domain:science.org
ThisIsSin
Liberty has an anti-privilege bias
5hr ago
“Consent to allow someone to view something that exists only as a moment in time or exists only in their memory is very different than consenting to allowing someone to have something that remains a permanent object and can be shared or viewed indefinitely,” Gibson said in the interview.
Or in other words, she can't monetize repeat viewings, and her self-worth is (in this case rationally) inextricably linked to being able to do exactly that.
Holding both that "sex work is real work" and that it's proper to be unable to function for 2 weeks because the only quality you could monetize has been made available for free is a logically consistent position.
Of course, a politician who [clearly demonstrates they believe] their main asset is something they got for free should not be a politician.
voters-eliot-azure
patently unbiased
5hr ago
When you say "almost everyone", who would you exclude? I agree directionally, but where we probably disagree is on who is included or excluded in "almost everyone".
YoungAchamian
5hr 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?
It's learned statistical representations and temporal associations between what Mozart is and what hip hop is. Statistically Mozart and Hip hop likely have no statistical co-occurrence. When you ask if Mozart liked hip-hop, the model isn't "thinking," "Mozart lived before hip-hop, so no." Instead, it generates text based on learned probabilities, where statements implying Mozart enjoyed hip-hop are statistically very rare or nonsensical.
Do you think that fact was explicitly hand-coded in?
I specialize in designing and training deep learning models as a career and I will never assert this because it is categorically wrong. The model would have to be very overfit for this to happen. And any company publishing a model that overfit is knowingly doing so to scam people. It should be treated similar to malfeasance or negligence.
To predict them well, it must compress latent generators: seasons, cities, typical temperatures, stylistic tropes. When we bolt on retrieval, we let it update those latents with fresh data.
I strongly agree that latent spaces can be surprisingly encompassing, but I think you're attributing more explicit meaning and conceptual structure to LLM latent spaces than actually exist. The latent space of an LLM fundamentally represents statistical relationships and contextual patterns derived entirely from textual data. These statistical regularities allow the model to implicitly predict plausible future text, including semantic, stylistic, and contextual relationships, but that doesn't amount to structured, explicit comprehension or 'understanding' of concepts as humans might interpret them. I'd postulate that GLoVe embeddings act similarly. They capture semantic relationships purely from statistical word co-occurrence; although modern LLMs are much richer, deeper, and more context-sensitive, they remain statistical predictors rather than explicit world-model builders. You're being sorta speculative/mind-in-the-clouds in suggesting that meaningful understanding requires, or emerges from, complete contextual or causal awareness within these latent spaces (Which I'd love to be true, but I have yet to see it in research or my own work). While predictive-processing metaphors are appealing, what LLMs encode is still implicit, statistical, and associative, not structured conceptual knowledge.
RLHF shapes behavior. It does not build the base competence.
RLHF guides style and human-like behavior. It's not based on expert truth assessments but attempting to be helpful and useful and not sound like it came from an AI. Someone here once described it as the ol' political commissar asking the AI a question and when it answers wrongly or unconvincingly, shooting it in the head and bringing in the next body. I love that visualization, and its sorta accurate enough that I remember it.
By insisting on “explicit, grounded, structured” you are smuggling in “symbolic, human-inspectable, modular”. That is a research preference, not a metaphysical requirement. Cognitive science moved past demanding explicit symbol tables for humans decades ago. We allow humans to count as having world models with distributed cortical encodings. We should use the same standard here.
I'll consider this, will probably edit a response in later. I wrote most of this in 10-20 minutes instead of paying attention during a meeting. I'm not sure I agree with your re-interpretation of my definition, but it does provoke thought.
George_E_Hale
insufferable blowhard
5hr ago
I know you're male, and I think you have a family. Beyond that you're one of the several on here for whom no picture emerges.
netstack
Texas is freedom land
5hr ago
“Directional whoring?” That’s ridiculous and insulting.
None of my half-dozen female cousins are whoring themselves out. None have any overlap between their careers and their relationships.
