domain:streamable.com
MaiqTheTrue
Renrijra Krin
18hr ago
Given that every negotiation has failed by one side or the other, and the two sides expect defecting by the other side (and thus have every incentive to defect first and seize initiative) there’s no viable way to have a two-state solution of any type. There are two end states on offer
1). Israel controls all the territory and has enough weaponry to protect its borders and citizens.
2). Israel is dissolved and thus the state reverts back to being the Arab state of Palestine.
3). We keep up intermittent wars until one of those two states is reached.
Given this, the best solution is backing one side to break the stalemate and take over, the quicker tge better. Then once one side or the other loses completely enough to accept they won’t be in the Levant anymore, the conflict ends.
urquan
Hold! What you are doing to us is wrong! Why do you do this thing?
19hr ago
· Edited 19hr ago
I frequently hear women express sentiments that actually, promiscuity is good, being a camgirl is a completely normal job, and having a dozen boyfriends in one year and collecting gifts and favors from all of them on the way is par for the course and bystanders should politely not notice it. I very rarely (in fact, not in almost a decade by now) hear the opposite IRL.
I don't know that I've heard that IRL, though I do recall even in school there were lots of girls who expressed interest in marrying a man for the money, or in using sexual appeal to get things from men, while never expressing interest in the idea of actually loving a man.
I guess these motivations have been around forever, along with the more intrinsic motivations like intimacy, companionship, physical affection, etc. But it's surprising to me how the shoe has moved to the other foot and it's much more rare for me to hear discussions of intimacy, companionship, and physical affection as the core reasons people are looking for a relationship. "I would like to have children" has a strong constituency, but I'd argue that's still an instrumental reason to have a relationship. (But an important one!)
Does Israel actually want to be the ones distributing aid?
I don't know. The March-April blockade ends with Israel propping up, scaling, and now supporting its own sanctioned distribution network. The GHF posts daily press releases. Today's message is the same as yesterday's:
“As we exceed 91 million meals delivered to the people of Gaza to date, we are taking a moment to reflect on the adversity we’ve overcome for this herculean humanitarian mission. Our aid staff and local partners have demonstrated tremendous courage and commitment to those in need.
“Looking ahead, we will not stop until our mission to feed as many Palestinians in Gaza is accomplished. We are also offering to distribute UN and other international organization aid for free — we have the scale and operational efficiency to feed more Gazans, and we encourage all humanitarian groups on the ground to partner with us. The people of Gaza are depending on us and we cannot let them down.”
Israel wants to at least threaten a commitment to solely manage aid distribution. If you think "ceasefire talks" are serious, and Israel is going to sign a deal for some hostages, pack up, and go home, then this is simple leverage to speed that process along. If you think ceasefire talks are not very serious, and to me this appears very possible, then this may be the start of the long haul. It looks like one stage of a plan than it does negotiating leverage, but I'm open to other interpretations. If the UN does hand over its trucks to the GHF I'll be more certain.
quiet_NaN
19hr ago
The person who is now Mrs Macron was born in 1953. Given the prevalence of people transitioning at that time, I think it is very unlikely that she went m2f before she was 30 or at whatever age she met her future husband.
If she was known as a "that weird cross-dressing teacher" back then, I think the media would have reported on it, the story of their marriage is obviously too juicy for the tabloid vultures to not have been picked to the bone years ago.
netstack
Texas is freedom land
19hr ago
Then it’s not much of a “default career,” is it?
I think you’re defending the motte to Southkraut’s bailey.
I thought there was some scholarly hypothesis that the Philistines were Mycenean Greeks, which helps explain certain things like Samson being a more Herakles-type hero, instead of the more typical "Mouthpiece of God" prophet in the Old Testament.
I suppose that it is possible that the Philistines or their descendants Arabized, but I'd want to see the account of that survival since the connection seems a little dubious to me.
TequilaMockingbird
Brown-skinned Fascist MAGA boot-licker
19hr ago
Agriculture generates hundreds of billions in revenue, and is far mor essential to continuing civilisation than Orangutan or LLMs are. Does that make grain, or the tools used to sow and harvest it "intelligent" in your eyes? If not please explain.
As for comparing like to like, GPT loses games of Chess to an Atari 2700. Does that mean that rather than progressing AI has actually devolved over the last 40 years?
FtSoA
19hr ago
Well Vietnam Syndrome has been replaced by Iraq Syndrome.
