AI fan wikis riddled with errors are muddying the waters of games culture

I don't think the boundary should live at human-written or AI-written, it should just be factual.

This weekend I was playing Stalker 2 and got frustrated that the trader prices in the game seem nonsensical. I spent a lot of time trying to research the mechanic but most guides and some wikis are outright wrong. They either buy into the games marketing or just assume mechanics work as expected.

So what I did was I let Claude read through my game files. 5 minutes later I ended up with a document that very concretely explains all the factors that influence trader prices with references to the game scripts to back it up.

My point is that having AI write game wikis is fine, as long as the information is verifiably correct.

We can put these together to conclude that every game and community that lacks that kind of dedication will die out and be replaced by psychologically personalized short-form dopamine-agonist videos, as exemplified in an infamous comic:

https://x.com/Merryweatherey/status/1395152040389713922

I think its working so far.

Wikipedia has always been different from other sites for leaning more on human labour instead of automated systems (so-called "soft security" https://meatballwiki.org/wiki/SoftSecurity ).

It needs to be verified. That's all. That could be by a human, but it could be by AI. The point is that any content that's written by AI might be wrong, so if you're not checking it you can't trust it. AI is pretty good at checking it's own work.

I imagine with a fan wiki (or company docs in my experience) no one bothers to put the effort in. That's probably because the rewards are for writing the content, and there's absolutely no reward for checking it's correct.

I'm afraid your dystopian future where the zone is flooded with shit and nobody can trust anything seems eminently plausible, partly because the internet was built without identity.

I think to solve this you'd have to have strong proofs of identity locked to biometric identifiers. We'll probably need that anyway for other reasons.

It's not a trivial problem. Websites like this could require webauthn tied to a hardware identifier of their choice (like a fingerprint or faceid) - I'm not sure if webauthn provides that level of detail to providers (which method was used), and how secure that attestation is. Of course that's not going to be popular but it might be the lesser of two evils.

Case in point: the proliferation of furcons.

The unspoken assumptions that underpinned the web included the idea that it was made by humans and that content generation couldn’t be mechanized.

There are already YouTube ads for services that Humanize AI output.

The best analogy is that we are at the start of the Dark Forest era of the net.

I assume people will adapt, based on the things that don’t change. If you cannot make out authenticity of participants, then you index on correctness of argument. If the output is useful to you, then you got what you needed and can move on. A more transactional and less community oriented web.

There is no solution (though there will be plenty of grifters trying to sell one).

The open internet we currently know will be lost to mountains of slop just like usenet and email were even prior to the AI era.

Maybe there are other people selling the tools, who try to convince the former people that what they're doing is smart and profit will roll in any day now.

Anyway it would be a mistake to think something is profitable just because someone bets on it.

Don’t get me started on Google’s propensity to show a full page of unwanted shopping results now without any way to show websites unless you turn verbatim on. “Brand picks” what garbage.