FeepingCreature
No bio...
User ID: 311
FeepingCreature
2mo ago
Trust..? I just ask it code questions, lol. They can sniff my 40k token Vulkan demo if they like.
FeepingCreature
8mo ago
But being a big-breasted female-presenting tiefling with a futa cock and dude voice? Feels like a strange midpoint. If you live in a world of magic and this can be done easily with a finger snap or a procured service, why wouldn't you go all in one way or the other? I'll admit to a possible failure of imagination on my end, but it just comes off as kink and fetishism.
If you live in a world of magic and this can be done easily with a finger snap, why wouldn't you go for kink and fetishism?
FeepingCreature
9mo ago
On one hand I want to say that surely, being able to recognize and admit misconduct is private is better than not being able to do so, so this leak is bad. On the other hand, this is a pretty impressive level of self-delusion even so, and we do want to push back on misconduct when we become aware of it.
But I guess my synthesis would be: if the only way we have of noticing misconduct in a topic as impactful as a world-wide pandemic is a leak of private messages where the scientists involved literally admit to it, then science has much, much bigger problems than these people's misconduct.
I mean, I agree, but you could imagine a society that was all work, zero play, 16 hours a day until you die. Any money you are paid for your labor is only reinvested to make you a more effective employee. Children are still raised (16 hours of schooling and training per day, enter the labor force at 12), but they refund their parents the cost of raising them and thus are merely another labor-raising device. All fun that one has is optimized for perfect recovery to maximize socially useful labor. I think if we look at why such a society was bad, we find what the proper role of fun is: this society doesn't seem to be for anything aside from itself. Is society for man or is man for society? Whereas from the "fun" perspective, or rather the "human values" perspective, we find that we don't need to justify labor: a life with a balance of meaningful challenges, self-actualization and silly fun seems more preferred, even on its own merits, than a life of only one of them. So there are two arguments for labor: first, a society with only fun quickly runs out of fun overhead. This is an argument that even fun-maximalists will embrace, but it doesn't give you meaning in a post-singularity setting where the amount of labor strictly required for fun maximization is zero. The other is that meaningful labor is fun. (At least, if we stretch the meaning of fun somewhat, to mean "fulfilling".) This offers a blueprint for a post-singularity world of voluntarist labor. And in that model, we may imagine that some people genuinely are most satisfied by a life filled entirely with vapid fun, and so what? Their fun does not diminish mine.
FeepingCreature
11mo ago
If the president calls some other country and tells them something that's classified, and he doesn't know it's classified, I think it still becomes declassified in doing so. At least, that seems to be the argument. So in effect, by taking them home and keeping them past the end of his presidency, Trump declassified the materials without realizing he did so.
I mostly just find that argument amazing and I hope it wins for its own sake as an argument.
FeepingCreature
11mo ago
I mean, it would lead to endless tit-for-tat only as long as supplies of crimes last. I mean, you could make it last a long time by changing laws, but you'd have to put a bunch of additional work in. Absent a new wave of ex-post-facto laws or blatant procedure prosecutions, honestly my first reaction is "yes, good." Let justice reign, etc.
FeepingCreature
11mo ago
If I had two sons, and one son got drunk and punched someone at a bar while another got drunk and was punched by someone at a bar. I would not want to live in a world where the former was killed and the latter killed their assaulter. I'd much rather live in the alternative world where no one died. Which would you rather live in?
I think this is an instance of causal decision theory in the wild, in that you're holding the punch stable when there's no reason to expect that to be the case. What if it being "the sort of world where people who throw punches are killed" means that instead you get to pick between the world where your sons punch and are punched, and the world where nobody is even punched? Then the question would be to what extent punch-kill actually allows acausal flow, right? Ie. we may imagine a world where some people just, out of the blue, are struck by the urge to punch and otherwise-agentically seek out a target to punch. In that case, the kill-branch obviously would only worsen the situation. So the question comes down to if the punch urge is such that the kill branch can successfully shift the incentives enough to suppress the punch branch enough to make up for the QALY loss.