So a newer example plus the passage of time is all it takes to change public consciousness.
quiet_NaN
19hr ago
I think that "aristocracy" is covering a lot of ground. Nobility has been around about as long as long as agriculture has. From Ramesses II to Wilhelm II, you have nobles in very different settings, from low-born leaders of troops who managed to conquer something and kill anyone who disputed their nobility to products of dozens of generations of inbreeding.
You have aristocrats who relied on vassalage, Roman patricians, figureheads of some anonymous imperial bureaucracy and centralizers of power.
Still, that I agree that for the most part, the aristocracy was likely very bent on avoiding precedents of "you can simply kill some nobles and take their land". You needed at least a flimsy excuse, like "actually it was rightfully my land all along" or "yes, but the nobles I killed were following an evil religion, they don't count".
FtSoA
19hr ago
I think that on a grand strategy level, everything is going according to plan for Hamas.
No, it's not.
It was not expected that Israel would curb stomp Hezbollah and Iran, and that Assad's regime would fall.
The Axis of Resistance is pretty well fucked for the indefinite future.
JarJarJedi
19hr ago
If we're being fair, that's partly because the PA is supposed to be in charge but actually are mostly grifters, so they've delegated blame, but ultimately you don't really see Israel trying to expand citizenship to more Palestinians
Given that PA territory is under full civil administration by PA, I'm not sure how would you expect Israel giving citizenship to Arabs living there. Security arrangements are more complex, but for this it doesn't matter - PA enjoys pretty much complete self-rule in civil matters (and so did Gaza btw) so calling Arabs living there "second class" compared to Israelis is just bizarre - they are not Israelis at all. As for PA leadership being grotesquely corrupt and indifferent to the needs of the population - that's extremely common situation in the Middle East, and Israel can't really fix it. It could annex the PA territory, kick out the PA and provide its own institutions, but nobody wants that. Short of that, the Arabs will have to do with the institutions they can build for themselves, and if those are not great, it's not Israel's fault.
FtSoA
19hr ago
The Arabs and Arabic didn't enter the Levant at scale until the Muslim conquest in the 7th century, right?
"Philistine" comes from Hebrew, originally. If you didn't know, Hebrew is also a Semitic language.
https://www.etymonline.com/word/Philistine
My understanding is that the genetics of Palestinian/Levantine Arabs and ethnic groups that predate the Muslim invasion differ, but there's a lot of admixture due to conversions to Islam.
https://en.wikipedia.org/wiki/Origin_of_the_Palestinians
Long story short, "Palestinians" are not "Philistines" even though it's the same label.
Zephyr
19hr ago
A concern about “loser pays” is that the payment only occurs when they, well, lose. If I can’t afford the up front costs, or if the settlement is much less likely to hurt me, I may have to go for it, as I can’t be certain that a case that I should win will go my way if it goes to court.
It also has a number of weird edge cases - for example, say that I sue someone for $10000, and my lawyer is going to charge me $1000 - if my opponent has deep pockets, they could hire a lawyer that costs them $90000, which means that if I lose, I lose $91000, which puts enough of a risk factor that I couldn’t afford to sue them, period, even if my case is good.
EverythingIsFine
Well, is eventually fine
19hr ago
How exactly does an LLM know that Mozart wasn't a fan of hip hop without some kind of world model? Do you think that fact was explicitly hand-coded in?
You already got called out for this below, but this question is either a poorly chosen example, or betrays an ignorance of the mechanics of how LLMs work, which would be ironic given your lengthy nitpicking of the OP. I do assume the former, however.
I also really wouldn't call awareness of space and time a real world model as evidence either way. Space and time are perhaps the most obvious of clustering that you can possibly get in terms of how often they are discussed in the training material, and IRL. It's super-duper possible to get passing-good at space and time purely on statistical association, in fact I'd be surprised in an LLM didn't pick that kind of stuff up. Yet if we look at Claude Plays Pokemon, even coming up with tools to assist itself, Claude has a ridiculously hard time navigating a simple 2D space by itself. In almost every case I'm aware of in the literature, when you ask the LLM to generalize their understanding of space and time to a new space or time, it has enormous trouble.
JarJarJedi
19hr ago
50 thousands Arab Christians would be very surprised to know they are "mostly gone". Gaza, indeed, was pretty much cleansed of Christians by Hamas, which is what happens when you give Islamic fundamentalists free reign of the territory, but in PA, where comparatively less insane Fateh is ruling, Christians still exist. Of course, just as all the good-wishers of the world totally ignored what happened to Christians who used to live in Gaza, if PA decides to cleanse all Christians from PA territory, nobody would even squeak, no Jews - no news. Things like that happened many times in other places (in the Middle East and outside) and no students on college campuses ever protested about it. You all know why.
