Buy-now-pay-later platform Klarna went public on the US stock market last week, sending its stock surging to well above its expected range.
The company’s CEO, Sebastian Siemiatkowski, has thrown the entire company’s weight behind artificial intelligence, infamously boasting that the tech was doing the work of “700 full-time agents” last year — only to regret his decision months later, admitting that humans play an important role after all.
But Siemiatkowski’s obsession with AI hasn’t disappeared. The CEO told the podcast Sourcery over the weekend that he’s been vibe coding, the practice of using AI to write code using natural language, to express new ideas to his firm’s engineers.
Siemiatkowski claimed that he went from being a “business person” to an amateur developer thanks to vibe coding, allowing him to come up with a prototype in just 20 minutes.
“Rather than disturbing my poor engineers and product people with what is half good ideas and half bad ideas, now I test it myself,” he told the podcast. “I come say, ‘Look, I’ve actually made this work, this is how it works, what do you think, could we do it this way?'”
Siemiatkowski claimed his newfound hobby is saving his staff significant amounts of time — but whether Klarna engineers would agree with that sentiment remains unclear.
In fact, we can easily imagine their eyes rolling out the back of their heads when confronted with a boss bearing hastily thrown-together code.
As Business Insider points out, Siemiatkowski is far from the only tech executive who vibe codes. In June, Google CEO Sundar Pichai noted that he’d been “messing around” with AI coding assistants to “build a custom webpage with all the sources of information I wanted in one place” — and came away feeling “excited.”
However, the practice has already turned into a major thorn in the sides of those tasked with making sense of the resulting inscrutable code. As 404 Media reported last week, freelancers are joking online about having become “vibe coding cleanup specialists” — and cashing checks from resulting gigs — in an ironic outcome of a technology meant to reduce the workload of developers.
Content delivery platform Fastly found in a recent report that at least 95 percent of the 800 surveyed developers had to spend extra time fixing AI-generated code.
... continue reading