AI agents are like “genies,” said Kent Beck, one of the authors of the seminal “Agile Manifesto” — they’ll often grant your wish, but not always in the way you’d like them to.
“They will not do what you mean. They have their own agenda,” Beck said on a recent episode of The Pragmatic Engineer podcast. “And the best analogy I could find is a genie. It grants you wishes, and then you wish for something, and then you get it, but it’s not what you actually wanted.”
After 50 years in programming, including a stint at Facebook and a hand in the foundational text about agile software development, Beck said he’s now having the most fun of his entire career — and it’s partly thanks to AI agents, even despite their unpredictability.
“I’m trying all of the tools,” he said. “Because right now, nobody knows what process is going to work best. Nobody knows anything. We should all be trying all the things that we can imagine, and then the truth will emerge out of all of that. So that’s what I’m doing.”
Beck was one of the 17 authors of the 2001 paper the “Agile Manifesto,” which outlines four values and 12 principles for faster software development. The paper inspired the “agile method” of software development, which became commonplace in Silicon Valley because of its efficiency-boosting techniques.
Agile Manifesto
Today, AI is already being used to replace human labor, particularly in programming, where it’s hitting early-career engineers the hardest while simultaneously speeding up software development. Given the current state of coding, Beck believes that fundamental organizational skills are increasingly important rather than particular technical specializations.
“So, having a vision, being able to set milestones towards that vision, keeping track of a design to maintain the levels or control the levels of complexity as you go forward,” Beck said. “Those are hugely leveraged skills now, compared to, ‘I know where to put the ampersands, and the stars, and the brackets in Rust.'”
Though Beck does incorporate AI into his process, he doesn’t necessarily trust it, he said. The technology isn’t consistent enough to be relied upon.
“Sometimes it even seems like the agent kind of has it in for you,” he said. “‘If you’re going to make me do all this work, I’m just going to delete all your tests and pretend I’m finished, ha, ha, ha.'”
Results are so inconsistent, he added, that using AI to code can sometimes feel like gambling.
“It feels like a rat and the pellet,” he said. “It’s like there’s just a run button and I have to click it every time. And I click it and it is a dopamine rush because this is exactly like a slot machine. You’ve got intermittent reinforcement, you’ve got negative outcomes and positive outcomes.”
Once in a while, though, the output will be just right — and Beck will be tempted to spin the wheel all over again.
“The distribution is fairly random, seemingly. So it’s literally an addictive loop to have it. You say, ‘Go do this thing.’ And then sometimes it’s just magic.”
