The whole thing where moderators have mostly-unchecked power in their little domain, and they can ban people or delete comments and make rules for what other humans in their space can and can’t say, is one of the most toxic features of Reddit. I think Lemmy copied it from sheer traditionalism, but it was really a mistake in both locations. It leads a bro to think they’re supposed to be in charge of the other little peons in their space, and that’s a pretty bad thing for a bro to start thinking like.

See, now you’ve got a problem. If you’d shot down the first one, Russia would have made a big noise and then it would have been fine. Now that you didn’t, now it’s weird if you start shooting them down.

My advice is to just be straight about it: Publicly announce what the line is where you’ll shoot them down, and then stick to it. Even if you just announced a date when the shooting-down will start any time they enter NATO airspace, that might be fine. But you have to stick to it. Right now you’re trying to figure out how to make them stop without shooting them down, and that approach just doesn’t work. Like you’re all surprised they don’t establish radio contact. Bro… that is not the game you are engaged in.

(You might also want to confiscate $10 billion in frozen Russian assets to give to Ukraine for each incursion, just to respond to what’s already happened… but again without shooting them down it’s not going to accomplish anything. The money’s already gone honestly, and they know that, they’re just waiting for you to figure it out and go through your whole “process” and make it official, and they think you’re stupid and weak for every year that goes by that you’re not doing that.)

Yeah, I get it. I don’t think it is necessarily bad research or anything. I just feel like maybe it would have been good to go into it as two papers:

1. Look at the funny LLM and how far off the rails it goes if you don’t keep it stable and let it kind of “build on itself” over time iteratively and don’t put the right boundaries on
2. How should we actually wrap up an LLM into a sensible model so that it can pursue an “agent” type of task, what leads it off the rails and what doesn’t, what are some various ideas to keep it grounded and which ones work and don’t work

And yeah obviously they can get confused or output counterfactuals or nonsense as a failure mode, what I meant to say was just that they don’t really do that as a response to an overload / “DDOS” situation specifically. They might do it as a result of too much context or a badly set up framework around them sure.

Initial thought: Well… but this is a transparently absurd way to set up an ML system to manage a vending machine. I mean it is a useful data point I guess, but to me it leads to the conclusion “Even though LLMs sound to humans like they know what they’re doing, they does not, don’t just stick the whole situation into the LLM input and expect good decisions and strategies to come out of the output, you have to embed it into a more capable and structured system for any good to come of it.”

Updated thought, after reading a little bit of the paper: Holy Christ on a pancake. Is this architecture what people have been meaning by “AI agents” this whole time I’ve been hearing about them? Yeah this isn’t going to work. What the fuck, of course it goes insane over time. I stand corrected, I guess, this is valid research pointing out the stupidity of basically putting the LLM in the driver’s seat of something even more complicated than the stuff it’s already been shown to fuck up, and hoping that goes okay.

Edit: Final thought, after reading more of the paper: Okay, now I’m back closer to the original reaction. I’ve done stuff like this before, this is not how you do it. Have it output JSON, have some tolerance and retries in the framework code for parsing the JSON, be more careful with the prompts to make sure that it’s set up for success, definitely don’t include all the damn history in the context up to the full wildly-inflated context window to send it off the rails, basically, be a lot more careful with how to set it up than this, and put a lot more limits on how much you are asking of the LLM so that it can actually succeed within the little box you’ve put it in. I am not at all surprised that this setup went off the rails in hilarious fashion (and it really is hilarious, you should read). Anyway that’s what LLMs do. I don’t know if this is because the researchers didn’t know any better, or because they were deliberately setting up the framework around the LLM to produce bad results, or because this stupid approach really is the state of the art right now, but this is not how you do it. I actually am a little bit skeptical about whether you even could set up a framework for a current-generation LLM that would enable to succeed at an objective and pretty frickin’ complicated task like they set it up for here, but regardless, this wasn’t a fair test. If it was meant as a test of “are LLMs capable of AGI all on their own regardless of the setup like humans generally are,” then congratulations, you learned the answer is no. But you could have framed it a little more directly to talk about that being the answer instead of setting up a poorly-designed agent framework to be involved in it.

Yeah it’s a bunch of shit. I’m not an expert obviously, just talking out of my ass, but:

1. Running inference for all the devices in the building to “our dev server” would not have maintained a usable level of response time for any of them, unless he meant to say “the dev cluster” or something and his home wifi glitched right at that moment and made it sound different
2. LLMs don’t degrade by giving wrong answers, they degrade by stopping producing tokens
3. Meta already has shown itself to be okay with lying
4. GUYS JUST USE FUCKING CANNED ANSWERS WITH THE RIGHT SOUNDING VOICE, THIS ISN’T ROCKET SCIENCE, THAT’S HOW YOU DO DEMOS WHEN YOUR SHIT’S NOT DONE YET

Yes which is why it’s so important to react firmly to someone who’s deliberately playing brinksmanship with a nuclear-armed defensive alliance, so as to correctly pass the can-I-push-the-boundaries test Putin is giving here. This is like lesson #1 about dealing with bullies.

