domain:betonit.substack.com
FtSoA
16hr ago
I feel like Nietzsche is cheating a bit. I'm familiar enough with some of his ideas to know he had some interesting things to say.
https://en.wikipedia.org/wiki/Ren%C3%A9_Girard
Undoubtedly you've seen this play out many times.
I have. Which is why I know continental philosophers seem to lean in to making it worse, as if obscurity and complexity is the point.
I'm glad some people like you do the reading to pass on some level of understanding.
RandomRanger
Just build nuclear plants!
16hr ago
intelligence is a degree of reactivity, and mechanically, current LLMs do not have this trait, they just 'make up' for it in practical usage by the sheer breadth and depth of their (text) knowledge base - but at their core they are simply good enough at the practical aspect that the lack of actual, true reactivity is partially obscured.
That's fair enough and LLM intelligence certainly is jagged, excellent at some tasks but weak on others. So is human intelligence for that matter, the mental arithmetic of our species is very weak.
Yet despite the jaggedness of that intelligence, we have unique abilities in story-telling, planning, reasoning, mathematics, problem-solving, coding that separate us from animals.
LLMs have a similar separation from PCs or traditional software, they're composers rather than merely simulating or spitting out a prefabricated concept.
The Aeneid is great fiction but it can only tell one story, exactly the same each time. A PC is more like a body than a mind, it's a vessel to be filled by something else. The Windows operating system has produced trillions in value but it's just a very long complicated bit of code that performs a fixed, planned function (with some errors in it). A video game is interactive but only in preset respects and very limited generality, it's just another piece of software.
With LLMs it's different, they have a special level of flexibility and interactivity that otherwise only humans possess. You're not buying a story or a piece of code that does one thing but a storywriter, a conversation partner, a technical assistant, a research assistant, a sycophant, a planner, a medical adviser...
Under any reasonable test of perceptiveness and reactivity I'm confident a modern LLM could pass albeit with caveats for their poor vision. But intelligence is not about abstract definitions but about concrete value added. Since it's writing code generally (unlike existing cookie-cutter tools), then that's general intelligence not just in words but in dollars. We can distinguish general intellectual labour like engineering, mathematics or software development tasks from specific limited transformations like Wolfram Alpha, a CAD program or a compiler. LLMs are doing the former, not the latter.
Talk is cheap in contrast.
More like saying that the soyuz rocket is propelled by expanding combustion gasses only for somone to pop in and say no, its actually propelled by a mixture of kerosene and liquid oxygen. As i said in my reply below, you and @self_made_human are both talking about vector based embedding like its something that a couple guys tried in back in 2013 and nobody ever used again rather than a methodology that would go on to become a defacto standard approach across multiple applications. You're acting like if you open up the source code for a transformer you aren't going to find loads of matrix math for for doing vector transformations.
The old cliche about asking whether a submarine can swim is part of why I made a point to set out my parameters at the beginning, how about you set out yours.
A Report on the Relatively Recent Desire to Kill for Preah Vihear Temple & Thai Politics
Cambodia and Thailand are shooting at each other over temple ruins again. The same ruins at the center of the last temple related border dispute. Admittedly, Preah Vihear temple looks like a pretty rad, old temple. Oddly, neither Thailand or Cambodia host substantial numbers of practicing Hindus. It's the principle of the matter, I'm sure.
I. Preah Vihear: Origins
The French, as a benevolent colonial neighbors should, sought to clarify French Indochina's territorial relationships with its neighbors in the early 1900s. This included what is now Thailand, then Siam. In the Franco-Siamese treaties of 1904 and 1907 Siam ceded some territory and Siam gained some territory. What's relevant for this post is that it was agreed a watershed would mark the border in a specific area on the Franco-Siamese border. A simple, natural barrier makes the cartographer's life easy.
Colonial overlords are popular for mucking up maps and France is no exception. True to stereotype, the French ignored terms set in the treaty. They either misunderstood the geography or decided to take the high ground for themselves. Of course no one cared about some muddy, forgotten ruins in the middle of nowhere. Not until the 1950's, a half century after the Franco-Siamese treaty of 1904, did Thailand challenge the claim. Thailand sought opportunities to build nationalist sentiment, centralize control, and develop ethnic asabiyyah. By what I am sure is mere coincidence, Thailand made its move immediately following Cambodian independence in 1953.
