If these things were clean cut, they would have been dragged to court already many times over. For messing with teenage girls for a laugh 10 years ago. For tacitly approving genocide in Myanmar. For cheating on their video views during the highly successful pivot to video. A good lawyer will get them out of this one too with but a slap on the wrist. They exist in a gray zone where they can fuck up as much as they want to without having to fear great consequences. Vote for politicians who want to regulate these companies more.
FriendOfDeSoto
joined 2 years ago
So what they're saying is that when you train a model on material biased against minority groups it will turn out to be biased against these minority groups? Truly shocking, totally unexpected, and really no one could have foreseen this. Because bias in so-called AI was really just discovered this morning! JFC.
I don't think this is right-wing specific. You could probably draw similar conclusions coming from an Islamist angle. And this so-called AI is going to be next "frontier" is not all that clear to me yet. It's a nebulous threat at this stage that starts with a lot of "imagine if" arguments. We don't know yet. It's worth paying attention. But we don't know if HitLLMer chatbots are going to cause more damage than the concentration camp simulation games that preceded them.
There is a good 15% of people who are drank the koolaid right-wing believers. I don't think that number has changed much in the last century. The number that changes is how many of the less extreme or undecided people in the middle they can convince they're right.
The internet is only as regulated as the least regulating country on this planet. So all it takes is a tiny island nation or a principality left over in time to break the chain. It's also conceivable (imagine if!) that a fine, upstanding citizen like Elon Musk uses the change he found in his couch cushions to circumvent any regulatory efforts anywhere to distribute otherwise regulated content via his private satellite network. The answer cannot be "let's limit speech more." The answer must be "fight back with truth and facts." If asshole ideologists use all the digital tools available to spread their bullshit, we need to fund initiatives that counter that speech with the same tools.
I say we stop listening to the opinion of people who are heavily invested in this industry for future prognoses.
It's not that clear really. His official titles would've been president and chancellor and he only got one of those in a manner the Weimar constitution legally envisioned. So the system, by which we would decide what an official title is today, was abused and then suspended all together. The title "der Führer" was basically a google translate from "il duce" in Italy and is not entirely honorific because he was leader of the Nazi party first. And he continues to be referred to by this semi-unofficial, semi-honorific title even in history books today and they don't always bother to disambiguate or add that they mean it sarcastically. So while Grok should be shot into space. And Nazi saluting Melon Usk deserves to be under this much scrutiny and more and can otherwise go eff himself as far as I'm concerned. The Ockham's razor for this gaff tells me the LLM just regurgitated book knowledge and nobody bothered to filter this with 2025 sensibilities. Not great but also more of a storm in a teacup. This won't make the top ten of atrocious things coming from the Melon.
I was also looking for a word other than 'honorific.' I find it has a positive connotation and should not apply to the titles of such infamous individuals as Hitler or Mussolini. But I could not come up with anything snappy.
Can we please stop paying attention to what this Nazi saluting man baby is saying?
God is surprisingly versatile in the good book. Old testament god is totally judgy and throws a plague of locusts at you. Or a flood if you wear the wrong fabric or something like that. Hissy fits to the hilt.
New testament, so Jesus-story god is more chill. He only kills his own son to make a point. Father of the year material.
Now, a lot of this stuff is open to interpretation. One might argue I have interpreted stuff in the preceding two paragraphs as well. I wouldn't argue against that and I'm not going to get drawn into a biblical discussion because I really don't care. I am a lapsed Lutheran protestant and the Jesus they tried to teach me about wouldn't have given a fuck about your sexuality. So if fake LLM Jesus says you do your thing, it ain't my biz who you love, I'm at least inclined to believe it was programmed with a similar interpretation of the story. And while the idea of a Jesus, Mary, and God LLM chat bot is absolutely, undeniably, ridiculously ludicrous, it is almost reassuring that there must be enough training material out there to get it to give that serene a reply.
To be fair to the former PM and author, all of these points are touched upon one way or another in the article and so-called AI is merely another layer of threat to the existence of such a small language community. She didn't consider Icelandic safe before the models popped up and just now got worried. And while it's probably too early to have scientific proof about the influence of models to back up her argument, I don't think it's nonsense to think that way. And I think the guardian headline is misleading.
Note this is not a final ruling yet and can and probably will be appealed.
I used to be more judgmental about this stuff too. Having little terrorists of my own has mellowed me. I do prefer a kid watching YouTube videos on a tablet over the kid throwing a temper tantrum at the cafe. All parents need a break. And I feel like if my parents had had the opportunity to let me watch sesame street on a handheld device in the 80s at a restaurant, they would have in a heartbeat. There were for sure other grownups heavily judging my folks for allowing me to play Tetris on the Gameboy in the 90s while we waited for our food! And I still managed to get a bachelor's degree.
At the same time, they thought watching too much TV wouldn't be good for you and I don't think that has turned out the way they feared. All gamers would automatically become sociopathic killers and they didn't. I do think "social" media has proven to be detrimental. But the internet is vaster than that BS. Most pacifying tablet use will not turn kids' brains into a rotting mess.
And as they get older you need to have numerous talks. The sex one, the online predator one, now the weird AI/chatbot one, the one that contextualizes pornography and other disturbing shit on the internet one. And at a certain point, maybe 16 or thereabouts, you have to let go and hope for the best. They will find all this stuff anyway. You probably did as well. I know I did. The only thing that has changed is ease of availability for most of this stuff. So that's what we need to prepare them for and then cross our fingers.
Are we elevating a find-and-replace script to the level of artificial intelligence now?
view more: next ›
I would say that's irrelevant for the crimes committed. And not just Americans would struggle to find Myanmar on a map. Or really care what's going on there unless it's rooting out phishing farms using abducted foreigners.
I commend your view on the matter, that when it comes to their children they will do something. That may turn out to be true. However, that's not going to be enough to get anyone at meta convicted under the current laws. They are running under a cover of diffuse authority and supervision internally and section 230 externally. Abhorent drug pusher comments are not admissions of guilt. They have good lawyers. We need new laws, more regulation, and fines that make Wall Street worried.