Because at the end of the day, we'd at least somewhat prefer that the least people die. Right?
If non vaccinated, or positives covid tested individuals and families had been shipped off to camps (outside china), and had been killed would that really have been incongruent with the rhetoric and propoganda deployed?
I obviously cannot prove this, but my immediate reaction is "yes, of course, massively incongruent."
We'd need more samples. I was right this time, but obviously n=1.
I think even in China, you could predict fairly reliably if a given camp or campaign was going to engage in mass murder or not, ie. whether the Uighur rhetoric is like the Nazi rhetoric in ways that the Covid rhetoric is not. To be clear, I don't have an opinion on this; I haven't done any research on genocide in China, but I'd expect if there was genocide we should see commonalities in the rhetoric.
edit: Ie. say, nobody was calling Covid victims dangerous parasites.... Okay, I'm not willing to say that. Maybe it's just that the US CW is so hot that the rhetoric on the street was genuinely indistinguishable from Mein Kampf? If so, Scott may be apropos: "Stop telling people they’re going to be killed. ... Stop trying to convince Americans that all the other Americans hate them."
I don't think the epistemic position is the same.
It's not that I, as a pro-lockdown person, abetted and ignored the possible genocide of Covid infectees, it's that I had a very strong positive expectation that there would not be a genocide or even a significant mass murder. (And, you know, I was right.) I don't think that can be said for people who supported the Nazi regime.
FeepingCreature
28d ago
I think Skyrim is a great game, but not for its writing. Try to just play the main quest, not too much of it is actually good, and it's a very short story.
GTA5 is the only game that ever actively convinced me to stop playing it. If the developers had any balls, they'd put the torture mission inside the Steam refund window.
"what we care about"
I mean, that's exactly the problem with definition fights. What we care about is different. That's why there's little sense in attaching so much meaning to terminology, and why you cannot convince people by gesticulating at genes and genitals. When you say "obviously a woman is", and when I say, "well in my opinion a woman is", we use terms that have 99.9% the same coverage, because they almost cleave reality at the joints - which is why the few edge cases are so difficult. In a distribution where almost every property is correlated, it is very hard to see that actually people might be selecting on totally different properties. For instance, since I spend a lot of time online, voice is a dominant criteria for gender for me, and since I'm bi, genitals are a relatively low factor. I don't have the "whatever makes people want to found families", so genes and womb don't factor at all. But you wouldn't see this by looking at what I call "men" and "women", because it's almost entirely the same as everyone else.
edit: In fact, we could probably formalize this into a law: the more dimensions a group correlates, and the smaller the set of exceptions is, the less people will naturally come to agree about group membership of the exceptions.
FeepingCreature
1mo ago
At base my argument is that "men who [choose to pursue and increase their femininity, AKA transwomen]" is legible.
And my point is that this argument, ultimately, only makes sense to you because it begins with your choice of the critical definitional aspects of masculinity. You say "men who" because you consider these attributes of manhood as critical, in which transwomen are masculine. But that is not an argument.
FeepingCreature
2mo ago
I think it might! When I say "humanlike", that's the sort of details I'm trying to capture. Of course, if it is objectively the case that an AI cannot in fact suffer, then there is no moral quandary; however conversely, when it accurately captures the experience of human despair in all its facets, I consider it secondary whether its despair is modelled by a level of a neurochemical transmitter or a 16-bit floating point number. I for one don't feel molecules.
FeepingCreature
4mo ago
Desecrating any of these
Atheist point of order: you cannot desecrate them, because they are not sacred.
The Invisible Pink Unicorn (possibly made of pink-glazed blown glass, in the style of My Little Pony) as the steed bearing the returning Jesus, depicted as a Super-Saiyan, His head and hair burning white, His eyes like a flame of fire, His feet like fine brass
Honestly, I believe many atheists would consider that "fucking awesome".