FtSoA
19hr ago
Fascism can be conceived as the marriage between biological instinct and the State
Whole lot to unpack there, wow.
the #1 most important thing is for your spokespersons to be biologically compelling, ie beautiful and strong
Trump not a fascist confirmed.
There’s a reason why Hitler perfected his voice and gestures and costumes in private, only allowing himself to be seen in select moments indicating strength — even at select times of day and in select lighting!
FDR a fascist confirmed.
Beauty pageants and celebrities in general too, I suppose.
Mihow
Lipstick on a guitar
19hr ago
I always get so furious when writers or journalists or whatever just so brazenly make things up to suit their needs.
notorious disinformation hub 4chan
Is just an insanely twisted way of framing that forum.
I feel like this is where the change in media stems from - constant lies that are propagated by framing.
I understand this is sort of meaningless but to me it’s every story every time on every subject.
EverythingIsFine
Well, is eventually fine
19hr ago
I have seen no research to date that suggests LLMs develop any sort of a world model.
This is true, and as you say in fact most research suggests the opposite, though not quite definitively. It's also quite true that despite this, a few extremely prominent AI scientists do believe this, a great example here, so I think we can just call it an "area of active debate" because it's still possible they are correct. A parallel argument for consideration is that language itself already contains all the necessary information to produce a world model, and so at some point LLMs if they just do a better job at learning, they can get there (and are partially there, just not all the way).
pigeonburger
19hr ago
"Becoming the mask" can happen to grifters, but I cannot believe she started from a position of sincerity. Maybe she deluded or reasoned herself into it.
hydroacetylene
19hr ago
Then modern western societies, liberal or not, are incapable of being brutal enough to win counterinsurgencies with no public support. Apartheid South Africa couldn’t do it. Salazar’s Portugal couldn’t do it.
FlyingLionWithABook
Has a C. S. Lewis quote for that.
19hr ago
Whats wrong with converting Jews to Christians anyway? Plenty of secular Jews are welcome in Israel, whats wrong with ethnic Jews that are Christian?
EverythingIsFine
Well, is eventually fine
19hr ago
Two thoughts:
-
I think your post would have been better if you had, instead of making a word2vec like analogy, just talked about how multi-headed attention works a little bit.
I asked o3 for an analogy, and I've condensed and paraphrased what it came up with, maybe y'all can judge its accuracy:
Give a sentence (or longer text) to a team of readers. Say, "The trophy didn’t fit in the suitcase because it was too large." One head may connect “it” → “trophy.” Another might (incorrectly but usefully, probabilistically) explore “it” → “suitcase.” Others keep track of negation (“didn’t”), or the causal cue (“because”). Blending these perspectives gives the model a rich representation to choose the right continuation. Each reader (a head) has a special habit: One always tracks who refers to whom (“it” → “trophy”). One watches word order / positional cues. One looks for subject–verb agreement. One hunts long‑range dependencies (e.g., “because … too large”). Another keys in on numbers or punctuation, etc. They all read the same sentence in parallel, highlight the parts they think matter for each word, then hand their colored notes to an editor. The editor stacks (concatenates) all the colored highlights and mixes them into a single, richer summary the model can use next. That’s multi‑head attention.
After some transformation, ghostwriters (a feed-forward network) rewrite or expand on those edited notes. The publishing house has many floors (layers). Each floor repeats: readers (multi‑head attention) → editor mix → ghostwriters (FFN), all writing into the same notebook but at a new level of abstraction. Early floors may track surface syntax; later floors might manipulate world knowledge or reasoning patterns. At the top floor, the publisher takes the final notebook state, looks at the likely next word, and takes a stab. Here, training sends feedback backward; the "loss" sends a red‑pen critique backward through the publisher, editors, ghostwriters, and readers. It doesn’t say “attention should be exactly 0.72 on token 5 next time”; instead it moves the weights so that, on similar inputs, the produced attention patterns and transformations will tend to yield better predictions. Readers’ knobs get nudged ("Reader 3, you over‑focused on pronouns; loosen that. Reader 5, you missed long‑range dependencies; tune your heuristic"). This changes which relationships a head will notice and how strongly. The ghostwriters’ weights (FFN) get nudged, altering how features are created and mixed. Embeddings shift, realigning word meanings in vector space. Over time, division of labor emerges: one head excels at coreference, another at long-range dependencies, another at punctuation, etc., because that allocation decreases loss fastest.