It's not in dispute that Thailand struck first in 1955. Thailand did not strike with bombs, bullets, or covert ops on behalf of its anti-communist benefactor. The government of Thailand did something more insidious. Thailand built a road. That road went to Preah Vihear, because it was easy and because they could. The same can not be said for Cambodia. The linked picture helps demonstrate why the Thai grievance is not without reason. According to the wording of the 1904 treaty, the demarcation of the border should be on the "line of the watershed." That big valley below the temple? That's where the water runs. The water lands on the high ground and makes it way South to Cambodia. Water-shed. The cliff? Not the watershed. What's on top of the cliff? Not the watershed. So, Thailand built a road into Cambodian territory. Despite the fact that they'd been relying on the "Annex I" French map for nearly 50 years
"You can't do that," Cambodia protested. To which Thailand responded, "You and whose French Foreign Legion?" Thailand then immediately moved troops to the temple on its fancy new access roads. In this way Thailand becomes de facto owner of Preah Vihear. Without firing a shot and for no great cause except the geography agreed with the action. Thailand can get to it, Cambodia faced a steep climb. The two nations bicker over this border dispute for the rest of the 1950s until the Cambodian government grew tired Thai intransigence.
II. The Peaceful Making of a Violent Grievance
The Cambodian government decided on a severe action. They took the case to the International Court of Justice (ICJ). The ICJ considered all the evidence and ruled in favor of Cambodia. The main reason the ICJ gave was acquiescence by conduct. Thailand knew of the map, used the map, and never bothered to object to it. Given enough time, use, and opportunity silence was, in the eyes of the ICJ, consent. Here's a PDF link to the 1962 ICJ ruling, but it's not mindblowing. Thailand begrudgingly accepted the ruling, but only after they made clear they have further cause to dispute other nearby areas. Thailand withdrew its troops from Preah Vihear Temple, but maintained a presence within several hundred meters.
Fast forward another half century and, in 2008, Cambodia lobbied UNESCO for Preah Vihear to be added as a World Heritage site. This upset Thai nationalists who were energized by turbulent, exciting times. A military backed coup a year or so prior had sent Thaksin Shinawatra -- Prime Minister, political dynasty patriarch, wealthy telecom mogul and oligarch -- into exile. If the perfidious Cambodian wasn't enough to demand action, then the fact Thailand's very own Foreign Minister supported the UNESCO bid certainly was. This led to the resignation of the Foreign Minister who, as an ally of ex-PM-in-exile Shinawatra, was basically asking for it.
On July 15th, 2008, five days after the Foreign Minister's resignation, Thai and Cambodian troops exchanged gunfire near Preah Vihear. One Cambodian soldier was killed in the skirmish. It is here that we can say Preah Vihear Temple claims its first casualty. He died not die due to complicated tensions of the Vietnam War, nor did he lose his life for reasons downstream of Khmer Rouge horror. The first man fell 100 years after a treaty led to the creation of a map. A map everyone used and a map the French likely fudged for a theoretical benefit they'd never realize. This was roughly 50 years after the ICJ decided the temple was Cambodian territory.
This conflict heats up and, over the next three years, dozens of killed and hundreds of wounded can be directly attributed to the fight for Preah Vihear. Cambodian representatives returned to the ICJ in 2011. This time they sought a ruling that would address the surrounding areas of Preah Vihear-- roads, hills, trails, and access stairways. The ICJ ruled in favor of Cambodia once more. The court declared the entire promontory as Cambodian clay which made the temple even more Cambodian. Thai officials begrudgingly grumbled.