That’s not because they’re all following the same script, either. One’s an accountant, another works for the government. A third got pregnant in college, but married the guy and started a stable family. She only has a job now that the kid is in school. Not exactly a gold digger.
What’s your justification for insulting the modal Western woman?
voters-eliot-azure
patently unbiased
6hr ago
FWIW, the opposite extreme ideology is easily dismantled as well: that the West in perfectly meritocratic and there is no need to study or even acknowledge power structures that affect and influence socioeconomic conditions. I suppose I could call this "right-wokism" and attack it as a strawman.
Woke, when I was first introduced to the concept from a leftwing perspective, would be the middle ground: an acknowledgement that arbitrary[1] power structures exist that continue to exacerbate adverse socioeconomic conditions. To be "awake", or "aware" of those power structures. It wasn't a call-to-arms, but more of a sly-wink of "Hey, be kind to one-another, because things don't have to be this way."
But now, woke as it's used from a rightwing perspective, is an extremism as you've described: that all socioeconomic conditions are due to perverse power structures that benefit only white men (self-victimization), and they are therefore thieves and exploiters.
My personal take, before anyone tries to paint me as a believer of a specific ideology, is not necessarily that government needs to play the role of dismantler of those power structures, but that it definitely should not continue to enable them to fester as open wounds in the social fabric of our society. E.g. don't test nuclear bombs near the indigenous peoples, but maybe also don't shoehorn social justice concepts into every bit of middle school curricula (just read a link from the Freddie de Boer post linked downthread).
[1] arbitrary, in the sense of an opposite of meritocratic
roystgnr
6hr ago
Do I just put my foot down and confront her, pushing her to be serious about her health?
This could horribly backfire, depending on what's underlying her behavior. I helped someone get over a very serious eating disorder, a very long time ago, and one thing that quickly became clear in that case was that the disorder had developed and been reinforced primarily because it provided a feeling of power and control in a life that had been very heavily controlled by others. Part of the solution was to logically explain just how self-destructive the disorder was, but a bigger part was to improve the level of and awareness of more wholesome ways to assert self-control, and to aid in that self-control in a way that made me seem like an ally rather than just another oppressive external source of control. I fear even the "logically explain" bit might have been counterproductive if I wasn't the sort of nerd who mostly interacts with the sorts of nerds that that kind of thing actually works on.
That all sounds ridiculously vague, partly because I'm trying to be respectful of privacy, but partly because your girlfriend may have a completely different underlying problem, and I don't want to give the impression that I'm recommending a particular fix rather than just a search for a deeper problem.
I also feel like it's cruel but necessary to point out that there may be no fix. A BMI of 16.3-and-decreasing is getting into the range typically associated with anorexia. Anorexia gets called "the most lethal mental disorder" because even when it's professionally diagnosed it's not always professionally remediable. Don't blame yourself if it turns out that you can't figure out a remedy here either. Getting her doctor and sister on the case may have been the best you could do, and encouraging and supporting them may be the best you can do now.
faceh
6hr ago
Yeah, I will say they might literally be the ONLY exchange I know of that was fully expecting, well in advance, the need to navigate regulatory environments and fight off attempts by regulators to bully them, and the plan was more than "ignore it until they're kicking the door in to serve a warrant."
I was in early enough to see what happened to Mt.Gox, so my choice of exchange way back then was very carefully reasoned, and Coinbase seemed like the only one that wasn't grown from tainted seeds (i.e. drugs, gambling, or money laundering).
Would make very little sense to chase 100x gains whilst ignoring the 20-30% chance of your preferred exchange getting fined or sued into oblivion or crashing due to incompetent leadership.
Kraken has been alright to me as well.
erwgv3g34
6hr ago
I was not prepared for their Wikipedia photo.
If she was born in 1951, she must have been 57 or 58 at the time of the photo. She aged incredibly well.
voters-eliot-azure
patently unbiased
6hr ago
There have been various attempts at defining "wokism", but for me the distinguishing characteristic is the set of tactics it employs and not just its goals.
