Comments

No it didn't. AI consistently writes junk code that takes longer to fix than if you wrote by yourself.

One of our interns used AI to vibe code a micro service. They were here a month and set us back 3

What did the microservice do?

When was the last time you used it and what model was it?

I feel like that’s exactly the point of the title - you can generate a ton of code, but if you care at all about the quality or the overall architecture, it’s made your job harder. It makes the easy stuff easier, and the harder stuff harder, exactly the reason I hate ORMs typically.

Incidentally, I say the title rather than the article, because I’m not going to waste my precious remaining life knowingly consuming AI output. That’s a hella long article that was probably generated off a few bullet points, and if the “author” can’t be bothered to actually write it, then I’ve got better things to do than read it.

It’s ironic to me because I’ve gotten so angry in the past at people who shallowly reacted and commented based solely on a title without reading the article, but these articles are usually so bland and devoid of meaningful insights that you can glean most of the idea from just the title or headline.

Boring time-series data processing stuff, can't really get too much into it.

Nothing worked, not the API Calls, not the database calls, the docs didn't even have real UUIDs for the "testing instructions". But it was just broken enough to return values. And nobody bothered checking if the values were valid or not.

Real gut wrenching realizing we had thousands of lines of lies instead of code.

Well, I think that was the developer not the AI.

I've been skeptical of ai, we're getting it pushed on us at work constantly, but with the latest iterations, I've seen what I feel is an inflection point in terms of producing good results and understanding more context than I ever provided it with and than I ever would have expected.

With access to the expensive models, you get something that produces good reliable results for the most part. I don't know how I'll continue a career in this field at this point.