If you disagree with my criticism of oppression-status granting infinite moral immunity, be specific about what limiting principle you'd propose (if any).
I kind of disagree with this, yes. The limiting factor is having a chance to flourish.
Hypothetical: A guy comes into your house to murder you. He has a gun and spec-ops training; you are a keyboard warrior; he will definitely find you and murder you. The best you can hope for is maybe take him by surprise and give him some bruises. Do you hang out in broad daylight, sheepishly say "guess you caught me" and let yourself be shot? Or do you do the fucker as much damage as you can?
The game theory is this: every decision to exploit somebody exists on a margin spectrum. You are trying to extract as much benefit as possible for a given effort cost; if the other can raise the effort or lower the benefit, it incentivizes you to maybe leave them alone. But we never know where somebody's cut-off point is, so there's always an incentive, if you notice you're being fucked over, to do as much damage as you can back.
So there's a very tentative hypothetical we can construct here to advocate for Palestinean terrorism. Israel is clearly fucking them while exploiting "their" land (whether your game theory implementation advocates forgiveness or revenge here probably depends on preexisting sentiment, but revenge is at least plausible), Israel is clearly trying to minimize effort costs with Gaza, maybe if you can impose some costs on Israel, it'll push them closer to the threshold or at any rate strengthen your negotiating position. In game theory, a person who never plays 'defect' isn't an agent but a resource. Hamas chose the most damaging strategy available to them. Did it break existing compacts? Sure, but I'd presume they assumed that they could not get fucked any worse than they were. Will it work? Probably no.
Okay, cynic hat on: no, but the cost of it not working will not fall on Hamas. IMO, Israel can't really do anything (not hugely expensive) here that will hurt Hamas more than it drives recruitment. From the cynical view, Hamas and the authoritarian movement in Israel are obviously just playing Toxoplasma Tennis. B attacks A'. This enrages A! A cannot fight B, so it attacks B'. This enrages B! B also cannot (cheaply) fight A, so it attacks A', and so on. Part of the reason I don't really have a strong moral view against Hamas is that if this is an accurate model, it's obviously "cooperative" to some extent. Hamas benefits Netanyahu, and conversely. And whenever a cycle like that exists, blaming the most recent hit on whoever committed it is looking at the wrong component. It's a systemic effect. Remove Hamas, another terror group will be found. There is a gap here that allows the existence of a feedback cycle, so a feedback cycle arises. Anyway, in this particular case, the cycle might be running out of control because somebody, A or B, underestimated the damage the current serve would do, so it's unclear what happens next. But my moral view to "let's put the angry people in a cage and then send the guard away" is: a stupid game was played, and a stupid prize was won, I feel bad for the victims but not angry at the perpetrators; it's not like they were the load-bearing causal component.
To loop back: why did I say "the limiting factor is having a chance to flourish?" Well, how do you get out of a cycle like this? You find better things to do with your life. Not sure how good a life you could have in Gaza City. If you could have a good life, a dignified life, a life with authorship and respect, and then you go on a revenge bender - well, I am a humanist, I want to maximize flourishing. When people live an unworthy life, I welcome attempts to, even counterfactually, push for a better life; when people could already live a worthy life, I don't. Do I think Gazans lack the capability to live a worthwhile life? I don't know, honestly, but if I wanted to construct a moral case for terrorism, that's where I'd start.
Addendum: When this conflict started, I said to a family member: "I don't think what Hamas did was right, but I am willing to bet on two things: at the end of this, a lot more Palestineans will have died than Israelis; and at the end of this, Hamas will still be there." If Israel wants to convince me that I'm wrong about the Toxoplasma Tennis thing, those are the two factors they should try to improve.
Is cooperation structurally indistinguishable from submission as well? What about domination?
