machine Learning Matt Ferrari told Fierce a year ago. That is, it’s not that coding will go away, but that developers will be able to automate repetitive and foundational coding tasks to focus on coding new features and creative innovations.
Indeed, Munton said this evolution is similar to earlier changes in the field. “No one writes binary anymore,” he pointed out. “As a craft, we’ve been abstracted from the ones and zeros. The problem space is where humans are.”
Hidden risks
However, Munton acknowledged there is a certain degree of risk that comes with using AI to build the same foundation for every house, so to speak. First, there’s the potential for apps and services to become more “vanilla” as they all tap into the same underlying code that the AI spits out.
And second, there’s the risk that apps and services end up with the same vulnerabilities – again due to using the same code.
The solution? Ironically, there could be an AI agent trained to detect that, he concluded.