Olives @olives@qoto.org

@caseynewton

https://qoto.org/@olives/111191543236620885 I haven't particularly specifically written much about "AI chatbots", but I go over it briefly here in my dive into bad faith conflations between fiction and reality.

https://qoto.org/@olives/111516011246609826 I've also been over here why puritanical anti porn rhetoric does not have a basis in science.

It's quite easy to fall into the "bothsidesism" trap too, where some random guy who doesn't really understand how things work might offer up his "hot take" (perhaps, someone looking at superficial associations and offering up garbage, or otherwise not factoring in anything other than a narrow slice of what they're looking at) and someone might try to "juxtapose" that against a more informed take.

https://qoto.org/@olives/111448111216663879 Also above the horrid "grrr, there are subhumans generating porn of fictional minors" take, there is also the "let's just focus on whether this increases or decreases crime" take which is, maybe better, although it still misses some key points (and is also prone to bothsidesism where some idiot, perhaps an ethicist or some other ignorant person, puts forth their own hot takes, and don't have qualms about doing it as they're not treating people as people). So, this unpicks that, although I haven't spent a great deal of time on that either, really.

I could comment further but it's not really a line of discourse which I really want to get into a dive about right now.

I suspect that OpenAI is most likely to run into the things which particularly rub people the wrong way, as far as copyright is concerned, more so than the "open source" options. I think it comes down to OpenAI trying to build more of a brand off "professionally finished style images".

Interesting theory that validating AI would cost so much that someone might as well just hire humans to do the job.

I mean, yeah, we're also probably tired of how the original concept of copyright has also gotten ridiculously stretched, thereby depriving the public from being able to use a work, practically ever.

Honestly, people criticizing training AIs on copyrighted material is better than someone holding the same opinion, but instead trying to swat companies for things which don't really matter, like what if it generates an image of Mickey with a knife.

There is nothing wrong with someone creating an artwork that has elements of violence to it.

For years, I never really understood Substack, and this big desire to go jump onto a big corporate middleman to... Basically host a blog.

My opinion is more nuanced than just "You should use Substack", in fact I think you shouldn't use Substack (self-host, if possible), but I don't see the value in piling onto platforms either because of apparently a few very unpleasant accounts (when historically that hasn't gone anywhere good).

Of course, like Elizabeth, I also have to question whether nit-picking a few accounts posting, alright, very offensive content, isn't doing a disservice to free expression (particularly when you consider how over the top Big Tech has gone). Is it really contributing?

Setting policies, quietly "reinterpreting policies" (to avoid the embarrassment of publicly backing down), these are like Big Tech shenanigans, and not particularly fun ones.

It's not even particularly good for free speech. The approach of the neutral hoster might be better (and would include things related to sex) than acting as if they're out there to save journalism.

It's also the sort of platform that is entirely optional.

At the end of the day, Substack is a blog hosting platform too. A blog is far easier to self-host than any of these other things.

It's not really clear why someone really needs yet another middleman coming along here. It's even less clear when using that middleman involves getting dragged into that middleman's own political drama.

At the end of the day, someone just wants to host a blog. Why so many games?

In fact, looking at the current dance going on, it seems like the same sort of political games which you had going on with Big Tech.

Perhaps, it's time to skip the games and exit early onto something that you can control yourself.

Hey, I support people using their own self-hosted solution (i.e. Ghost), rather than yet another middleman (i.e. Substack).

https://web.archive.org/web/20240116024448/https://docstore.ohchr.org/SelfServices/FilesHandler.ashx?enc=6QkG1d%2FPPRiCAqhKb7yhsrHPiif0%2F1kumQo%2BD50%2F9nZB8sh%2BSXw2rU9z6OACw0gFuqtxLLzbGbY8zvV%2BJN1f3Rk%2B1b99SwrYcHSLeZkcUZFS0pl7G25fMIfpsB3dX7S%2FEqbFNqVFSe1ZgEMGnW5yiA%3D%3D

Commenting on U.N. / Senegalese silliness this time.

The term "mere representation" is extremely suspicious and should be deleted. There is no legitimate purpose for this whatsoever.

The term "realistic" is somewhat suspicious and likely problematic (i.e. chasing fictional scenarios). Something needs to be done about this.

https://www.eff.org/deeplinks/2023/10/eff-and-45-organizations-tell-un-reverse-decision-host-igf-saudi-arabia
@eff seems to agree with me that hosting the #IGF in Saudi Arabia is a bad idea.

https://web.archive.org/web/20240106034714/https://indico.un.org/event/1009455/
https://web.archive.org/web/20231114005554/https://www.intgovforum.org/en/content/mag-2024-members

The next #IGF (an internet policy event hosted by the U.N.) is being held in Saudi Arabia, a country known for brutal repression and chopping critical journalists up into tiny little pieces.

Talk about out of touch.

You might remember the IGF from this critique of Kadokawa.

https://web.archive.org/web/20231227071126/https://www.iso.org/standard/81467.html

Nothing like a standards organization like ISO engaging in silly virtue signalling, like telling you what "SDGs" this supposedly "contributes towards" (this seems to be done in a "this vaguely intersects with this, let's toss it in" manner). Also, is it ethical to charge for these standards?

https://reason.com/2024/01/15/the-newly-unveiled-hhs-rationale-for-rescheduling-marijuana-underlines-drug-warriors-dishonesty/

"Last week, in response to a Freedom of Information Act request by Houston lawyer Matthew Zorn, the Department of Health and Human Services (HHS) revealed the rationale for its August 2023 recommendation that the Drug Enforcement Administration (DEA) move marijuana from Schedule I to Schedule III of the Controlled Substances Act."