How we cover AI-assisted projects - GBAtemp community check-in

GBAtemp Banner.png

It’s about that time of year again where I check in with the community to see if we can get the front page aligned with what people are interested in knowing and what people are interested in talking about. As the title might have given away, there’s one topic more hot than most at the moment, and that’s AI.

I am a C++ developer by trade, and my place of employ has just started to roll out Claude licences for those interested, so the topic of AI assisted programming is very much on my mind. I’m not just going to go out there and say “AI bad”. I don’t necessarily think it’s a productive conversation to have. The reality is that we’re going to see more projects that make use of it in some form, and some are likely to warrant attention.

Going forwards, I think it’s important that we highlight when projects we do feature make use of AI, and to what extent if such information is available. My hope is that this will strike a reasonable balance of being able to share projects that might look interesting, and giving people who want to avoid AI related projects a simple indicator.

Obviously not every project is going to be so open with their usage, and we have no intention of starting witch hunts by deep diving code and looking for em dashes in the comments. The aim here is to make whatever public information there is accessible for people reading here, and just make things a touch clearer for the people who are looking for that kind of disclaimer.

With that aside though, I am interested to know what the general consensus is among the user base. Are you open to playing ports assisted by AI if it’s the only port available, or happens to offer something unique? Do you draw the line at vibe coding, or do you avoid any project that has Claude listed as a contributor? We sadly don’t have access to polls on the front page any more, but I will make sure to keep up with any replies as they come in.
 
environmental hazards and stealing from humans
AI is probably going to be a major pain—and perhaps even a matter of life and death—for many people, much worse than social media. But helping people make free games or free tools for gaming certainly isn’t one of the things I’d worry about.

PS: There won't be many games that remain completely free of AI involvement, just as you don't really have a choice about living in a world untouched by the influence of social media.

I would like to have a “No AI-generated” video switch on YouTube, though. Also, a “Fact-checked by AI” switch would be nice too. I’d certainly try it to see whether those blatantly obvious fakes could be screened out.

To not waste time or money on bad games check the reviews.
 
Last edited by TomSwitch,
My take is that while GBAtemp should cover every homebrew game and port, regardless of whether it's been written with or without LLM, it would be nice to give more space to human-coded projects. How? Well, the way I would do it is to report all projects produced with LLM support into a single news post (something like "Best games and ports of this week generated with AI-assistance"), whereas continuing to report fully human-generated projects into a well-deserved individual post.

Why? Well, because LLM allows people to produce more with less effort, and what I'd like to praise is not just the end result, but the effort it took to achieve it.
 
My take is that while GBAtemp should cover every homebrew game and port, regardless of whether it's been written with or without LLM, it would be nice to give more space to human-coded projects. How? Well, the way I would do it is to report all projects produced with LLM support into a single news post (something like "Best games and ports of this week generated with AI-assistance"), whereas continuing to report fully human-generated projects into a well-deserved individual post.

Why? Well, because LLM allows people to produce more with less effort, and what I'd like to praise is not just the end result, but the effort it took to achieve it.
Nobody is stopping anyone from having a “Handmade” logo. You’re welcome to make it bold and prominent so everyone is clear about it, rather than asking silly questions about why easy features (at least as far as AI is concerned) are missing.

A well-made homebrew is simply well-made. Whether the author uses AI or not makes absolutely no difference to me. If you care about it that much, just ask whether it is “Handmade”—and remember, they don’t owe you an answer. Assuming it isn’t handmade unless explicitly stated otherwise is perfectly sensible.

Why do you care so much about effort? Do you also need someone to declare that they’re not gifted? Or perhaps say, “Please don’t thank me too much, because this was easy for me”? This isn’t school.

If someone comes here looking for praise, they clearly won't stay for long, because they won't get much of it here.
Post automatically merged:

In a manner of speaking. The engineering side isn't what's controversial.
For instance automated tools like web search engines require in essence rudimentary "intellect" to not run amok, returning useless results.
Treat AI like a new employee and you can’t go too wrong. The key difference is that humans don’t become smarter as quickly as AI does. AI felt like an intern not that long ago; now it feels more like a new hire coming from the competition. Soon, they’ll feel like talent you poach from the competition.

Last year’s AI was dumb—good mainly for cheap labor. This year’s AI learns fast, and after some investment in training, they become surprisingly productive. Next year, you may just find that you can’t live without them.
 
Last edited by TomSwitch,
environmental hazards and stealing from humans
The environmental hazard of slightly heating some water in a heat exchange? Stop regurgitating propaganda. Your apps have been doing the same thing for decades and the issues are already mitigated by controls - because they have to be. It's not the wild west.

And learning isn't stealing. I'm not having another argument about what steeling is. The last one (piracy isn't theft) took a quarter century to settle, and there are still holdouts. Learn more words.

