Technology

Why not AI?

| #AI | #MiSTer | #technology | #videogames |


I hang out on the MiSTer Discord chat. It’s where programmers and users gather to discuss, work on and troubleshoot the MiSTer FPGA system that allows very authentic reproduction of old arcade games and consoles. From zero to N64, the MiSTer provides a near perfect, and often better, experience that software emulators running on a PC cannot match.

But lately an issue has been coming to a head, where people using AI are cranking out cores – the chunks of code that represent the hardware being reproduced – rapid fire. Some users and some devs are upset about this, and others don’t understand why.

I don’t understand why people are so against AI in coding

I’ll tell you why I’m upset. And probably some of these reasons will resonate with others as well, but I don’t presume to speak for them.

Consent

AI companies do not believe in consent. They need vast amounts of input data to create any kind of AI worth using, and they suck the internet dry, over and over, all the time. First they’ll hammer your server under their own banner, ignoring all requests to pace themselves, and causing a massive amount slowdown as their requests run the server ragged trying to keep up. Users suffer.

So as a server admin, you block their banner, and the crawlers – the AI bots that do the sucking up of the data – just change it. They pretend to be something or someone else. So you block the source of the requests, and they bounce their requests off of another system, in order to mask their origins. The servers get slower. Users suffer.

And so more walls are thrown up. Massive services like Cloudflare (Check here to prove you’re a human!) are introduced, or things like Anubis are thrown up, requiring every first request to do some heavy computation before they can view the website being requested. AI bots used to bounce off these, now they just do the math and get through. The difficulty of the computation is cranked up, users suffer – especially on mobile phones, which will get warm as they do heavy math just to view the page. In theory this causes the AI users a little bit of pain as they do so many requests, but they don’t stop coming. They don’t care. They want it, they’ll take it.

This is the AI way, and yeah, people are upset.

Resources

AI data centres use as much electricity as small countries. Each one. And there are many. They need huge infrastructure installations to suck up all the power, and still they run (usually illegal) gas generators, creating harmful emissions and creating walls of noise that anger their neighbors. There’s talk of massive amounts of water use to cool the computers in these centres, and even if that water is not consumed but is returned to the rivers from whence it came, it’s hot on the way out. Rivers heat up. Things die.

Attitudes

AI users are, all too often, complete assholes. “Please don’t use my code for AI” one developer on the MiSTer discord wrote today.

“You can’t stop AI” was the reply. “What did you expect?” “[…]not something you have the right to enforce.” It’s like photoshop, or calculators, they say. It’s like an electric drill.

if you aren’t using it for at least reviews, you’re probably just objectively worse than people who are

If you don’t use AI, you’re bad at what you do. This is often the response to anyone who has a problem with AI. Get with it, get with us, or get left behind. You have no choice but to join the thing you hate, you idiot. You loser.

These same developers will usually be firm believers in open source. MiSTer was built on developers releasing their code under permissive licenses, so other devs can learn from their efforts, double check their code, contribute fixes. And these same devs will get very upset when someone ignores the open source licenses. They’ll get very upset when another developer locks their cores behind a paywall, or doesn’t release the code in a timely manner. If you make consumer electronics and don’t include the source code with your TV, the authors of that code – and many others as well – will get quite loud about it. And, I guess, they’ll use AI to write the letter of condemnation, because it’s easier for them. Some violations don’t matter as much.

Someone pointed out that this part of the rant is a bit wobbly, and that’s true. I have deleted some stuff.

to me AI is a fancy compiler with a 60%-80% success rate

That’s what they’re defending. A machine which is literally a fancy autocomplete word generator that produces output that’s likely to be correct based on all the stuff it’s collected. Ask any expert on any subject and they’ll tell you AI gets their field totally wrong much of the time. Ask any novice and they’ll tell you that AI is right a surprising amount of the time.