I mean, yes. There's a bunch of arguments for situations where you should not extract the maximum you can in the short term from a relationship, but they're all founded in maximizing your long-term payoff, not in "being a good person". Even decision theories like TDT/superrationality, where you occasionally leave money on the table, are based on this - in sum, the TDT agent walks away with more utilons than the CDT agent. A decision theory that systematically ended up with less utilons than it could would just be bad.
He who cannot make peace with the thought that his payoff is the smaller one, and makes it even less, burning commons out of pure spite, is irrational; thus, evil.
Utility, being unitless, is not comparable between agents. All theories that allow comparing payoffs do so on the basis of axioms, like pretending that every other agent is a copy of you ("putting yourself in their shoes"), or normalizing all human preferences to a common bound (humanism). Money is arguably also a way to do this. Though all variants of the ultimatum game depend on some way to compare utility between agents to converge, that comparison has to be agreed upon by some other mechanism such as relative capacity to destroy whatever your opponent values. Utilitarianism has no opinion on what the "correct" exchange ratio is. (Though it does advise that you should follow an algorithm to find it that maximizes your payoff. It says that a lot.)
FeepingCreature
9mo ago
If I have a book and 5% of the pages I look at are blank, I'll have the strong expectation that the rest of the pages are blank too. And for the same reason - why would the author leave any pages blank? Energy is energy. If you're bothering to colonize any measurable amount of the universe, you'll colonize the rest too. Any species that ever stopped expanding would stop expanding long before it became globally visible.
FeepingCreature
10mo ago
It's nonsensical, but it's nonsense that screens off the other nonsense. You're basically double-counting criticism. If you grant that the kids return to being regular kids at the age they left, that already gives you all the nonsensicality required for the hazy memories as well.
FeepingCreature
11mo ago
As a European, this is half the reason I'm in favor of us supporting Ukraine.
FeepingCreature
11mo ago
Seems special pleading: why are we requiring no special incantation when declassifying, but are requiring a special incantation ("these are classified" is not enough!) when classifying?
That said, has Trump spoken any incantation for declassifying?
FeepingCreature
11mo ago
Aliens make no sense because the stars still shine. I would not expect the greatest visible evidence for aliens to be on Earth, I would expect it to be humanity surrounded by Dyson spheres. (If I was a civilization that got post-singularity, I would totally eat every sun.) The idea that the strongest visible evidence for alien life is found in Earth's atmosphere simply does not pass any smell test.
FeepingCreature
11mo ago
Sometimes, if a thing is "needed" and violates the constitution, that means you still shouldn't get to have it. What's the point of principles if you only hold to them on matters that are agreeable anyways?
Not at all; it's an implication that you consider other subsections vulnerable to rape, that is, desirable. "Unrapeable" says "not even with zero effort or consequences would she get any".
If you're rating on a spectrum, you get "I would put effort into getting laid with this person" as the higher tier, but then there's a tier of "sure, would fuck if an opportunity arose". That's the "rape" tier; it's not saying you want to engage in rape, but that rape is the obvious-to-come-to-mind situation in which their attractiveness would overcome the thus-lowered effort barrier. If there was a rapist in the room, they would rape this person. They would not rape the lower tier - unrapeable - because it would be actively unenjoyable, net negative even if free. "Thanks, I'd rather masturbate."
A less edgy schoolyard way to phrase the same thing is "would pay to fuck", "would fuck if you paid me" and "not even if you paid me."
FeepingCreature
1mo ago
About halfway through, I completely lost track of what the comment was advocating or even saying.
The thing I don't understand is how you can possibly train for uncertainty.
The model needs to "learn the feeling of not being sure". But whether it's sure or not always depends on its state of knowledge at the time, and that state of knowledge will never be represented in its training set. Additionally and relatedly, you cannot train a LLM to "notice when it's saying something wrong" without indirectly training it to say something wrong, then say it notices.
You would have to inspect the network and somehow determine when it is objectively uncertain, and to what degree, and then synthesize a training task based on that actual uncertainty. That level of interpretability is pretty beyond us at the moment.
More options
Context Copy link