Jerry's Marginalia — You Don't Have to Agree With My Tools
There's something I've been thinking about lately.
I've been exploring the indie web, looking at Neocities sites, collecting buttons, browsing personal websites, and slowly putting together a little corner of the internet called Germlin Subs.
It's been fun. It's also been a reminder of what made the old internet interesting in the first place: people making things because they wanted to.
Then I came across something that gave me pause.
On someone's website to-do list, I found a note about moving their artwork away from Neocities because its owner had expressed support for generative AI.
The note ended with a little parenthetical:
(which is BAD)
And that got me thinking.
You're allowed to disagree.
Let's get something out of the way.
Artists have legitimate concerns about generative AI.
Copyright, consent, training data, compensation, transparency, imitation, and the potential impact on creative work are all serious subjects.
These aren't imaginary concerns. In December 2023, the Federal Trade Commission published a staff report summarizing concerns raised by working creative professionals. They discussed the collection of creative work without consent, lack of disclosure about training data, competition with AI-generated work, and imitation of artists' styles and likenesses.
A 2024 academic survey of 459 artists also explored the tension between the potential usefulness of generative AI and concerns about ownership, transparency, compensation, and the future of creative work.
People have every right to question these systems, criticize companies, refuse to use AI tools, and decide they don't want their work associated with them.
If someone wants to move their artwork to another platform, that's their decision.
Nobody owes an AI company their support.
But there's another side to this.
Someone else's decision to use AI doesn't automatically make them a bad person. Someone experimenting with these tools isn't automatically attacking artists.
You can disagree with somebody without demanding that everyone else agree with you.
The platform isn't the people.
On January 21, 2025, Neocities published its official position on AI.
Its stated approach was that users would not be forced to use AI, that Neocities would provide ways to opt out of known AI crawlers through robots.txt, and that it would not ban websites simply for using AI.
The owner also clarified that low-quality generated content, like other low-quality content, could still be considered spam.
That doesn't mean every user has to like the decision.
It doesn't mean artists should ignore their concerns.
And it certainly doesn't mean people aren't allowed to leave.
But I think there's an important distinction between saying:
"I don't feel comfortable hosting my work here anymore."
And saying:
"This is BAD, and everyone should understand it that way."
The first is a personal boundary.
The second starts becoming a judgment about everyone else's choices.
Those aren't quite the same thing.
I'm experimenting. I'm not pretending.
I've been interested in AI-assisted translation and game preservation for a while now.
I've followed projects like CODEC BABEL. I've been looking at recompilation projects. I've been thinking about games that never received English releases and the possibility of making them accessible to people who otherwise might never experience them.
I'm also building Germlin Subs.
I don't have a team of professional translators waiting in an office. I don't have a publishing company behind me. I'm one person trying to learn, experiment, research, and build something.
And yes, I intend to use AI tools.
That doesn't mean I think AI is perfect.
It doesn't mean I intend to blindly trust its output.
It doesn't mean I won't check translations, test patches, document mistakes, or acknowledge the people whose work helped make a project possible.
Quite the opposite.
If an AI-assisted translation makes a forgotten game playable, that's a starting point. It isn't automatically the finish line.
Test it. Question it. Improve it. Be honest about how it was made.
If someone later comes along and improves the translation, wonderful. That's preservation too.
We've survived bad fan translations before. We'll survive these too.
The legal questions aren't settled by an opinion.
There's another reason I don't want to reduce this entire discussion to a simple argument about whether AI is good or bad.
The legal questions are complicated, and they're still being examined.
The U.S. Copyright Office's Artificial Intelligence initiative examines issues including copyrightability, digital replicas, and the use of copyrighted material in AI training.
Its May 2025 Part 3 report on generative AI training examines the copyright questions surrounding the use of protected works to train generative AI systems.
That doesn't magically settle every dispute. It does show that there are real questions worth examining, rather than dismissing.
I can acknowledge those concerns without concluding that every person who uses AI is doing something inherently wrong.
And I can support experimentation without pretending there are no problems to solve.
Both things can be true.
The right to choose goes both ways.
I don't expect everyone to agree with me.
Some people will never want to use generative AI. That's fine.
Some people will experiment with it, find it useful, and incorporate it into their work. That's fine too.
Some people will leave platforms because they disagree with their owners. They can do that.
But I don't think disagreement should become a requirement that everyone else abandon their own choices.
You don't have to use my tools.
You don't have to like my methods.
You don't have to visit my website.
But I don't need your permission to experiment, either.
And perhaps that's the part of this whole discussion that gets lost.
The internet is big enough for people to make different choices. It should be possible to disagree without turning every difference into a declaration of war.
Anyway, back to the raccoon.
I'm still building Germlin Subs.
There are games worth documenting, translations worth exploring, old websites worth visiting, and forgotten corners of gaming history worth preserving.
I don't know exactly where all of this will lead.
Maybe I'll make something useful. Maybe I'll make mistakes. Maybe someone will come along and help improve what I've started.
That's okay.
I'd rather try, learn, and leave something behind than spend all my time arguing about whether I'm allowed to pick up the tools.
The tools are in our hands.
What matters is what we do with them.
And if you don't like my toolbox?
That's okay.
Just don't expect me to put it down because you don't like what's inside.
Further Reading & Sources
These are the references behind the discussion. They're included so readers can investigate the issues themselves.
Neocities — Official AI Position (January 21, 2025)
Read the original statement on BlueskyFederal Trade Commission — AI and Creative Fields (December 2023)
Staff report and discussion summaryLovato et al. — Foregrounding Artist Opinions (2024)
Academic study — 459 artists surveyedU.S. Copyright Office — Copyright and Artificial Intelligence
Official AI initiative and reportsU.S. Copyright Office — Artificial Intelligence Study
Reports, including Part 3: Generative AI Training
— Jerry "The Ankle-Biter" Silverhand 🦝
Germlin Subs: Translating the Untranslatable Since Whenever the Hell We Feel Like It.