AI Scuffed Programming

For small code snippets and small tasks its great. Use it all the time for it. The longer and more complex the problem and code gets the worse it gets (obviously). But it drops of very quickly.

Whenever I use AI generated code and it looks like it should work. Then it doesnt and I have to debug it, I always wish I would have just written it myself from scratch 😂

Thats why its so good at writing comments and unit tests. Because the shitty programmers never write them

you just forgot to prompt it for best quality, high quality, masterpiece, top-tier, elite-level, error-free, premium code.

"this is shitty code which is also most of the code" I'm doing my part

I primarily use GPT as a ‘super Google’ I.e., for learning new languages/libraries. And for taking my thought process and quickly querying library documentation in search of an answer. The basic code examples are just a bonus to me

Manual programming: *writes bug; debugs bug; doesn't write bug again* Ai programming: *writes bug; debugs bug; writes bug again*

The difference between those who understand the mathematical underpinnings of models like those used by Copilot, and software engineers, is that the former didn't need to use it for a second to arrive at the same conclusion.

That's right. Copilot is like an aggressive 8 year old for coding. For small tasks, brilliant. Incredible. For larger more complex tasks? Let it take the wheel but just know you'll be adjusting about 75% of the initial suggestion.

it gets even more glorious once you realize that since AI code is not really flagged in any meaningful way it will find its way into the training data. the longer we use AI in coding the more AI will reinforce its own coding habits. And the more lopsided towards crappy code we get.

From my point of view, as an amateur learning to program, it's way better to have an AI assistant that has the knowledge of maybe a 40th percentile programmer than having no assistance at all.

As a person who has worked with one of the biggest ai companies there is another reason why its shit because we are unable to train it on code where context is missing, suppose you have a big project with multiple files and you just ask about a small piece of code which has the error and the model gives you an incorrect answer or hallucinates and you click thumbs down, now we have no way of knowing whole code structure or all the files so that we can debug and fix the mistake that the model was making , so in these cases we usually skipped these thumbs down chats and the model never gets trained on it so as long as this is the case the model will always fail to answer complex questions pertaining to project where multiple files are involved and the issue can be rooted into any of the files

THANK YOU! Everyone around me keeps saying that they use gpt for coding and I don't understand how because I spent more time correcting its shitty code than it would've taken me to write that code.

I find myself talking to ChatGPT / copilot like a stressed partner in a long relationship.😂 I’m like “ I don’t understand what’s gotten into you but I don’t want to keep going around in circles anymore “ 😂

I think that’s why the way supermaven gives suggestions is better. Supermaven only provides small one liners that are usually tedious to write instead of providing an entire ginormous piece of code that you probably don’t want. It also results in better performance and less shitty code cause by pure laziness.

"Let me hit you with some facts". Proceeds to postulate

I love these AI tools for extremely simple stuff, and starting from scratch. It makes getting the first draft of a prototype very easy. But once I get into the refinement steps, the AIs start to drop off in usefulness. Not just them being error-prone, but even knowing what i want. If i break my work down into a lot of small functions the AI is usually pretty good at figuring out what it should be doing, but at that point there is no productivity gain

Finally someone says it. The longer I work I realized that most programmers are pretty bad at what theyre doing. However, it is still mostly "good enough" to get the job done, which is why most of them never needed to actively put in any effort to actually get good.

Yeah this isnt just a coding thing, this is universal. People consistently overestimate their knowledge and skills.

I just script as a hobby in Godot and I've attempted to use AI to help and I've found that for "generating code" AI is horrible. However it has REALLY helped once I started using it as a pseudo-code generator, it spits out some code when I am at a writers block that I can look at in an abstract way to help me brainstorm actually helpful ideas.