I think they should have shot down the fighters and then sent footage of the falling wreckage to Putin overlaid with the “Curb Your Enthusiam” music and fade out with “Stay in your own airspace pls and thank you.”

Yeah. In retrospect, it always seems pretty stupid that we thought we could just resolve things with diplomacy and the rabid dog hanging around the building will probably just calm down and stay in the back section and we just won’t go there.

I feel like escalating to war because that’s clearly what’s up, and whatever happens happens, oh well, is one of the few areas where Churchill had it 100% right in the lead up to the global war.

The balls will be fastly sucked into the machine

Incarceritis

Honestly, I do agree with you on that but I also feel like any “system” is just never going to be the answer. The big lesson of all those Nepalis electing their new government over Discord isn’t that Discord is the way, it is that if huge numbers of people are engaged and willing to fight for what they need and then engage to set up something sensible, there’s a pretty good chance they’ll get it.

The American people – almost all of them, not 100% but the great majority – have just been checked out of the concept of being involved in making sure their country runs right for too many decades now for things to stay on the rails.

I think the crisis of Trump is likely to be worse than any crisis in the Western world for the last 50 years. I think the closest analogue is probably the collapse of the USSR. So yes, some of the rich people upped their wealth by orders of magnitude, and honestly you might be right that Zuck might manage to be one of that category, but also some of them lost everything or got thrown out windows, or had to survive in reduced capacity within their new walled fortresses in the horrifying new meta. I feel like more likely is that the MAGA world will remember Facebook censoring their posts about ivermectin, and not feel like Zuck needs to have a seat at the table, no matter how many ass-kissing sessions he shows up at the White House to do.

For example I feel like breaking up Meta and mandating Truth Social and TikTok as the only new sanctioned social media going forward might be one possible outcome. It’s kind of hard to say and I won’t swear that you’re definitely wrong that he might come out way ahead in the end. I’m just saying that this type of crisis is a very different type of crisis.

Part of my point is that the damage Trump is going to do will cost them tons more money than if they had helped to prop up the fairly safe and civil society they previously were allowed to exist within, under which secure umbrella they’ve been able to rake in money like leaves in autumn on a wide country estate.

My sister had a cat who was declawed (not by her), medicated, old, sort of inbred, and had lived inside all his life. A more harmless cat you could barely imagine. He regularly found it a challenge to drink his water without fucking it up, and spent a lot of time just staring at the wall.

He once found a moth that was already damaged when he got to it, and successfully killed it. He carried the moth around in his mouth, clearly basking in the intoxicating flood of pride of the hunter he was feeling, for almost an hour before eating the whole thing.

Hopefully the idiots who run these tech companies will learn their lesson soon.

“If the thing is free, you’re the product” applies just as much to dinners at the White House as it does to social media apps.

Rumsfeld more or less bragged about it too, a lot of that crowd did. We haven’t had a serial killer in the White House for a few years, so it’s an adjustment back (and of course Trump is such an overtly vicious person that it’ll have a new flavor now), but this stuff is not new for Republicans.

(Actually Biden and Bush Sr. are the only two presidents since Carter that I can remember that weren’t fans of blowing up random people directly with US forces for no reason. In Afghanistan they did an airstrike on a wedding.)

I mean honestly, US presidents killing random people and families the world over for literally no reason is nothing new. On the scale of Trump stuff, this is pretty minor, as sad as that is to say.

This wasn’t your question exactly, but here, a good horror movie for you:

https://www.dailymotion.com/video/x8hpcx5

“Pikunikku” from 1996 is supposed to be pretty good too.

First possibility that comes to mind is a backed up federation queue on lemmy.world for some reason.

It could also be some failure on sh.itjust.works, but if I had to guess I would say it’s more likely an issue on the sending end.

Honestly it seems like a fine solution on both sides.

“You have secured yourself an honored platform and you interact with it. Please buy me my own honored platform, so that we may remain as equals in the matter of the computer. Thank you for your attention to this matter.”

Ones I have not seen in other comments, although you probably have enough at this point:

• The Fifth Element - Yes I know the director is a sexual predator, it’s still really really good. One of the most creative movies ever made
• Run Lola Run - Surreal modern-style impossible narrative movie from way before movies really did that, also very well done
• O Brother Where Art Thou - Damn, we’re in a tight spot
• Schindler’s List - Warning, very very very heavy obviously, but there is a reason it won all the awards
• Conan the Barbarian - You may or may not like B movies, but if you like “Sword and Sandal” B movies, this one is the king
• The Sting - Absolutely top-notch gangster movie from before your parents were born
• Office Space - If you have ever worked a crappy tech job this movie will become your spirit animal in time of need

I’ve also heard great things about “Highlander.” Don’t watch the sequels.

There are also miniseries. “Stranger Things” season 1, “Chernobyl,” and “Game of Thrones” up until about season 5. Yeah, it’s a shame they never made any more after that, but you can just let it be. You’ll know roughly when. If the three female assassins from Dorne come on the screen, you went too far, just turn it off immediately and be sure to take a shower right away.

Edit: Additions