If you have actual concerns, let's talk about beef, almonds, coffee, shampoo, smartphones, golf, gasoline, diamonds... a lot of the stuff you can see has a horror story behind it. AI is one of the least harmful things to enter the world in half a century, and it has the potential to solve the world's biggest problems.
 
The environmental hazard of slightly heating some water in a heat exchange? Stop regurgitating propaganda. Your apps have been doing the same thing for decades and the issues are already mitigated by controls - because they have to be. It's not the wild west.

And learning isn't stealing. I'm not having another argument about what steeling is. The last one (piracy isn't theft) took a quarter century to settle, and there are still holdouts. Learn more words.

If you have actual concerns, let's talk about beef, almonds, coffee, shampoo, smartphones, golf, gasoline, diamonds... a lot of the stuff you can see has a horror story behind it. AI is one of the least harmful things to enter the world in half a century, and it has the potential to solve the world's biggest problems.
Environmental impact is a valid concern. AI taking people's livelihoods is a real concern. Whether we get Skynet in the near future is still unknown, although it feels more plausible than ever. AI depriving app makers of the encouragement they need to hand-code their apps, however, is so far down the ladder of concerns that it's practically like picking up sand to fill the sea.

GBATemp is not a school, never was, and probably never will be. It's a place where we share our toys, projects, and hobbies. Our playthings change with time. AI is the new toy here. Some people have started having fun with it, and there is nothing wrong with that.
 
Last edited by TomSwitch,
Environmental impact is a valid concern. AI taking people's livelihoods is a real concern.

(...) AI is the new toy here. and there is nothing wrong with that.

Do you see how you are contradicting yourself?

Yes, there is something wrong with AI and it's not about how well-made the code was at the end. The problem is the technology itself and how it works. The concercing is ethical.

AI technology hurts people. This is an undebatable fact, if you don't believe it, go watch some news. AI should be labeled AI. Its the bare minimum.
 
Story time

Killing Skynet​

November 2006

In 2047, they called it Skynet.

Not because it thought. It never thought, never woke, never launched a single missile. The name was a joke in the resistance — the kind of joke you make when the truth is too embarrassing to say plainly: humanity had built the machine tyranny from the movies entirely out of a supply chain. One stack. One vendor. One green eye logo on the substrate of every thought the world outsourced.

So when the cell found a way back, nobody called the mission anything fancy. It was the Skynet job.

Mara arrived in a world that did not know what was coming.

She stood in the rain behind a strip mall in Santa Clara, wearing stolen clothes and a stolen name, watching a man in a leather jacket park his own car. No motorcade. No security cordon. No glass hive of servers behind him. Just a mid-sized office building with that same green eye, and a banner that meant nothing to most of the world yet: the graphics card company.

In 2047, that logo ran civilization. Here, it was stapled above a ballroom full of gamers.

GeForce 8800 GTX. The future of graphics.

Not the future of everything. Not yet.

Kai met her at the motel with a paper map and a phone book, because that was what this era ran on. "I found him the way they taught us," he said, tapping the listing. "He's not hiding. Nobody hides when they think they're only selling toys."

Mara studied the timeline on her wrist display, its future-tech screen dimmed to look like a cheap watch. One line pulsed red: CUDA ships this week. The seed of the stack. Kill Skynet in the cradle.

"This is the window," she said. "Not 2026. The cell chose wrong before. You don't cut the branch at its widest — you cut it at the root, when nobody's watching. When the downloads are in the hundreds and the industry laughs."

"And the guardian?"

"There's no guardian in 2006," Mara said. "Nobody in this decade believes he's worth protecting."

She was wrong.



Elias found her on the third day, in the press row of the launch event, three seats away, wearing a lanyard that said BENCHMARK REVIEW. He didn't look at her. He watched the stage, where Huang paced under the lights, talking about a strange new idea: that a graphics chip could compute anything. That the parallel engine in a gamer's rig could run science, medicine, markets.

"You're from the cell," Elias said quietly. "The ones who think 2026 was the wrong window."

Mara's hand stayed very still inside her jacket. "You're early for a guardian."

"I'm not from your future," Elias said. "I'm from the one you're about to make."

He slid a battered handheld across the empty seat. On its cracked screen: a timeline she recognized and then didn't. The Skynet job, completed. Huang, removed in 2006. The stack, leaderless — then bought, fractured, absorbed. A different logo rose on the same architecture. A different name on the same choke point. By 2041, the world ran on one stack again.

"We did the job," Elias said. "I carried the payload myself. Skynet died in the cradle, exactly like you want. And it didn't matter — because the choke point never lived in the man. It lived in the fact that nobody builds a second road while the first one is free. Kill this Skynet, and the world just compiles another one."