III. Thailand's Turbulent Politics and the Shinawatra Dynasty
This past spring Thai soldiers shared misgivings with visiting Cambodian tourists. This escalated to violence in May of 2025. Then, come July, a metaphorical bombshell: Prime Minister Paetongtarn Shinawatra was suspended from her role by a high court. She is to await trial to defend her actions. The catalyst for this attack was a leaked phone call the PM made to the Cambodian head of state. PM Paetongtarn, a Cambodian translator, and the leader of Cambodia Samdech (Khmer honorific) Hun Se were all present for the private call. Their conversation concerned diplomacy in the hopes of de-escalating the violence. PM Paetongtarn's opposition leaked the call, framed it as overly friendly (calling him "uncle"), accused her of appeasing a hostile nation, and wants to investigate if she committed treason. The PM claims she did nothing but represent her nation's interests. She still awaits trial.
Here is one translation I found of the controversial phone call snippet. I didn't see anything I would consider treasonous, although if the translation is accurate, then I can see to how the PM's opposition could smear her with it. It might be too informal, too chummy for diplomacy with a nation that's shooting at you. But, if it works it works. She even states she needs to consult with her military before concluding the matter. That sounds deferential towards the military-royal establishment, her main political enemies, that are set on tearing down another Shinawatra. One small bit of irony in the translated excerpt:
PM Paetongtarn: Right now, the administration is at its weakest, ever since I took to the office, it is this matter on Cambodia which I myself chose to not respond to any allegations posed to me yet. This is because I both love and respect Mr. Hun Sen and therefore if there is anything you want, please tell me directly. Just lift up the phone and tell me. Whatever it is that isn’t news will not become news, what you saw leaked was the product of the press, when you haven’t talk with me one-on-one, when we talk as a group, these things can leak. If you talk to me personally one-on-one, there’s no way this could leak anyways.
Nope, no way this could leak. Very suspicious!
The Shinawatra family is a powerful political dynasty. The aforementioned ex-PM Thaksin Shinawatra, Paetongtarn's father, was exiled after being removed from power in 2008. PM Paetongtarn's aunt, Thaksin's sister, was similarly ousted in a 2014 coup. I'd call the Shinawatras New Money elite. The Shinawatra family's electoral power lies in rural, populist support. They are extremely wealthy and curry electoral favor by maintaining regional patronage networks in the country's North.
Shinawatras are in competition with the more conservative establishment elite. This is the traditional royalist establishment. The Old Money, Old Guard wield power from Central and Southern Thailand. It collects this power, places it in the capitol and coast, then manages much of the country with unveiled force. The Old Guard also maintains the marriage between monarchy and military. The establishment has control of the judicial branch such that ousting a PM, changing constitutional law, or dissolving political parties is barely an inconvenience. The censure and prosecution of political enemies has become routine in Thailand. The party dissolution is often predicated on the basis of criticizing King Rama X, thinking about criticizing the monarchy, or not supporting the monarchy enough.
IV. Why?
I sought out information on this conflict, because I was curious about the history of what was commonly reported as an old feud baked in blood. It's not an unreasonable assumption. We saw similar skirmishes to today just ten years ago. While the border dispute is technically old the willingness to send boys to fight and die over it is developing history. There is some military value of the land disputed, but not any significant amount when compared to 1904. These days there is some value in tourism at the temple, though I doubt enough to offset the costs of F-16 sorties.
These neighboring nations share a long border and a longer history. Each carries its own motivations that encompass more than any single, simple item like a temple. A conflict cannot be reduced to a 9th century Hindu temple. Still, the fight for Preah Vihear most resembles a manufactured conflict of political convenience to me. A nationalist narrative to be pulled out of pocket as political winds dictate or as an opportunity to cook up a kind of conflict comfort food. A home cooked war. 1000 year old feuds have to start somewhere.
Huh! Interesting. What about a siege of a medieval city - not a fortress with just a military garrison, but an actual significant permanent settlement, probably including nearby farmers, villagers, etc who fled for the "safety" of the town when the attackers showed up - when the attacking army encircles the city and doesn't allow anyone or anything in or out?
Thanks for the reply. Speaking in general, slippery slopes are tough because as I mentioned, they can be legitimate expressions of cascading effects, or they can be a rhetorically devious tool without a sound foundation. The key differentiator is if you hand-wave the actual progression too much, it effectively functions as a fallacy. Basically, fear-mongering! There are cases where "if you give an inch they'll take a mile" but there are also cases where small changes are catastrophized, so at the end of the day you kind of have to take it case by case.
