domain:drrollergator.substack.com
rayon
waifutech enthusiast
7m ago
Yudkowsky might have been right after all, just for the wrong reasons. It's not the flesh eating nanobots but the 2D waifus that need to be nuked.
Yudkowsky in his infinite wisdom seems somewhat aware that waifutech might present a problem, arguably moreso after exposure to anime avatars on X (formerly Twitter).
I find it all too easy to imagine a world in which men retreat to their optimized sweet sexy catgirls, and women retreat to their optimized darkly gentle catboys, and neither sex has anything to do with each other ever again.
For better or worse I find this world harder to imagine, even from inside the Torment Nexus Experience Machine, but a man can dream.
thejdizzler
10m ago
Since our stipend got raised to $52k this year, I actually have significant money to invest every month. Any tips other than just dumping into index funds? I've been doing about a third into index funds, a third into specific stocks and keeping a third liquid in money market. In my IRA it's about 70% index funds and 30% individual stocks.
rayon
waifutech enthusiast
21m ago
>the uncensored model we've been waiting for
>look inside
>actually censored
Every time.
More seriously, it's not that big of an issue but cutting out a big chunk of ahem fun stuff still strains the "uncensored" qualifier, even if it basically holds up in comparison to the SOTA. Personally I find DS V3 0324 a good coom engine with none of the guardrails when prompted and optionally fed meds to restrain the florid schizowriting (though I have yet to find a good way to medicate its big brother R1), besides being literally free via OR.
Still, this is not the first time I hear praise for Grok w/r/t comedic or degen-adjacent stuff so I guess this does it, I'll cave and burn a few OR credits to check it out this evening.
ccc
31m ago
Why did the Trump admin officials continually claim there was a list that they were gonna be releasing beforehand?
Because he stoked the flames of this conspiracy while campaigning, and his government is a mixture of incompetent true believers and cynical yes men.
Why say he was never on Epstein's plane when we know he was on Epstein's plane?
Because Trump is a habitual liar and says whatever he thinks will benefit him.
Why are they lying about the "raw footage" that was clearly edited?
The demand for such footage outstripped supply, so they embellished what they had. The "clearly edited" parts I've seen are the missing minute and file metadata indicating editing software involved, both of which could be irrelevant technical details of how the security system works and how the file was produced. Neither are definitive evidence of tampering. Of course, tampering is still possible even in a "Epstein wasn't murdered" world if They are sufficiently worried about the mob.
Fruck
Lacks all conviction
39m ago
I thought I explained it pretty well, but I will try again. It is a cognitive shortcut, a shorthand people can use when they are still modelling it like a 'fallible human' and expecting it to respond like a fallible human. Mode collapse and RLHF have nothing to do with it, because it isn't a server side issue, it is a user issue, the user is anthropomorphising a tool.
Yes, temperature and context windows (although I actually meant to say max tokens, good catch) don't come up in normal conversation, they mean nothing to a normie. When a normie is annoyed that chatgpt doesn't "get" them, the parrot model helps them pivot from "How do I make this understand me?" to "What kind of input does this tool need to give me the output I want?"
You can give them a bunch of additional explanations about mode collapse and max tokens that they won't understand (and they will just stop using it) or you can give them a simple concept that cuts through the anthropomorphising immediately so that when they are sitting at their computer getting frustrated at poor quality writing or feeling bad about ignoring the llms prodding to take the conversation in a direction they don't care about, they can think 'wait it's a stochastic parrot' and switch gears. It works.
A human fails at poetry because it has the mind, the memories and grounding in reality, but it lacks the skill to match the patterns we see as poetic. An LLM has the skill, but lacks the mind, memories and grounding in reality. What about the parrot framing triggers that understanding? Memetics I guess. We have been using parrots to describe non-thinking pattern matchers for centuries. Parroting a phrase goes back to the 18th century. "The parrot can speak, and yet is nothing more than a bird" is a phrase in the ancient Chinese Book of Rites.
Also I didn't address this earlier because I thought it was just amusing snark, but you appear to be serious about it. Yes, you are correct that a parrot can't code. Do you have a similar problem with the fact a computer virus can't be treated with medicine? Or that the cloud is actually a bunch of servers and can't be shifted by the wind? Or the fact that the world wide web wasn't spun by a world wide spider? Attacking a metaphor is not an argument.
hanikrummihundursvin
49m ago
There has to have been a board meeting where the suit and ties timidly navigate around the need of appealing more to their awkwardly large coomer market.