And, according to several courts, anything produced by AI cannot be copyrighted. Human hands only. And what happens when you take something written by a human, under an open source license that requires derivatives to be published under the same license, and you publish your own code with AI contributions that may not be similarly licensed? Yeah I’m still figuring this part out, soz.

Humanity

AI has decimated the artistic fields. Every artist I know is struggling. Every writer I know has lost work. Every translator is looking for another job.

Because AI, the machine that mostly gets the number of fingers right, that mostly produces very pretty women that all look vaguely similar, that can be relied on to get things right just often enough to make it seem like you’re saving time while you proofread everything, because that AI stole everything, refuses our requests, refuses our demands, refuses to stay out of our software, our machines, our communities.

That same AI that saved you some time is fucking killing our friends.

I used to love hanging out in the MiSTer Discord because very smart people were doing very difficult work to re-create old things in modern ways. It was beautiful, it was people coming together and making things I appreciated, and the results were accurate, the body of knowledge was improved, and I supported them with my money and any other way I could. I wrote guides to installing tricky things, I wrote documents that made some software easier to understand.

And now, people come in and wave their wands around, and conjure up cores based on old code that the AI machine found, or was pointed to, that is often wrong or based on poorly understood observations. These cores are appearing rapidly, without the effort, the toil, the tears, the community efforts.

Someone once said “Why should I bother to look at something no one could be bothered to create?”

Some of these wand-waved cores are for games that I love, that I have spent hundreds of dollars buying, fixing, preserving, playing in very specific ways. Some of these cores are for games I can’t afford to buy. And I won’t play them now, these magic’d up AI cores are tainted.

I don’t understand why people are so against AI in coding.

My friend, I don’t understand why you’re not angry like we are.

The Rage

Because life wasn’t really super easy in 2026, what with the fucking everything and the fucking everything else. And now I’ve got to scrutinize every piece of art, every written article, every piece of code. I have to evaluate the specific amount of AI that was used to determine whether or not I feel like I can use it without getting the icks.

It’s hard to know who to trust and who to slip a stick of dynamite under, when it comes to AI use.

Good devs doing good work, creating real PCBs to attach probes to real hardware to get real results, say things like “I asked claude to find errors in my design and it found a part wired backwards”

And half of me is like “That’s lazy but valid” and the other half is “light the fuse LIGHT THE FUSE“.

Meanwhile my website is still struggling under the AI onslaught, consent still means nothing, data centres are still poisoning us.

How can any AI use be ethical, fair, or honest? I hate this shit. I hate having to doubt people, I hate that governments are begging companies to build their datacentres in their regions. I hate that big money has nothing else to do but throw cash at AI.

Every AI user is fucking complicit, IMO. Your use may be honourable, but the thing you’re using sure as fuck is not.

--NFG
[ Aug 31 2026 ]
Navigation

Comments

Just some guy

Sep 10 2026

Yeah. In the 90’s I done a high school project on AI as part of my computing course at a level that led to university entry level qualifications. HIGH SCHOOL. I wrote a trivial program in True Basic to play noughts and crosses against the user and then discussed whether it would pass the Turin test. Years later, degree and PhD statistics in Bayesian methods – basically the foundations of some branches of inference and AI. Now, working in a field where everybody who knows anything and is actually good despises AI, but we get hounded by management to use it and give examples of cases where it has improved productivity and been useful. The good people do the same work they always would, then ask the AI to do it, they then correct it (!) as it’s guaranteed to have problems, and then tell management here is some AI use in just enough of a frequency that they don’t get replaced by some new graduate who thinks AI is the future but simply has not been burned and doesn’t know enough yet to understand that even the top AI models today are not even production ready.

You are not wrong on the rest of your points that are far wider than whether the technology is even suitable for professional use. On the plus side, the cost of AI is about to bring down the generic mass use cases as even if you believe LLM’s are the future, the cost of use is only now starting to be put on the end user, and once that hits we will see management baulk at the reality of the situation.

Got something to add?

Your Comment
Name:
Email:
Website:

 

© 2026 NFGworld Subscribe: ::