Pulling at this thread more - wokism isn't (strictly?) an ideology, but a set of tactics to bring about social change. I agree that many of these same tactics are being used - or have been historically used - by the right. And, might I add, for every tactic to bring about social change "the left" has that "the right" doesn't, there also seems to be one "the right" has that "the left" doesn't, e.g. evangelism.
That does make its comparison to Marxism interesting, though, if one views Marxism as an ideology to bring about revolutionary social change to end the class struggle under capitalism. But apart from self-described leftwing revolutionaries, I don't personally know anyone "woke" who desires revolutionary change rather than incremental change, because incremental change seems to have been working pretty well over the past ~60 years or so. Someone recently posted "capitalism, but nice" in this thread and that's pretty much the extent of "woke" that I experience. Otherwise we would just call them communists. But if we're saying that woke = communist then we're back to the original strawman position.
gattsuru
6hr ago
It may be worth evaluating whether it's a fear thing, a social thing, or a genuine lack of hunger; all are problems, but they'll have different solutions.
You're treating it like it's a social thing (habits, how others see her, so on), so assuming that's the case, I'd caution that it's usually more effective to work within existing habits than around them. Try to negotiate a slightly higher calorie intake, or add a protein requirement (even if vegan proteins), or have one meal once a week away from the cell phone, rather than get rid of the calorie-counter app entirely. Suggest vitamin or macronutrient supplements rather than changing what's on the dining room table. That'll not only avoid problems with being controlling; it should also make it easier to acclimatize toward.
Regardless of approach, be aware that sustained significant increases in calorie intake (or most macronutrients) aren't much easier to actually do than decreases.
quiet_NaN
6hr ago
Not every capital allocation decision is beneficial. Housing in cities is famously supply-inelastic: if you increase the prices of houses by another factor ten, this will not result in much increase in the supply. If we magically prevented billionaires from investing in cities with high rents, I doubt that there would be bad consequences.
Or take the stock market. Nvidia has a net profit of 76G$/year and a market cap of 4T$, so it is worth about 50 years of profit. If there was less capital around to be invested, it might only be worth 2T$ instead, but I fail to see what would be so bad about that.
Some "capital allocation decisions" are actually better seen as consumption in disguise. When Bezos invested in Blue Origin, or Musk bought Twitter, that read to me as much as a consumption decision as some nerd buying Magic boosters. Sure, it is always possible that the cards will appreciate over time, but the real value for the buyer comes from the joy and prestige of ownership itself.
I think that it is good that people who for whatever reason are good with investing money have capital to invest, at least assuming the investments are done is broadly pro-social endeavors (which can be controlled through regulations to some degree). I do not however think that this is the only good use of money, and for example would be opposed to giving taxpayer money to successful investors so that they can invest even more.
MaiqTheTrue
Renrijra Krin
6hr ago
Personally, I think that it’s an identitarianism form of Marxism. The idea being that those at the top of the hierarchy got their wealth, power, and positions by exploiting those who are not in the dominant group. So in the West, white people, particularly white men, got everything they have from either past theft and exploitation or current theft and exploitation. And thus the belief suggests that the way to solve this injustice is to take from those who have and give it to those who don’t.
WhiningCoil
Ghost of Quokka's Future
6hr ago
Coinbase has always been more professionally run that other crypto companies. They brought on experienced corporate lawyers, auditors and executives early, and were probably the most serious and normie coded crypto company. I think they are the only one who's been regularly audited by an accredited outside firm as well, something FTX and Binance and others refused to comply with claiming it was impossible.
WhiningCoil
Ghost of Quokka's Future
6hr ago
I mean, at the end of the day, she's only your girlfriend and it's none of your business. I can't pretend to know your relationship, but imagine if you will she came to you and started taking a deadly serious and negatively slanted interest in something you did? Pathologized how much you played video games, or arguing with people on the internet, or watched porn? If you don't actually do any of those things, replace it with something you do do.
snigger do do.
I guess my point is, even if you are right, you're still just a boyfriend. Probably the latest of many, others of whom never had a problem with this before. This is more of a fight for her family and doctor to wage.
That said, my wife had some eating disorder adjacent behavior that didn't stop until we decided to have kids. After putting a baby in her 6 years ago they haven't re-emerged even slightly.
