Dominus wrote on Yesterday, 17:58:

I didn’t mean to offend you, sorry if I did. There are just certain mannerisms that made me think that. Thank you for clearing this up 😉

Not at all, I just found it interesting you were prompted to put it that way.

For sure: no computer system is ever going to replace Yours Truly [so long as my deliverables can keep outperforming theirs]!

I think we're all in general agreement that the human element is superior, but the point I was prompted to make is AI is not the big threat its detractors portray it as because the responsibility to assert our inherent human worth belongs to us alone and the process of that sliding into decline began well before AI came along to potentially catalyze it. I mentioned I quit social media nearly a decade ago, yet in a way I have found that puts me more in-touch with what goes on in the world around me as well as clearer-thinking on how to respond, and not just because subsequent studied observations have vindicated the notion that there is something sinister about there being screens literally everywhere with inherently-addictive apps that have all but replaced direct human connections with technology-filtered ones. I've never had a problem with technology making us more productive (or more fun) and I never will, but count me among those who draw the line at being expected to assimilate myself into what is essentially a cybernetic collective.

Ars Technica: AI game clones are making indie developers afraid to show their work

Delayed Steam Next Fest demo is a sign of a growing problem for indie devs.

McPixel developer Sos Sosowski has been documenting numerous examples of fellow indie devs complaining of AI-generated clones making wholesale copies of their efforts. Sosowski says the situation puts indie developers in a hopeless bind when trying to promote their work: “Don’t share your game—no sales. Share your game—gets stolen,” he wrote. “How do you even survive as an indie in the current landscape? This is insane.”

Still on the original subject, here's a relevant Kenney post with relevant (and predictable) reactions.

enaiel wrote on Yesterday, 22:30:

Ars Technica: AI game clones are making indie developers afraid to show their work

Delayed Steam Next Fest demo is a sign of a growing problem for indie devs.

McPixel developer Sos Sosowski has been documenting numerous examples of fellow indie devs complaining of AI-generated clones making wholesale copies of their efforts. Sosowski says the situation puts indie developers in a hopeless bind when trying to promote their work: “Don’t share your game—no sales. Share your game—gets stolen,” he wrote. “How do you even survive as an indie in the current landscape? This is insane.”

Looks like it has started. How exactly are people supposed to pitch a game for crowdfunding when the idea can be co-opted completely by AI in a couple weeks? (assuming effort is put in)

gerwin wrote on Yesterday, 15:45:

megatog615 wrote on 2026-10-14, 02:24:

Trashbytes wrote on 2026-10-14, 00:51:

AI Slop is a different debate and since that slop required human input to even exist ..well ...garbage in garbage out and there is a huge swarth of human created Slop out there and we are happy to give that a free pass. (Just visit your local Art gallery to see the human created slop or Steam for the human created slop games there too, or just check out Netflix which is peddling slop like its going extinct)

It's not a different debate and I refuse to treat it as such. All GenAI output is slop. No amount of "prompt engineering" will turn such output into gold. The only media or software or anything else worth consuming is made without plagiarism and GenAI is purely plagiarism.

Sort of. Stupid thing is that reverse-engineering is part of my hobby, and a big thing in relation to things vogons. It was like that long before AI existed. So in the past me and others just managed with datasheets and many hours of contemplating, guessing. and trying. Now with AI one can give it a binary, maybe do some ghidra beforehand, then ask: How does this ancient driver interface to certain old hardware, at which code offset, which registers and what does it set by default. It can answer that in a minute. Like having X-ray vision at your disposal. You can test that answer, and usually it appears to be the right answer.

Don't get me wrong, I am not gonna be a cheerleader for AI. In some other fields (non-software) it is more like a propaganda machine. I still use a Nokia dumb-phone and such. But it is tempting to get some things done that seemingly stalled. I got like 8 or so stalled projects to the finish line in a month. AI assistance is always disclosed there, right at the top of any readme, I don't need AI to appear smarter. (I also curb any attempt of AI trying to "behave social", yuk). In general, the whole new situation is bizarre. The software world is in for a rocky ride, and that is just that world.