I do not think that a broad assertion that all politics is a maximalist, existential struggle is accurate as a general worldview, nor a common enough viewpoint to be assumed.
In politics, victory leading to stalling out is actually more common than you might imagine. It's partially related to the idea of "political capital", where there's actually only so much appetite/time/attention/money for change to go around. Not uncommon is the situation where a major change leaves everyone exhausted and further efforts lose their urgency, or even provoke a counter-reaction in a kind of rubber banding effect. Honestly I think it's more fair to say that societies are generally biased towards the status quo, rather than constantly hopping on runaway trains. This is especially true the more lower-d democratic a society is! So clearly Weimar Germany is a bad example. I think people forget that politics is ultimately downstream of the actual opinions of regular people, not the other way around.
So when it comes to Israel-Palestine, a one-state solution would need to be accompanied by a ground-up swell of support and persuasion to co-exist. A two-state solution is almost by its very nature a compromise, and as they say, the best compromise leaves everybody at least a little angry. And didn't you yourself say that true escalation comes when both sides lose? So at least in my eyes, any two state solution, if actually implemented, is definitionally a détente. I will however concede that the higher the violence level, and the more disproportionate the representation, the less moderating influence there is. Again though I would ask the question: would a genuine attempt at a two-state solution, under Israeli-preferred lines, be accomplished via a high degree of force? I think the answer is a clear no, but I'd be interested to hear if you disagree and think it's really a plausible end-state of naked maximalist agenda-seeking by both sides. Furthermore, geographic national boundaries in particular are, historically, way more sticky than you might think. Just look how awkwardly persistent the British and European decided lines are in the Middle East overall, despite their in many cases obvious unsuitability to match the facts on the ground! (I will however concede that I'm not quite aware of anything quite as swiss-cheese as the current scenario of settlements, partitions, and general strangeness that is current Israel (broad definition) and that virtually any 'solution' requires at least some people to relocate in practice, though I assume a halfway equitable solution could be found, ideally with plenty of money)
Supah_Schmendrick
17hr ago
Does Israel actually want to be the ones distributing aid?
Yes, because they can prevent it from being directly hijacked by Hamas. (What hamas fighters do to Palestinians who take food to homes outside of Israeli-controlled safe zones is different, but that's a lot less efficient for Hamas than just seizing/being handed the aid while it's still loaded on the trucks as the UN was doing).
It was my impression that they kind of like the current situation, where Gaza mostly starves but it's not their fault directly
No, because that way Hamas controlled the aid and could continue to starve the population while keeping all the goodies for themselves to enable further armed resistance/rebuilding after the ultimate end of the war - and it will end, because the IDF doesn't have the full-time soldiery to keep up this occupation without calling up large numbers of reservists and disrupting the civilian economy.
TequilaMockingbird
Brown-skinned Fascist MAGA boot-licker
17hr ago
Well said.
Primaprimaprima
...something all admit only "TRUMP", and the Trump Administration, can do.
17hr ago
· Edited 17hr ago
Which half though?
Just read Nietzsche. If you don't get anything out of him then don't bother with the others. He represents the "continental mind" at its best.
Girard
Never heard of him.
Continentalists seem to get very mad at Analyticals misrepresenting them, without themselves having a consensus about what was "really" meant by any given thinker.
Think of philosophy as being like one big internet argument. (It basically is that, quite literally. Many of the questions we discuss regularly on TheMotte are philosophical questions.) There has never been any time in history when someone made a forum post on a non-trivial political question and everyone thought "yep, that's correct, there's nothing to clarify or add, he simply got it right". At minimum, there will be a dozen replies telling the guy how he actually got everything wrong. Frequently, these posts will wade into interpretive matters -- asking for further elaboration or clarification on point X, asking if in this particular sentence he meant Y or if he really meant Z, asking if his arguments really support W or if that's really what he even wanted to argue for in the first place, etc. Undoubtedly you've seen this play out many times. This is just what happens when you discuss complex matters using natural language. So it goes for philosophy in general.
