The Daily · The New York Times

Can A.I. Already Do Your Job?

February 18, 2026·31 min·3 clips
Watch AI build Natalie's personal website with an Eagles-themed design in just 96 seconds.
The Daily host Natalie Kitroeff opens by asking New York Times tech podcast Hard Fork host Kevin Roos to explain vibe coding, a term she had encountered but did not understand. Roos describes vibe coding as a practice coined by Andrej Karpathy, a former programmer at OpenAI who is one of the most respected figures in the AI research community, in which a user describes what they want in plain language and AI tools like Claude, GPT-4, and others write all the actual code. The vibe coder does not need to understand the code being written or even be able to read it: they describe intent, the AI produces code, they test whether it does what they wanted, and they describe further changes. Roos argues this is fundamentally different from previous AI coding assistance, which helped programmers write code faster, because vibe coding eliminates the programming knowledge requirement entirely. He describes building functional apps and tools in minutes without writing a single line of code himself, and discusses the broader implications for who can create software. The episode covers both what vibe coding enables, rapid prototyping, hobby projects, and small business tools that previously required hiring a developer, and what its limitations are, including security vulnerabilities in AI-generated code, debugging difficulty when things go wrong, and the risk of building on code you do not understand and cannot maintain. Roos and Kitroeff discuss whether vibe coding will displace professional programmers or primarily expand the universe of people who can build things. Roos is cautious about strong displacement predictions, arguing that professional engineering judgment remains valuable for complex systems. The tone is accessible and curious, aimed at an audience that may be technically naive but is interested in understanding what is actually happening in AI. The audience is anyone curious about how AI is changing programming and what it means for creative and technical work.
Listen to the show on