The bug AI missed — never assume it figured it out

Transcript preview

Okay, so it finally opened up. So here's a very interesting thing. I think it kind of assumed that every question is going to end with question mark. So this is very important when you're building and white coding things. Don't assume that AI has figured everything out. As I have specified, if you have been watching me on Instagram, that the main role now when AI is able to do the code part is for you to test all the edge cases. Over here, it was trying to do that through the replica site, it created a replica page and was trying to do that. And I specifically mentioned that, hey, you got to go there and check. And when it did that, it was able to find its own flaws that the questions aren't ending in question mark. And it also had this issue of not able to figure out. So you see, it's able to figure out what's exactly happening. When I actually worked on this extension, I had to take the whole code and I had to research about multiple things. I still remember there is fine by ID, fine by script, fine by this and that, class name, something like that. So I used to understand the code and try to figure out the consistency of how things are happening five years ago. And today, I'm just sitting here watching at Claude terminal and I'm like, just do it already, man. It's been two minutes. It's a total different time, guys. And I hope you're able to understand how amazing time we are living in. I really hope I had some video of me building that time, you could have been able to understand. But I do have, I'll attach that I don't want. But the point is now the real test is running, okay, so it's going to rerun that I usually go there and reload for it to be able to show me that I reloaded the extension for it to…