I think I may be convinced of its usefulness to experts in programming if the datasets they used for training the LLMs were all from public domain code and reading material. But even then, having a machine gobble up such works still feels a bit gross. It might take decades for LLMs, if they were to go down this route of respecting copyright and human creativity, to shed the stigma related to slop code.

We're never putting this genie back in the bottle, so a) we remove the techbros pushing for AI datacenters(every single one of these tech billionaires is guilty of crimes worthy of life imprisonment and if we can't have that then we tax billionaires until they don't exist anymore) and remove such datacenters, b) enforce copyright law against tech companies scraping human content.

Unfortunately I'm way too jaded to think any of that will even come close to happening.

tcaud wrote on Today, 03:19:

How exactly are people supposed to pitch a game for crowdfunding when the idea can be co-opted completely by AI in a couple weeks? (assuming effort is put in)

They're not, i dont where this ends, but i suspect as everything moves online, including core game logic, any one who creates games to run locally will be at a continuous disadvantage if they mean for their game to provide income. Even online code could be inferred by training ai on gameplay and getting it to "do something like that"

I can even imagine games being genai streamed in part, maybe some gameplay templates and mechanics that get mixed up along with generated assets. never play the same game twice... doesnt sound so good 🙁

megatog615 wrote on Today, 10:35:

I think I may be convinced of its usefulness to experts in programming if the datasets they used for training the LLMs were all from public domain code and reading material. But even then, having a machine gobble up such works still feels a bit gross. It might take decades for LLMs, if they were to go down this route of respecting copyright and human creativity, to shed the stigma related to slop code.

We're never putting this genie back in the bottle, so a) we remove the techbros pushing for AI datacenters(every single one of these tech billionaires is guilty of crimes worthy of life imprisonment and if we can't have that then we tax billionaires until they don't exist anymore) and remove such datacenters, b) enforce copyright law against tech companies scraping human content.

usefulness is separate from provenance. I'm not sure how copyright can be enforced, one of those permission to access versus permission to reproduce issues, AI doesn't reproduce as such. But there is an argument for "you lot learned from all of our data and benefit from it, so pay more tax", rather than solve for each specific copyright its a woolly attempt to address is at large (brings its own issues to)

Dominus wrote on Yesterday, 15:29:

I noticed that your posts read as if they are AI generated not written by a human 🤷‍♂️

BaronSFel001 wrote on Yesterday, 17:51:

I have written the way I have for at least the past 15 years (right about the age my personality type got set). This is not the first time I have had some aspect of my humanity questioned; it has occurred to me that I may be on the spectrum, not that I see a point to getting such a thing checked at my age. But since I know I am not lying when I declare that I personally refuse to use AI I take this as a compliment to my orientation to detail, desire to keep focus on all pertinent aspects of a matter, and understanding that we all can still get by with our own skills absent the aid of computers [which, I will keep saying, are ultimately too DUMB to replace human beings].

The funny thing about these kinds of remarks is; you have to realize how these LLM's are "trained" in the first place. Especially local models employed by one off users: They see someone on the internet with a very intelligent manner of wording their posts, etc... so they decide to continuously train a model after this user. They'd rather have a machine clone someone's mannerisms and grasp of language than try to hold such a grasp themselves over language.

It's a pretty disturbing practice honestly, I've seen friends of mine getting turned into "AI chatbots" behind their backs even. Apparently "ethics" goes beyond everyone's minds this current decade on the internet.

DracoNihil wrote on Today, 13:25:

The funny thing about these kinds of remarks is; you have to realize how these LLM's are "trained" in the first place. Especially local models employed by one off users: They see someone on the internet with a very intelligent manner of wording their posts, etc... so they decide to continuously train a model after this user. They'd rather have a machine clone someone's mannerisms and grasp of language than try to hold such a grasp themselves over language.

Oh, let me tell you: several months ago I was trying to research something regarding Gravis Ultrasound, and saw the Co-pilot results (I use Bing as my search engine of choice because it gives free Xbox points) quoted off one of my own posts here on VOGONS where I was speculating rather than stating known fact.

Just more evidence on top of the pile of how dumb computer systems are.

By the way, I never found a definitive answer to the question: does GUS MIDI support some form of capital-tone fallback like AWE32 does?