Comment by applfanboysbgon
2 days ago
> I’m not meaning to say “it works on my machine” by this, but simply that I couldn’t have caught this before.
If you knew what you were doing you could have reasoned about what your code did, though. Such is the danger of trial-by-error vibe coding. I personally closed the tab because the audio kept cutting off arbitrarily after some minutes. Not to criticise you unduly because this was a cute concept, but I hope you might take away that LLMs are not only not perfect, but are also only "good enough" if your standards for "good enough" are particularly low, and use this as motivation to learn more in the future instead of always relying on them.
I think no interaction is good for an offline pond, but I do think some tiny little bit of interaction would be nice online since it adds a touch of humanity, being able to see other people and interact with them in some small way.
I think you are being too harsh. Even before AI, software and games would often work great on one computer and poor on other computers. I believe that isn't because they understood the programming less or their standards were low, but because there is so much variance between computer hardware, the firmware inside, and the software running on it that you can't guarantee software to run identically across many combinations.
Similarly it's not unusual for one website to behave differently across browsers.
Sure, to some extent. What I really took issue with is saying "I couldn't have caught this", which is a sort of learned helplessness. It's very often the case that, independent of inscrutable platform differences, LLMs just produce extremely shoddy code that obviously does not scale, and there is a whole class of problems you can solve before they become problems if you actually know how to program.
My intent was not to be harsh, though. I was trying to make a point that one can learn and do better. Coding is very much not solved.
I agree that couldn’t is a tad too far, you could always have tested more. But even with more testing you could’ve missed this, completely irrespective of the use of AI.