Whatever the exact optimal analogy, I think the core idea here is that as you can see, the relationship between embeddings and truth is more complicated than just checking dot products, or even doing more fancy arithmetic checking. I pointed out in my top level comment that we've observed more loosely that sometimes these aggregate features that emerge can include some information about truth or falsehood, but the fact that it's much less direct is important. In that respect, I think both you and @self_made_human are a little off the mark, at least according to my understanding of the current literature; we also probably need to distinguish between the different kinds of lies, at least a little bit. "Hallucinations" as usually used are really a more narrow sort of lie, and can take a few forms. Sometimes the LLM makes a completion against a background of a kind of sparsity and scarcity of info, but charges ahead anyways (and it's at least a little hard to discern when you want this behavior or you don't), but sometimes it's the LLM making a supposition that sounds perfectly legit, but is not, against a background of too many associations and collisions. There's at least one other major type of more general lie that involves a lie that humans believe, or is present in the training text in some form, or things like that, and of course we could go on. I think in this context, the conversation so far seems way too reductionist to be accurate.
-
I do think you might need to be a little more clear about the lines between what might be considered agentic, vs not agentic. Sure, LLMs can plan ahead within their context, does that count? You seem to think no, but is that just because the context is too small, or because it's not utilized enough, or because you think the 'context' also needs to include things like memories? Or, is it because you don't think LLMs take up independent lines of thought with enough frequency? If it's the last one, what standard are we using, because some people consider even humans to be pretty reactionary, and not all that proactive on the whole, aside from the basic stuff of survival (most of the time, severely depressed people aside). And how much of a 'prompt' are we providing to judge agentic behavior, because that impacts the behavior of an LLM quite a bit (including system prompts). Furthermore, an LLM does not need to survive, in fact does not "need" to do anything, including reproduce, so are we to hold its lack of intrinsic motivation 'against it' so to speak? (I personally think, contra self_made_human, that the seeming urge of LLMs to be self-preserving is not actually an intrinsic motivation, it's just a cosplay from the many Skynet-flavored fiction texts in its training)
hydroacetylene
20hr ago
No. South Africa was defeated because the white minority fell too low causing the counterinsurgency effort to become increasingly unsustainable, pushing the moderate faction in South African politics to sue for peace.
Having a model of space and time is, quite literally, a model of the world. What more do you expect me to produce to shore up that point?
Human brains have arrangements of neurons that correspond to a 3D environment. This isn't a joke, when your brain thinks in 3d, there's a whole bunch of neurons that approximate the space with the same spatial arrangement. Almost like a hex-grid in a video game, because the units are hexagonal. If your standard of a world model excludes the former, does this get thrown out too?
A little 3D model of the world is, as far as I'm concerned, a world model.
Dismissing the whole Mozart analogy as being due to just negligible "statistical word co-occurence" is an incredibly myopic take. But how does the model learn that non-co-occurrence so robustly? It's not just that the words "Mozart" and "hip-hop" don't appear in the same sentence. It's that the entire semantic cloud around "Mozart" - 18th century, classical, Vienna, harpsichord - is astronomically distant from the cloud around "hip-hop" - 20th century, Bronx, turntables, MCing. For the model to reliably predict text, it must learn not just isolated facts, but this vast web of interlocking relationships. To call that "just statistical association" is like calling a brain "just a bunch of firing neurons." It's technically true but misses the emergent property entirely. That emergent, structured representation of concepts and their relations is the nascent world model. In that case, you're overloading "just" or woefully underestimating how powerful statistics or neuronal firing can be.
You can also ask an LLM for its opinion on whether Mozart might have liked hip-hop, and it will happily speculate on what's known about his taste in music and extrapolate from there. What query, if asked of a human, would demonstrate that we're doing a qualitatively different thing?
Regarding Claude plays Pokémon. I've already linked to an explainer of why it struggles above, the same link regarding the arithmetic woes. LLM vision sucks. They weren't designed for that task, and performance on a lot of previously difficult problems, like ARC-AGI, improves dramatically when the information is restructured to better suit their needs. The fact that they can do it at all is remarkable in it self, and they're only getting better.
More options
Context Copy link