Mara looked at the stage. Huang was holding up a graphics card like a religious artifact, and a room full of gamers applauded a technology they would use for everything except what it was actually for. The idea was already out of the building — burned onto master discs in a warehouse off San Tomas Expressway, sitting in the laptops of a hundred engineers, waiting in the minds of every graduate student who would download it next spring.

"Then nothing works," she said. "Not 2006. Not 2026. Not the man, not the moment."

"No," Elias agreed. "One thing works. You don't kill the architect." He nodded toward the back of the ballroom, where a young engineer stood alone, sketching compiler notes on a napkin. "You build other architects. Starting tonight."

The lights came up. The crowd surged toward the demo stations. In the noise and the dark between them, Mara's hand came out of her jacket — empty.

She walked out into the California night, past the banner, past the green eye, past the future that would arrive on schedule no matter what she did. Behind her, through the glass, she could see Elias shake the young engineer's hand.

The Skynet job was over. Nobody had killed Skynet. Nobody ever could.

The first road was already paved.

So she would start the second.

END

Prompted by Tom, generated by AI
 
Story time

Killing Skynet​

November 2006

In 2047, they called it Skynet.

Not because it thought. It never thought, never woke, never launched a single missile. The name was a joke in the resistance — the kind of joke you make when the truth is too embarrassing to say plainly: humanity had built the machine tyranny from the movies entirely out of a supply chain. One stack. One vendor. One green eye logo on the substrate of every thought the world outsourced.

So when the cell found a way back, nobody called the mission anything fancy. It was the Skynet job.

Mara arrived in a world that did not know what was coming.

She stood in the rain behind a strip mall in Santa Clara, wearing stolen clothes and a stolen name, watching a man in a leather jacket park his own car. No motorcade. No security cordon. No glass hive of servers behind him. Just a mid-sized office building with that same green eye, and a banner that meant nothing to most of the world yet: the graphics card company.

In 2047, that logo ran civilization. Here, it was stapled above a ballroom full of gamers.

GeForce 8800 GTX. The future of graphics.

Not the future of everything. Not yet.

Kai met her at the motel with a paper map and a phone book, because that was what this era ran on. "I found him the way they taught us," he said, tapping the listing. "He's not hiding. Nobody hides when they think they're only selling toys."

Mara studied the timeline on her wrist display, its future-tech screen dimmed to look like a cheap watch. One line pulsed red: CUDA ships this week. The seed of the stack. Kill Skynet in the cradle.

"This is the window," she said. "Not 2026. The cell chose wrong before. You don't cut the branch at its widest — you cut it at the root, when nobody's watching. When the downloads are in the hundreds and the industry laughs."

"And the guardian?"

"There's no guardian in 2006," Mara said. "Nobody in this decade believes he's worth protecting."

She was wrong.



Elias found her on the third day, in the press row of the launch event, three seats away, wearing a lanyard that said BENCHMARK REVIEW. He didn't look at her. He watched the stage, where Huang paced under the lights, talking about a strange new idea: that a graphics chip could compute anything. That the parallel engine in a gamer's rig could run science, medicine, markets.

"You're from the cell," Elias said quietly. "The ones who think 2026 was the wrong window."

Mara's hand stayed very still inside her jacket. "You're early for a guardian."

"I'm not from your future," Elias said. "I'm from the one you're about to make."

He slid a battered handheld across the empty seat. On its cracked screen: a timeline she recognized and then didn't. The Skynet job, completed. Huang, removed in 2006. The stack, leaderless — then bought, fractured, absorbed. A different logo rose on the same architecture. A different name on the same choke point. By 2041, the world ran on one stack again.

"We did the job," Elias said. "I carried the payload myself. Skynet died in the cradle, exactly like you want. And it didn't matter — because the choke point never lived in the man. It lived in the fact that nobody builds a second road while the first one is free. Kill this Skynet, and the world just compiles another one."

Mara looked at the stage. Huang was holding up a graphics card like a religious artifact, and a room full of gamers applauded a technology they would use for everything except what it was actually for. The idea was already out of the building — burned onto master discs in a warehouse off San Tomas Expressway, sitting in the laptops of a hundred engineers, waiting in the minds of every graduate student who would download it next spring.

"Then nothing works," she said. "Not 2006. Not 2026. Not the man, not the moment."

"No," Elias agreed. "One thing works. You don't kill the architect." He nodded toward the back of the ballroom, where a young engineer stood alone, sketching compiler notes on a napkin. "You build other architects. Starting tonight."

The lights came up. The crowd surged toward the demo stations. In the noise and the dark between them, Mara's hand came out of her jacket — empty.

She walked out into the California night, past the banner, past the green eye, past the future that would arrive on schedule no matter what she did. Behind her, through the glass, she could see Elias shake the young engineer's hand.

The Skynet job was over. Nobody had killed Skynet. Nobody ever could.

