"Earlier this year, former OpenAI exec Andrej Karpathy coined a new term - "vibe coding" - for using artificial intelligence to rapidly develop software using natural language prompts. But the approach comes with some glaring shortcomings that have gradually come to light, from major cybersecurity problems leading to mass leaking of sensitive personal information to rampant hallucinations that turn vibe-coded projects into a buggy mess that has to be painstakingly fixed by human programmers."
""It's basically entirely hand-written," Karpathy wrote in a followup. "I tried to use Claude/ C odex agents a few times but they just didn't work well enough at all and net unhelpful, possibly the repo is too far off the data distribution." In other words, even the godfather of vibe coding doesn't trust the tech enough to use it on his own project."
Andrej Karpathy coined "vibe coding" to describe using LLMs with natural-language prompts to rapidly build software prototypes. The approach enables fast, throwaway projects but produces significant shortcomings, including cybersecurity flaws that have leaked sensitive data and frequent hallucinations that create buggy, unreliable code requiring manual fixes. Karpathy later released Nanochat, a minimal ChatGPT-like interface, and reported building it mostly by hand after attempts to use Claude and Codex agents failed. Karpathy acknowledged that LLMs sometimes cannot fix bugs and that vibe coding is best suited for quick experiments, while overreliance risks costly failures as organizations cut human development resources.
Read at Futurism
Unable to calculate read time
Collection
[
|
...
]