Think about how people still, after all these decades, can't agree over whether pro-lifers "really" believe that abortion is murder. I mean, they even say in very plain language that they think it's murder, and people still can't agree what such utterances "really" mean! Skeptics will say, well they can't actually mean that, because it's not consistent with their other beliefs/behavior, or their arguments clearly don't support that conclusion, so they have to mean something else. On and on it goes.
Sometimes the interpretive difficulties with a philosophical text are literally at the level of "I don't know what this sentence is saying". Typically they're more subtle than that though. There's a persistent interpretive difficulty with Kant's Critique of Pure Reason for example over whether his metaphysics is a dual substance theory (two types of objects, appearances and things-in-themselves) or a dual property theory (only one type of object, but it has phenomenal and non-phenomenal properties). Kant was an unusually meticulous thinker, the CoPR is 800 pages of densely pedantic arguments, but on this one (rather fundamental) issue he simply never addressed it explicitly. When we're writing, we can't predict every question that every reader will ever come up with; sometimes we think something is perfectly clear even when it's not, or it just never even occurs to us to ask that particular question at all. I'm sure you can again think of many examples from your own experience.
That being said, although interpretive difficulties in natural language debates can never be entirely eradicated, some interpretations of a text are clearly better than others, and Russell was notorious for not being a particularly careful reader of the thinkers he profiled in A History of Western Philosophy. See this for example for a criticism of Russell's interpretation of Kant.
Supah_Schmendrick
17hr ago
At the exact same time Israel declared independence, an ethnic cleansing/population exchange a literal order of magnitude larger was already going on in another part of the same empire.
RandomRanger
Just build nuclear plants!
17hr ago
Intelligence is very spiky
Absolutely right. I have no doubt elephants and other advanced mammals are better than humans at some mental tasks. But the applications of these advantages are negligible, it's just an academic curiosity.
Some materials have a thermoelectric effect where if you heat them it produces electricity, no need for steam pressure. But it's very inefficient compared to boiling water and turning steam into motion. So there are only a few niche use-cases, we could live without it. The effect is unimportant.
Supah_Schmendrick
17hr ago
Israel is a small state that is going to be in constant conflict with everyone and everything around them.
RandomRanger
Just build nuclear plants!
17hr ago
Grain, combine harvesters and so on do not do intellectual labour for us. A combine harvester is a perfect example of what modern LLMs are not - excellent for a highly specific usecase and terrible in all other areas. The AI equivalent of a harvester might be something that can only write the exact same format of SQL database code with a few variations. That would be a glorified cookie cutter.
I specifically specified 'harnessing the intellectual skills' of orungatans. Nobody is doing this. Agriculture is a totally different matter.
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?
Chess is just not what ChatGPT is supposed to do. LLMs are notoriously poor at spatial reasoning tasks, this is a legitimate weakness but does not preclude intelligence. See here: https://dynomight.net/chess/
There is variation even in chess. Gpt-3.5-turbo-instruct is somehow far better at chess than o1-mini or gpt-4o or other versions of gpt-3.5. OpenAI rightly concludes that nobody wants their AI to play mediocre games of chess for 1000x market prices and devotes resources instead to making it better at what people do want to use it for. Code, maths, sycophancy, creative writing.
Supah_Schmendrick
17hr ago
They have dug the world's largest tunnel network to shelter military personnel, deliberately intermingled it with the civilian population and vital civilian infrastructure, and denied those civilians the ability to shelter in it. They are trying to get their people killed.
George_E_Hale
insufferable blowhard
17hr ago
I feel like it's a show I'd have watched when younger, and I understand their interest in it. My wife's stereotype of me is that I like dark depressing films that require thought. This isn't wrong, but it's hardly the only genre I am interested in. I expect they (my sons) have some of my traits, but also Boys.
