I for One, Welcome Our New Digital Rubber Duck Overlord
ChatGPT is our new Rubber Duck, and it’s here to stay!
Lately, I’ve been using ChatGPT to write code, however, not in the way you’d expect.
I’m not crazy enough to blindly trust the generated code. In fact, as you can read from this other article, the code is usually buggy, so you have to help it find the problems with its own creation.
It’s true that it’s definitely one of the most advanced conversational models we’ve been able to play with for a while. But it’s also true that it’s far from perfect, people are highlighting their problems constantly, so why spend time with a faulty AI?
To paraphrase a popular saying:
It’s not the size of your AI, it’s how you use it that matters.
Or something along those lines. Let me explain.
The good ol’rubber duck
The rubber duck is that iconic figure you use to debug your code by explaining what your problem is to it and then figuring out how to solve it.
It works because you’re externalizing your problems and by doing so, you hear them from a third-person perspective.
