AI Scuffed Programming

### Twitch
Everything is built live on twitch
### Editor

Join this channel to get access to perks:

### Links

#coding #neovim #typescript #programming #vim #softwareengineering #codinglife #webdesign #webdevelopment #webdev #javascript #rustlang #rust #twitch #twitchstreamer #programmerhumor #codinghumor #software #softwareengineer #softwaredeveloper #softwaredevelopment

[brunospasta]

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.

[dandogamer]

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 😂

[hjewkes]

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

[TADevelopment]

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

[schonsense]

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

[jonathanwells-]

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.

[scroll-anarchy]

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.

[boredSoloDev]

"this is shitty code which is also most of the code"


I'm doing my part

[NinjaLobsterStudios]

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

[IAmNotASandwich]

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.

[MagicAndReason]

That's why you have people curating the inputs to a LLM, and not just doing automated input. Also, most LLMs have a feature where you can vote up or down the replies. That gets fed back into improving the quality of the model.

[pikaprisma]

I only use it for comments now, saves me a bunch of time.

It's faster to read the comments and then make changes than to write them all from scratch

[wristcontrl]

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.

[Shonicheck]

Ai is great at applying pre-known algorythms to the problem. Ai code tends to be cleaner than my own if created as sorta small snippet(since it puts great emphasiz on styling and redability a lot of times, like examples in books), and he makes work way faster than it may have been otherwise just by doing that, wven with all the mistakes and needed refactoring.

My point is that AI is great at making arbitrary decision and working as a sort of a code monkey - you can't trust it to write long or structuraly complex code, but it will do with simpler widely aplicable things tuned to the set restrictions. Its also good at guessing things and looking at things at weird angles which in turn creates sort of a rubber ducky effect on steroids, since you have a rubber ducky that can provide unique input.

[benjoe]

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.

[SHADOWGAMER-niiv]

Spent hours digging into the CYBEROPOLIS whitepaper, and wow, I'm blown away. The idea, roadmap – meticulously detailed. I'm usually wary of presales, but after this, I threw a big bag into it. This project's got serious potential!

[thibauldnuyten]

Yeah this isnt just a coding thing, this is universal.

People consistently overestimate their knowledge and skills.

[Veptis]

The same is true for general language model.

Which is why you pretrain on a large corpus to have the model be confident with language in general (or code in this case). And then you fine-tune a model on a well curated subset of highest quality and the specific format you want. The literature has proven this does work for code.

I am writing my thesis on code generation benchmarks, and finetuned models do a whole lot better. They are also contaiminated and overfit a lot - but that's some good chapters to write about.

[paulverse]

The undervalued part of LLMs and other approaches is them not just absorbing all inputs 1:1, but compressing it first on multiple levels, causing it - on large enough datasets - to get the quintessence of the inputs, rather than the exact sequence of words or code. This allows them to come up with actual original outputs and improve on whatever crappy inputs they had.
This is always left out, even though it's not a technical detail but responsible for them outputting not complete garbage.
Tho, yeah, no tech is perfect instantly (or ever).

[bogdanbarbu]

You don't need to train your model against a random sample of code, which is the only situation in which your plot is accurate. You can put only high-quality code bases in its training set.