There’s been a lot of talk recently about “vibe coding”, which is the practice of using large language models to write software, describing the problem to the AI which then generates the code; humans hone and test it afterwards.
Here at Hackaday, we wondered whether this is a good or bad thing, so we asked our readers what they thought.
Responses varied, but many comments highlighted that using AI for coding still requires a solid grasp of the intricacies of programming and articulating the problem in the first place, so it’s not a shortcut for less experienced coders.
Some commenters also highlighted that the process of using AI is more about speeding up the busywork of coding, with the large language models proving useful for taking care of the more tedious aspects of programming.
It could be dangerous to use AI-produced code that nobody quite understands though, and could contain bugs that cause disasters.
Overall, Hackaday concluded that, while AI coders shouldn’t be treated as senior developers just yet, they are certainly better than the worst summer interns.
