Developers Are Spending More Time Fixing AI-Generated Code

Senior software developers have been spending more time fixing AI-generated code as trends like “vibe coding” rise. Seasoned programmers have described the modern task as “worse than babysitting.”

According to a recent TechCrunch report, developers are increasingly dealing with AI-generated code that, while produced quickly—especially during vibe coding sessions—often contains bugs and errors. In some cases, fixing this output takes longer than correcting code written by junior developers.

A recent survey conducted by Fastly, which included nearly 800 participants, confirmed that senior professionals are spending significant time fixing and editing AI output, addressing issues such as security risks, hallucinations, and missing information.

TechCrunch also noted that the problem has grown so widespread that it has even given rise to a new role in the industry: “vibe code cleanup specialist.”

“Using a coding co-pilot is kind of like giving a coffee pot to a smart six-year-old and saying, ‘Please take this into the dining room and pour coffee for the family,’” said Carla Rover, a senior web developer who has been using AI for developing software for her startup. She explained that while AI is capable of generating code, the results are rarely clean or correct, calling the task of fixing AI output “worse than babysitting.”

Another developer interviewed by TechCrunch, Feridoon Malekzadeh, agreed that generative AI often behaves like a child, describing it as “hiring your stubborn, insolent teenager to help you do something.”

Malekzadeh said he spends 30% to 40% of his time fixing AI-written code. “You have to ask them 15 times to do something,” he told TechCrunch. “In the end, they do some of what you asked, some stuff you didn’t ask for, and they break a bunch of things along the way.”

While professionals criticize AI-generated code for inaccuracies, hallucinations, and errors, cybersecurity experts warn of broader consequences. A few days ago, researchers reported that a security flaw in one of the most popular AI code editors among developers, Cursor, allowed hackers to execute malicious code.