Supah_Schmendrick
17hr ago
War crimes are only crimes insofar as both sides can agree to - and actually do - abide by particular rules of warfare, either customary or explicit by treaty. Insofar as one party either verbally refuses to or actually breaches those rules, they lose the protection of the rules and are subject to the whim of whatever the opposing party wants to do them (and can actually do/get away with doing).
Even the Red Cross accepts the concept of reprisal as a means of forcing non-conforming belligerents to shape up and fly right.
C'mon dude. If this is the third draft of the essay, I really expect more substantial rebuttal than this.
You misunderstand me. My response was not the third revision, it was the third attempt.
I don't know if you realize this, but you come across as extremely condescending and passive-agressive in text. It really is quite infuriating. I would sit down, start crafting a response, and as i worked through your post i would just get more angry/frustrated until getting to the point where id have to step away from the computer lest i lose my temper and say something that would get me moderated.
And that illustration was wrong.
As i acknowledged in my reply to @Amadan it would have been more accurate to say that it is part of why LLMs are bad at counting, but I am going to maintain that no, it is not "wrong". You and @rae are both talking about vector based embedding like its something that a couple guys tried in back in 2013 and nobody ever used again rather than a methodology that would go on to become a defacto standard approach across multiple applications. You're acting like if you open up the source code for a transformer you aren't going to find loads of matrix math for for doing vector transformations.
Why is the opinion of the "average American" the only standard by which to recognize AGI?
Why isn't it a valid standard? You are the one who's been accusing society of moving the goalposts on you. "the goalposts haven't actually moved" seems like a fairly reasonable rebuttal to me.
I had forgotten how much of your previous weak critique to the same evidence was based off naked credentialism. After all, you claimed:
I understand how my statements could be interpreted that way, but at the same time I am also one of the guys in my company who's been lobbying to drop degree requirements from hiring. I see myself as subscribing to the old hacker ethos of "show me the code". Its not about credentials its about whether you can produce tangible results.
The companies that spend hundreds of billions of dollars on AI are doing just fine.
For a given definition of fine, i still think OpenAI and Anthropic are grifters more than they are engineers but I guess we'll just have to see who gets there first.
EverythingIsFine
Well, is eventually fine
17hr ago
Attack on Titan can be loosely understood as a kind of mix between Game of Thrones, where the cast is vulnerable to sudden death (sense of foreboding and unpredictable world), Inception or the Matrix, where the action scenes are unique, tense, and cool (flying through the air with powered grappling hooks and swords is not a concept you'll see anywhere else), and Lost, where there's a compelling mystery underneath everything that's going on (and like Lost, arguably has a bit of a let-down of an ending, although not that bad).
But yeah, not for everyone. Maybe The Walking Dead would be another analogy, though the thrills and horrors of that series are a little more performative in my opinion.
Mantergeistmann
17hr ago
"Defensive" in that it's a direct response to offensive action, not that they're currently holding ground and trying not to get pushed back by a sustained opposing offensive.
EverythingIsFine
Well, is eventually fine
17hr ago
good art = deals with ‘themes’ that can be summed up in one word
Oh no, my actual #1 pet peeve from English class rears its ugly head again! (I can still hear a particularly passionate teacher of mine desperately trying to explain that a theme actually needs to say something, like a sentence, it can't just be a word)
Aside from your nice impromptu mini-essay, I will say that for whatever reason, the "first anime" someone watches, for most people, ends up desensitizing you more than you'd think to whatever flavor of strange anime logic it employs (ask me how I know, or rather, maybe don't). Also, sometimes it's just fun to have an anime that doesn't take itself too seriously, although the exact line is sometimes a little hard to discern: Eminence in Shadow for example is a terrifically fun time, poking fun at the tropes, but also... takes its responsibility to represent the tropes seriously enough that it becomes part of the attraction, thus coming a bit full circle? Almost like a stupider cousin of the thing where a too-sharp parody is mistaken for the real thing, a la Fight Club's take on fragile masculinity. Actually, now I really want a Motte thread on the movie to see that in action. A Motte Movie club night would be hilarious.
