I’m not a fan of vibe coding, as it’s inherent in human beings to do things more lazily and then not do proper verification of the code in regards to function, bugs and security. And the biggest threat is coders who aren’t skilled enough using vibe coding to put out software they otherwise couldn’t, who then lack the skills to vet the AI generated code. I’ve played with some Bitcoin Android wallets this week that were just buggy, couldn’t connect to and use my Electrum server backend properly, didn’t show some coinjoin transactions even though it had the proper balance… And the worst offenders seemed to be the ones putting out new features quickly, making me suspect vibe coding and lack of vetting, which has me leaning on more tried and tested apps that are more stable and not trying to add so many new features. And for a fact one project was using vibe coding as it was mentioned in the comments of my bug report on their Github. Consequently, I’m thinking we’ll eventually see some Linux distributions and projects who will reject AI to get back to having better and more stable code, as even running opensource, we’re going to have to be more concerned with how well vetted the code is we’re running on our computers as well as who actually coded it.