The Unbearable Lightness of Vibe Coding

Aditya Agarwal has written an account of how it feels for a successful, highly skilled coder and entrepreneur to live through AI’s rapid expansion. His essay in The Information , “ A Claude Catharsis ,” describes a weekend spent coding with Anthropic’s AI assistant after 20 years of building software, line by line, at Facebook and Dropbox. Agarwal says that with automated coding he produced more code in five days than in the previous five years, and that the work was better and more ambitious than what he could have undertaken alone. This was a nearly existential moment with his feelings gyrating between a sense of lost professional identity to what he describes as a kind of “wild, almost reckless energy” anticipating opportunities he could not previously have imagined.  Agarwal’s coming to terms represents one possible resolution to similar crises bearing down on many in the knowledge sector. Some take a resistance stance. Workers whose careers were built on credentials and deep experience in technical matters have the most to lose from a system that no longer takes credentials at face value and seeks to automate their expertise. Degrees from elite institutions, while potentially meaningful, say less than they once did about readiness. He recounts the experience of a company that ran twenty week-long work trials for engineering hires and found zero correlation between years of experience and adaptability to AI tools. What credentials don’t capture, Agarwal says, is an intangible he calls restlessness. He describes another firm that is giving coding interviews intentionally too long to finish by hand. The company found that people using AI tools daily, rather than just reading about them, produced ten times the code volume of those who weren’t. Coding skills count, but what counts more is building solutions to under-specified questions with open-ended outputs. At the moment, there’s no credential for that. Agarwal’s framing is hopeful and daunting at the same time: adaptability is the new currency, and unlike a Stanford degree, it’s available to everyone. In her 2016 book Grit , University of Pennsylvania scholar Angela Duckworth highlighted the importance of passion and perseverance toward long-term goals as predictors of success. Agarwal’s experience is grit in a different form. A coder who abandons hand-coding yet still remains committed to building useful software has changed tactics not goals. The adoption of AI coding tools is akin to recommitment to a life project, with new energy and new vistas if the coder doesn’t turn away from them. The disposition Agarwal describes involves a tolerance for mess and uncertainty. A saying often attributed to Edison applies here: Opportunity is missed by most people because it is dressed in overalls and looks like work. Restless tinkering eats time, produces dead ends, broken prototypes, and false starts. Tolerating mess conflicts with our innate caution against risk, disorder, and unpredictability. But our willingness to accept risk is what needs encouragement; our survival instinct is usually strong enough to take care of itself. The hardest part of this transition, Agarwal writes, has always been letting go of the thing you were in order to become the thing you might be. That is a human problem rather than a technological one, and it is one AI is making impossible to ignore. The post The Unbearable Lightness of Vibe Coding appeared first on American Enterprise Institute - AEI .