P-Necromancer
18hr ago
But slopes are slippery! It's the literal, physical nature of a slope (and the relationship between static and kinetic friction) that, once you start to move down one, you tend to continue. The argument is, I suppose, that a lot of things people treat like slopes really aren't... but aren't they? I'm struggling to think of a case where a political movement, having achieved its proximal objective, declares victory and goes home. Actually, I'm not just struggling; the idea is absurd. Individuals can do that; amorphous groups never can.
Victory draws interest because everyone loves a winner, and to divide up the spoils -- power, but mostly cachet -- you get purity tests, which rapidly become purity spirals. The intra-group dynamics drive the inter-group dynamics: if you don't keep pushing for more, you get pushed out. This is what we see in real life: victory only emboldens movements, and a couple decades down the line, they're demanding things their forebears' mocked as slippery slope arguments. They reach and reach until, finally, the public's patience runs out... then their opponents get a turn.
(This is just one mechanism. There are others.)
The civil rights movement, the moral majority, the LGBT movement, anti-communism, progressivism, interventionism; just a handful of the many, many examples from recent history.
To put it in concrete terms: obviously bullet point 2 makes bullet point 3 more likely. Well, I very much doubt it'll follow such a clean progression; there's generally more momentum to these things. Palestinians don't exactly hide the fact that a supermajority want the last point; how could letting them organize and regroup not make it more likely? It might still be unlikely -- not like any of the other Arab nations have proven able to enforce their will on Israel -- but I think it's very hard to argue it would become less likely.
But, you argue, isn't Israeli oppression a slippery slope too? If Palestine just lets Israel establish settlements in the West Bank (or whatever), doesn't that just make more thorough depredations more likely? Yes! Both sides accuse the other of starting down a slippery slope, and both are right!
(You frame this as 'backsliding' from the two state solution; because you think it's more fair, presumably? But why would Palestine see it that way? Backsliding would moving towards an Israeli-controlled single state. A Palestinian-controlled single state would, obviously, be continuing to slide forward down the same slope: Palestine achieving it's goals.)
In Germany, the Nazis rose in large part to oppose the communists, who were, at the time, the dominant political force in the country (not in terms of votes, certainly, but in terms of organization and political violence. Which was, after all, their stated path to victory). Then the Nazis, having achieved power, ruthlessly suppressed the communists; they would do the same to them if they got the chance, they said. Which was thoroughly borne out the moment the communists did get the chance!
So how, in this model, can de-escalation ever occur? Well, one side can wipe the other out, either literally or in terms of group membership; this is how the conflict between slave owners and abolitionists ended, for example. But true de-escalation mainly happens when both sides lose, I think. The Good Friday Agreement was a tacit admission from both sides that neither could achieve their full aims. And sometimes, when the swings are too quick and dramatic, the public can simultaneously lose patience with both.
(I don't know, do the modern models even have them today?)
They do; operating directly on one-hot tokens would be prohibitively expensive.
But they're not central to the power of modern LLMs. You can even run an ablation where you use unlearned, static, entirely random embeddings (so, nearly every embedding is approximately orthogonal to every other embedding; semantic similarity would have zero relation to cosine similarity). The later layers are still able to learn syntax and semantics on their own, albeit with significantly increased loss.
Which speaks to the power of transformers: you'll get far more coherent text out of transformers with even random embeddings than some novel architecture made of simple linear combinations of word2vec.
Loosely, yes. Although in the case of Israel and Gaza, Israel fully controls all the entry and exit points, and in theory controls (and asserts the sole right to control) all the internal area as well, whereas WWI Germany still had options, just worse ones, so we can't shift quite the same burden of blame on their opponents when they themselves can pick some up. So for those reasons I'd shy away from using the term as such; overall however there's a reason WWI is a major step toward "total war" as a concept. I mean we could get in the weeds about the different 'axes' for which we judge a genocide, but I'd say it falls on the 'spectrum' somewhere.
You cannot do so without, as your cat example demonstrates, having a holistic understanding of all the relevant factors at play.
The proposed loss ratio standard is a metric worth considering, but it is hardly a good single metric. If Hamas was better at fighting, more Israeli soldiers and civilians would be dead because they have tried, but not succeeded most of the time.
More options
Context Copy link