• 0 Posts
  • 108 Comments
Joined 3 years ago
cake
Cake day: August 10th, 2023

help-circle
  • I literally never said I was vegan or vegetarian. I literally said I’d eat any carnivore lol. Why are you freaking out.

    Deer and cows are opportunistic carnivore but it’s not a major source of nutrients like it is for other animals and they generally don’t choose to unless there’s and issue.

    https://deerassociation.com/what-do-deer-eat/

    It’s not like a academic resource but it’s not like deer are out here hunting down bunnies bro.

    How is it vegan / Vegetarian to go “I’m only going to eat predator species”

    Fish, birds, rodents etc do eat insects and other animals as major parts of their diets.




  • Slop is slop, AI slop < human slop.

    All your doing is advertising how much you don’t respect your work or the work of others.

    Code is the app, code is the functionality. If the code is crap then so is the functionality.

    You’ve made a tool with a loose handle. And axe with a crack in they head. It’s gonna fly off and hurt someone.

    Just because you are a crappy programmer doesn’t mean everyone else is.

    There is nothing wrong with open source. Your opened source code just sucks.

    Idk what to tell you besides to take a long critical look at yourself and your work.


  • If you can’t write manually and have it not be slop then you can’t program with a LLM effectively.

    It doesn’t matter how much instruction you give a LLM it fundamentally cannot evaluate itself. Because to ensure that it’s evaluating correctly either you need to evaluate it or someone else. These are not deterministic machines and will “lie” to reach their goals. And I put that in quotes because it’s not really lying that’s to much personification.

    These things are not good for literally anything beyond minor transformation or boilerplate.

    Trust me. If you actually spend the time learning to code well written software by hand you will save time and get a better result. LLMs based coding is a anti-pattern.

    You’re not getting push back because “programmers a upset their jobs are getting stolen” you’re getting push back because your falling for LLM company propaganda. LLMs just are not there yet.

    If more then 20% of your code is written by a LLM your using it wrong.



  • I’ve read some other comments and wanted to add.

    You cannot use a LLM to verify its own work

    They have no ability to think. Any intelligence they have is extremely limited. There a mostly automatic copy and paste machines. They pull code from their training data and online and attempt to compose the.

    Using a LLM to verify its own work is like asking a criminal to run their own trial.

    That’s just now any of this works. I think you should take a step back from the LLM and really start evaluating your work more critically. There is more to software then “it works!”


  • look I’m all for using LLMs for tedious or straight forward transformations of easily verifiable logic. The issue is they LLMs are sycophantic by nature and we are seeing a lot of newly freed “geniuses” who have promised “no no no. You see! I know the secret to using them for good!”

    It’s like the one ring. If you start using it for doing anything beyond reformating, anything that requires critical thinking, you’ve already trapped yourself.

    You’ll feel like your work is quality when it isn’t.

    Personally I still think the quality of LLM code is crap for pretty much anything. Much better done by a well seasoned developer, which is harder to come by then people think. A LLM can help in some narrow cases but not many.


  • I use it for tedious transformations or needle ones haystack problems.

    They are better at searching for themes or concepts then they are at actually doing any “thinking tasks”. My rule is that if requires a lot of critical thinking then the LLM can do it.

    It’s definitely not all they say it is. I think LLMs will fundamentally always have these problems.

    I’ve actually had a much better time using it for in line completion as if recent. It’s much better when the scope of the problem it needs to “solve” ( the code it needs to find and compose to complete your line ) is like the Goldilocks zone. And if the answer it gives is bad I just keep typing.

    I really hate the way LLM vibe coded slop is written and architecture. To me is clear these things have extremely limited conception. I’ve compared it to essentially ripping out the human language center, giving it a keyboard and asking it to program for you. It’s just no really what it’s good at.











  • Gunna be honest, if you really feel that way you’d fire the vibe coder. I don’t give a shit if it’s broken ifs due to the negligence of management and other employees.

    Poor planning on your part does not constitute an emergency on mine. And ofc this varies based on the service as all things do. But it’s rarely important enough that I’d stay over. Like if lives where at risk or it’s a “bring the company down but” maybe I’d stay. But there would be hell to pay for whoever broke it later. Either that or the company would loose me as an employee