Not that I'm suggesting you rush anything.
self_made_human
Grippy socks, grippy box
6hr 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?
Anyway:
Instead, current research strongly suggests that LLMs are primarily pattern-recognition systems that infer regularities purely from text statistics rather than internally representing the world in a structured, grounded way.
This is the core of our disagreement. I'd argue this is a false dichotomy. How does one become a master pattern-matcher of text that describes the world? The most parsimonious way to predict what comes next in a story about balls falling or characters moving between cities is not to memorize every possible story, but to learn an implicit model of physics and geography.
Which we know happens:
Large Language Models develop structured internal representations of both space and time
The researchers discovered that large language models (LLMs) develop structured internal representations of both space (geographic locations) and time (historical dates/periods) during their training, even though they’re only trained to predict the next word in text.
There's a whole heap of mechanistic interpretability research out there, which finds well-ordered concepts out there, inside LLMs.
You can find more, this Substack has a good roundup.
You say: “The LLM cannot know today’s weather, only the scaffolding can.” True. That does not bear on whether the base model holds a world model in the predictive-processing sense. The base model’s “world” is the distribution of texts generated by humans who live in the physical world. To predict them well, it must compress latent generators: seasons, cities, typical temperatures, stylistic tropes. When we bolt on retrieval, we let it update those latents with fresh data. Lack of online weight updates does not negate the latent model, it just limits plasticity.
The report it is producing was RLHF-ed to look correct
RLHF shapes behavior. It does not build the base competence. The internal “truth detectors” found by multiple groups are present before RLHF, though RLHF can suppress or amplify their influence on the final token choice. The fact that we can linearly read out “lying vs truthful” features means the base network distinguished them. A policy can still choose to ignore a feature, but the feature exists.
On your definition of a world model:
By insisting on “explicit, grounded, structured” you are smuggling in “symbolic, human-inspectable, modular”. That is a research preference, not a metaphysical requirement. Cognitive science moved past demanding explicit symbol tables for humans decades ago. We allow humans to count as having world models with distributed cortical encodings. We should use the same standard here.
ABigGuy4U
6hr ago
Florida and many other states have a mandatory two-semester Holocaust class.
faceh
6hr ago
They had assets on the book worth more than their market cap
I'm sure FTX also had assets 'on the books' large enough to cover all their liabilities and even lawsuit/regulatory risk if they got targetted.
But the books were cooked, which is seemingly a common factor in crypto exchanges, with insider trading and light fraud showing up with some regularity. Ironically given that one use of crypto is keeping all books open for easy scrutiny at all times.
Mostly joking. Coinbase is, I think, the longest running American Crypto Exchange and has weathered several storms so I'd not be the one betting against them. I've had an account with them since they used to hold private keys printed on physical paper stored in physical lockboxes. But there was definitely a period of time around and after the IPO where it wasn't clear if the Gov't was going to really crack down on exchanges or liberalize the rules.
We're clearly in the liberalizing timeline.
Freddie's post sounds like ravings of exhaustion from having to fight a broad and deep set of ideological concepts that all have shared roots in 20th century social liberalism (feminism, civil rights, etc.), and his solution is to pigeonhole all those ideological concepts into a single overarching theory that can be attacked directly without having to get into the weeds and nuances of any individual ideology. But also, he says that it's not his responsibility to perform this abstraction, but that all of these separate ideologies must bring themselves under a single banner? For his convenience?
I don't see the appeal of his writing, either. This is the only snippet I've read, but I've stumbled across his name.
Edit: I've read more of his writing. This post seems to be written in an intentionally exasperated voice.
TowardsPanna
7hr ago
unfortunance
Had to check if this is a real word. It is not. :(
Yeah and there's constant double standards about this stuff. "Manosphere" and all that gets maligned like it's automatically slightly evil and should be combated. Bullshit harmful sources of programming for girls/women? crickets Just some concern about whether it's not quite girlbossy enough??
TowardsPanna
7hr ago
Ask her how she arrived at the calorie goal in the app.
I might as well ask.
More options
Context Copy link