The first road was already paved.

So she would start the second.

END

Prompted by Tom, generated by AI
I need to tell you, maybe others have told you already but still needs to be told.
Your long boring AI generated texts are... well, exactly that, boring, TL;DR stuff.

I would guess people are not even reading past the first paragraph, so you're wasting time and you're not communicating whatever you're trying to say.

If you want to communicate something, try some other away, people then may read and get what you're trying to say.

If you are NOT trying to communicate something, then stop wasting your and everybody else's time with these long winded collection of words, if there is something in there it was lost.... better just share the prompt, I believe that may be shorter and we will understand better what you tried to say reading the prompt instead of reading that random book from the library of babel you posted.

TL;DR: Just say straight what you want to say instead of posting a boring AI generated novel instead of getting straight to the point.
 
  • Like
Reactions: mudrik
I need to tell you, maybe others have told you already but still needs to be told.
Your long boring AI generated texts are... well, exactly that, boring, TL;DR stuff.

I would guess people are not even reading past the first paragraph, so you're wasting time and you're not communicating whatever you're trying to say.

If you want to communicate something, try some other away, people then may read and get what you're trying to say.

If you are NOT trying to communicate something, then stop wasting your and everybody else's time with these long winded collection of words, if there is something in there it was lost.... better just share the prompt, I believe that may be shorter and we will understand better what you tried to say reading the prompt instead of reading that random book from the library of babel you posted.

TL;DR: Just say straight what you want to say instead of posting a boring AI generated novel instead of getting straight to the point.
What is your problem? Is this some “your PS5 is better than my Switch” type of thing? I like my Switch. That doesn’t stop you from licking your PS5 or whatever you do with it.
 
What is your problem? You PS5 is better than my Switch type of problem?
I mean exactly what I said, if you tried to say something it was lost because the text was boring to read.
If you want to say something say it simple and in little words, or make it fun I don't know.
 
You are just a troll.
Look, what I am telling you is honest, I am not trying to troll you.

I find the text boring to read because it feels like random filler put together to make an idea that could have been put into two or three sentences take the size of an essay; but it doesn't say more than the two or three sentences would have said, it is all filler, so it is very boring.

I think what you wanted to say in those two or three sentences maybe was important (I don't have a clue because trying to read the filler text was a drag so I didn't do it). If you wanted to communicate (I mean, to reach somebody else with your ideas) it would have been better to just post the two or three sentences instead of the fluff-filler laden text.

I will quote this, in topic, from "The Reverse Centaur's Guide To Life After AI":

The Reverse Centaur's Guide To Life After AI - Cory Doctorow said:
Let me provide an analogy. A colleague of mine is a law professor. Prior to the proliferation of chatbots, law students were loath to ask their professors for letters of reference, because these were understood to be a choresome labor. You’d ask your prof to go to all this trouble only if you were damned sure that they thought very highly of you.

But as chatbots infiltrated higher ed, everyone—profs and students—understood that the way a letter of reference was produced in today’s world was for a prof to feed three bullet points to a chatbot along with a prompt to expand them into a law school reference letter. The chatbot would dutifully inflate these three bullet points with five or six paragraphs of florid nonsense, which the prof could paste onto letterhead and sign.

This has led to an explosion of reference letters, and a concomitant cheapening of their value. Today, when my friend’s department has an opening for a postdoc, they are inundated with applicants’ reference letters, in such profusion that no one could possibly read them.

You may have already guessed at how they solve this: that’s right, they feed these letters back to a chatbot and ask it to reduce them to three bullet points.

Obviously, this is very perverse. For one thing, the three bullet points the receiver’s chatbot generates are only going to be passingly related to the three bullet points the sender fed to their chatbot. This is a very lossy game of broken telephone.

Your text feels exactly like the three-bullet-points inflated reference letter.

This is the last reply I will give you because otherwise again the thread goes off topic and it will be my fault.
 
Last edited by sarkwalvein,
Do you see how you are contradicting yourself?

Yes, there is something wrong with AI and it's not about how well-made the code was at the end. The problem is the technology itself and how it works. The concercing is ethical.

AI technology hurts people. This is an undebatable fact, if you don't believe it, go watch some news. AI should be labeled AI. Its the bare minimum.
No contradictions. AI causing harm is no different from a knife being capable of killing. The technology is here to stay, whether you like it or not. AI is unfair competition to people who write code for a living is beyond doubt. Industrial revolution harm a lot of people. AI probably will harm even more. Industrial revolution raised living standard to the majority of people. Whether that will happen with AI I don't know. Globalization used to be a good thing, now the thinking is mixed. There is no need to stand on a side about whether AI is good or bad, it's more complex than that. What good does saying AI bad do?
 

Site & Scene News

Popular threads in this forum