I Vibecoded Morrowind, Duke Nukem, and Serious Sam Onto an iPhone

It seems like you're bending over backward here to alienate the author from his work
On the contrary. I'm pointing out the real author here, which is the AI generating the code (or the people whose code was used as training data for the AI), not the man telling it to do so.

You know, like when you order your dinner in a restaurant, you'll not magically become the "author" or creator of the food you're eating, just because you told the server what you want to eat. Instead that title will still go to the chef who cooked it or to the person who invented the recipe.

The rest of your comment in similar nonsense, the result of not understanding even just the basics involved here.
 
You know, like when you order your dinner in a restaurant, you'll not magically become the "author" or creator of the food you're eating, just because you told the server what you want to eat. Instead that title will still go to the chef who cooked it or to the person who invented the recipe.
But you can give the chef the credit - even if the sous-chef did the actual cooking? Hmmm… sounds like you’ve lost your own argument :)
The rest of your comment in similar nonsense, the result of not understanding even just the basics involved here.
I think you’re grasping at straws now… just give up…
 
But you can give the chef the credit - even if the sous-chef did the actual cooking?
So, you didn't even get the parallel, and that the question is whether the customer ordering the dinner is to be credited with having created the food. Why am I not suprised?

I think you’re grasping at straws now… just give up…
So, you are now _begging me_ to "give up", because at this point even you know that you've lost this debate and have no way to recover? How cute.
 
The "did the human really do anything" framing in this thread is missing the more useful question, in my opinion: not who typed the code, but who's on the hook when it breaks.


A weekend port of an already-open-source engine and a piece of software real people depend on daily are different regimes entirely. The AI can generate working code for both. What it can't do is decide whether an edge case matters enough to fix before shipping, or take responsibility when it doesn't. That's still squarely a human job, AI-assisted or not, and it's the part of this whole debate that tends to get skipped in favor of arguing about authorship credit.
 
So, you didn't even get the parallel, and that the question is whether the customer ordering the dinner is to be credited with having created the food. Why am I not suprised?
You didn’t understand that you basically restated my point I posted a few posts back - and are simply beating a dead horse with your WRONG idea.

Credit/blame goes to whoever was in charge of any activity. If I use a tool to do something, I am still on the hook for the blame. And I get credit if it succeeds.

Hence the pizza - if I liked it, I credit the restaurant - not the individual who happened to prepare it. And if I get salmonella, I also blame the restaurant - not the individual who prepared it…

No one goes to Pizza Hut and blames “Franklin the cook” if it sucked…
So, you are now _begging me_ to "give up", because at this point even you know that you've lost this debate and have no way to recover? How cute.
I’m not begging - I’m advising you to do so as you’re clearly wrong and just grasping at straws to salvage your already battered reputation…
 
Yes, that's exactly what I told you. You didn't understand. Still don't. Your entire comment is a statement on that.


Of course you are. I even quoted you on that.
So now you can’t even argue - just slander… I know it sucks to be proven wrong on a forum - but it’s just words… move on…
 
The "author" had really played no measurable role in this. ... I'm pointing out the real author here, which is the AI generating the code (or the people whose code was used as training data for the AI), not the man telling it to do so.
Your first claim about the author playing no measurable role is clearly false, by even a cursory reading of the article. The LLM could not on its own create a finished product; it required multiple iterations and inputs from the author. That is distinctly measurable.

Nobody here is arguing that the code wasn't generated. We're taking issue with you making the leap to denying all credit to the mod creator simply because they did not hand-write the code.

By your logic, a ship captain can't take credit for a successful voyage because he only gave direction to the crew who carried out his orders. But I think we all understand that the captain's value is not him physically turning the wheel, but rather deciding where to go and how to get there, and how to deal with the problems that arise on the voyage.
I ask you directly then: should the captain of a ship be given credit for the success or failure of a voyage, and why?
 
