Generative AI’s ability to write software code has quickly created one of the technology’s first real use cases for business.
Professional software engineers and novices alike are using AI coding assistants to produce, test, edit, and debug code, reducing the amount of time it takes to complete the often tedious steps required to finish projects. And Big Tech is fully on board: AI now writes as much as 30% of Microsoft’s code and more than a quarter of Google’s, according to the heads of those companies, while Mark Zuckerberg aspires to have most of Meta’s code written by AI agents in the near future.
Meanwhile, powerful new AI tools like GitHub Copilot, Cursor, Lovable, and Replit have given even people with little to no knowledge of coding the ability to knock up impressive-looking apps, games, websites, and other digital projects using little more than a series of prompts detailing what they want to build.
Some practitioners are even allowing the software to take the lead when it comes to writing code and accepting some or all of its suggestions, a method known as “vibe coding.” But there’s still no substitute for good old human know-how—because AI hallucinates nonsense, there’s no guarantee that its suggestions will be helpful or secure. Researchers at MIT CSAIL highlight how even AI-generated code that looks plausible may not always do what it’s designed to. AI tools also struggle with large, complex code bases—though companies such as Cosine and Poolside are working on that.
We’re also beginning to see the early effects on other parts of the industry—including fewer entry-level jobs for younger workers. So while coding assistants may help you in your existing job, they won’t necessarily help you land a new one.
Correction note: This original version of this story referred to Microsoft Copilot. It has been updated to use the correct name for the product we’re referring to, which is GitHub Copilot. We apologize for the error.
