everyone knows "but if we ban LLMs the AI bros will just lie about them, so we should let them through" is a bad faith bad excuse and the person saying it knows that every time, right?

[gavin57@toot.wales]

@davidgerard I think the Americans tried that with firearms. The only way to stop a bad guy with an LLM is a good guy with an LLM, etc.

[davidgerard@circumstances.run]

@agowa338 and yet i see it from people who are closely resembling advocates

as i say above, I'd expect modeling future behaviour as if it were bad faith would give accurate predictions.

[agowa338@chaos.social]

@davidgerard

Oh, ok, I see.

[bms48@mastodon.social]

@davidgerard I'd say the argument allows for the possibility of "Black Propaganda" in the original sense of that term. But the argument itself starts from a false premise, when it is now a matter of public record that the "AI bros" are actually lying and prompting models to lie (Claude source code leak, "undercover.ts"), as well as "Copyright law for me, but not for thee" (subsequent DMCA takedowns of leak, vs IP land grab for training); it's been established.

[kitkat_blue@mastodon.social]

@davidgerard

Yes, and for another reason-- because *even if you cannot tell*, banning LLM crap and the people who promote it sends a valuable message that we -distrust- and are personally -repelled- by those who would trade their (and by extension, our own) humanity for convenience and kicks.

Banning LLM ai from any space is an ETHICAL statement.

[petko@social.petko.me]

@davidgerard how about we tax the everliving fuck out of them instead? To a point where LLMs will fund the universal basic income they will necessitate...

Facelss corps: NO NOT LIKE THAT

[gbargoud@masto.nyc]

@davidgerard

"If there isn't a foolproof 100% effective way to enforce a rule it should not exist" is a wild standard to have

[davidgerard@circumstances.run]

@catch56 @agowa338 or you could not make the exception and tell them to just fuck off, that works too

[cap_ybarra@beige.party]

@davidgerard there are no good faith arguments to be had here anymore; only the lottery-ticket holders and people with any ethics or sense

[lykso@tiny.tilde.website]

@davidgerard In the same vein as "if we shame the 'ephebophiles,' they'll just engage in even WORSE behavior, so we should accept and normalize them."

[lumi@snug.moe]

@davidgerard it's such a dumb argument. free software already relies on things that are even harder to prove or enforce

what if someone wrote their code on company time, meaning they don't own it? that's way harder to enforce than a genai ban

[miss_rodent@girlcock.club]

@davidgerard Yeah, and tbh... if they're the sort to lie about it, they are a problem. Just like someone lying and saying they 'wrote' something they plagiarized some other way would be a problem.

[miss_rodent@girlcock.club]

@davidgerard "You can't stop people from copying and pasting code from elsewhere"
Yeah, not always, but we *do* deal with it - when we find out they're doing it without proper permissions, credit given, licensing, etc., their "contributions" are generally rejected or removed, and usually so is the person that lied about it.

[lonsfor@woof.tech]

@davidgerard
What's next, ban murder?

[fayedrake@furry.engineer]

@davidgerard

The friction is important. As Amazon will tell you even a fraction of a second can make a significant difference to people’s decisions.

Also I’ve seen it posted here that the ethical stance is important, and yes, making it clear that it’s seen as socially disadvantageous is also important to putting the barest control on the cost asymmetry issue.

[toerror@mastodon.gamedev.place]

@miss_rodent @davidgerard It's literally impossible to prevent wire fraud, and frankly - why bother? It's one of the fastest growing stimulus to the global economy. Countries should cease onto this growth oppourtunity and open scam centers or be left behind.