Your first claim about the author playing no measurable role is clearly false, by even a cursory reading of the article. The LLM could not on its own create a finished product; it required multiple iterations and inputs from the author. That is distinctly measurable.
Right. Like when you order a birthday cake, you tell the bakery what kind of flavor or text you want on the cake. Still, it's not you who will be considered the creator or maker of the cake, but the bakery.

Nobody here is arguing that the code wasn't generated. We're taking issue with you making the leap to denying all credit to the mod creator simply because they did not hand-write the code.
For one, that's your straw man, not what I actually said. Also, by definition, someone who hasn't written the code of a program can't have "coded" it or can't be the coder or the author of the code.

I ask you directly then: should the captain of a ship be given credit for the success or failure of a voyage, and why?
The captain will surely not be credited as the author of the "voyage". Despite his contributions to it being far larger than the prompter's to the end result in this case.
 
What jumps out to me here, is how well vibecoding matches this specific use case: one person, one machine, no users but yourself, and if it breaks tomorrow the cost is just your own evening. That's a completely different risk profile from shipping something other people depend on, where the AI-written code you don't fully understand yet is the code you'll be debugging at 2am when someone else's data is on the line. Both are real, they're just not the same activity wearing the same name.
 
For one, that's your straw man, not what I actually said. Also, by definition, someone who hasn't written the code of a program can't have "coded" it or can't be the coder or the author of the code.
You started off technically correct, saying that the author didn't write the code for their project. Something everyone has acknowledged as true and obvious.

But then you shifted into how to assign broader credit, your mask slipped, and you erred by stating "The "author" had really played no measurable role in this." That is you explicitly denying them such credit.

Right. Like when you order a birthday cake, you tell the bakery what kind of flavor or text you want on the cake. Still, it's not you who will be considered the creator or maker of the cake, but the bakery.
"The creator or maker of the cake" is a reductive assumption which implies that only one party can be a creator.

Creative and high-level work is work, and such work is necessary for any project to be worth doing in the first place. If your friend is having a birthday, and you get them a special cake that reflects your relationship and their preferences, you're a creator. Maybe you provide the baker specifications like design, message, colors, flavor, size, shape... and then offer direction when the baker is unclear how to proceed. In that case, you have objectively and measurably contributed to the work, because the bakers could not have created it otherwise.

It's the exact same thing for coding with AI. You don't get credit for physically writing the code. But you do get credit for identifying the right problem to solve, deciding how to solve it, responding appropriately to problems as they arise, and for the quality (or lack thereof) of the finished product. We are trying to get you to understand this nuance and you seem determined to reject.

The captain will surely not be credited as the author of the "voyage". Despite his contributions to it being far larger than the prompter's to the end result in this case.
You answered with a remarkably confident No, but did not explain why. And the why was a crucial element of my question. So I repeat it. Why would you surely not give a ship's captain credit for the success or failure of his voyage?
 
Repeating the same nonsense several times won't change it being a nonsense. You already said these things. It was already explained why you are wrong. You still are.
Maybe you should practice what you preach? You have said the same thing over and over- and many of us have proved you wrong… yet… you continue to show how wrong you are again and again 😜
 
I'm confident FF222 didn't read the article at all. Just read the headline and first paragraph and made his conclusion.
 
AIs (LLms) are only good at doing things _that have been already done several times_. It's because they can only do things that they have been taught to, by previous instances of the same thing. They can't achieve something completely new, because that's not how they work.
You made a vast over-simplification. Do you use AI for anything meaningful (work, research, etc)?

Have you ever heard of inference? AI references across disciplines and is able to reach conclusions and theories on topics it's NOT been specifically trained in. It's one of the amazing features! Also, I can train it (like a student) to help me meet my goals!

Is it always right? Nope, usually not even on the first few tries for the complex scientific puzzles I'm working on. However, with the right prompting you can make it correct and better than a human or machine could alone. The process is truly intriguing and powerful - but it takes experience, skill and knowledge to yield.
 
Thanks for confirming what I wrote - despite you obviously not realizing that you did just that.

Also "inference" - in relation to LLMs - doesn't mean what you think it does, and doesn't work the way you think it does.
 
Last edited:
Back