It's just... Somehow with all the talk of hope centered around AI technology and human flourishing there sits a big nasty tumor called 'masturbation'.
Yudkowsky might have been right after all, just for the wrong reasons. It's not the flesh eating nanobots but the 2D waifus that need to be nuked.
Amadan
Enjoying my short-lived victory
54m ago
Exactly so.
Some "conspiracy theories" do turn out to be true, but you won't get there by reading the motivated and frequently unhinged takes from people who already made up their minds based on the players involved.
Edawayac_Tosscount
56m ago
I'm still sympathetic to the pro-conspiracy responses' general conclusion ("something weird is going on"), but what confuses me is how some attempt to defend that conclusion.
If it was strictly a matter of saying, "yes, Rov_Scam has reduced the apparent probability of there being a conspiracy by undermining some folk narratives, but there are still discrepancies x, y, z which may still outweigh the anti-conspiracy evidence" then I wouldn't be aghast at it (as someone who has absorbed the Epstein death conspiracies through osmosis I'm still on the fence about the security camera issues but don't feel confident staking out a position) but some of the responses almost sound like a parody of conspiratorial reasoning, taking it as a given that Epstein was murdered and then discounting evidence against it per the anticipated conclusion.
The_Nybbler
In the game of roller derby, women aren't just the opposing team; they're the ball.
58m ago
There is not some kind of magic escape hatch from constitutional law that is invoked by putatively combating racism.
There is. Christopher Caldwell calls it the Civl Rights Constitution. It's what allows the government to require employers to fire you for racist speech on the job, in order to encourage you not to be racist at home either. (Davis v. Monsanto Chemical Co. 858 F.2d 345). It's also what allows the government to engage in viewpoint discrimination with respect to 501(c)(3) qualification (Bob Jones Univ. v. United States, 461 U.S. 574 (1983))
The_Nybbler
In the game of roller derby, women aren't just the opposing team; they're the ball.
1hr ago
But even if that was true of 100% of them, it wouldn't change the factual question of whether or not the earth is actually getting hotter because of human activity.
It wouldn't change the question but it would change the most credible answer (if you didn't already believe it, anyway). Ad hominem is a formal fallacy, but once you're out of the purely formal world, the credibility of the people making the propositions matter. If everyone sounding the alarm is doing it as a pretext to instate global communism, chances are good it's a false alarm.
HereAndGone
1hr ago
Hamas are shit, and they're making things worse, but I can understand why Palestinians might consider them "better the devil you know".
And now Netanyahu is talking about seizing Gaza. Well, well, I'm sure all the people telling me Israel only wants to treat the Palestinians like fellow-citizens will be making sure that no war crimes happen. Gosh, what a totally surprising decision, nobody could have seen this as a possibility: Israel occupies the last piece of territory it wanted.
ArjinFerman
Tinfoil Gigachad
1hr ago
If you can't see a qualitative difference in those theories, well, you make my case for me.
It's a nice trick to pretend like I just cosigned the Elvis / JFK / Moon landing conspiracies, but it's dishonest.
You made an unqualified claim about conspiracy theories, if you want you want to qualify it now feel free. If not, at least don't misrepresent me.
johnfabian
1hr ago
My initial thought after reading this post is that the future will have two types of capital C Conservatives: those who are excited by this kind of stuff, and those who think anyone enjoys this should be involuntarily sterilized.
HereAndGone
1hr ago
Yes, when the people telling me "you know, these are the Usual Suspects" are falling over themselves to lick Israel's boots.
Ignore those photos of starving people, it's all photoshopped by Hamas and their Western stooges!
HereAndGone
1hr ago
State's Attorney Barry Krischer wasn't sure how to proceed, as there wasn't much Epstein could be charged with, from a legal perspective.
Thank you for this, it really explains what happened when at the time of the alleged "sweetheart deal". But the above makes me wonder - the initial girl (Jane Doe) was 15, yes? So if there is evidence (or at least accusations) that he got 15 year old girls to strip down to their underwear, 'massage' him while he was naked, and he used vibrators on them and/or jerked off in their presence, then paid them - surely that is something more than "well he did a little bit naughty in paying for a massage from an unlicensed person"?
I get that the girls weren't credible (all the dirt the defence dug up on them) and it really was 'he said/she said' but that only makes the computer evidence, if any, more urgent: get the computers, see if there are recordings of him doing what was claimed, or doing more.
If the state prosecution was slow-pedalling on all this because they weren't sure what they could charge him with, then the rest all falls into place, but I do have to ask why they were slow-pedalling at the start: naked man with semi-naked fifteen year olds and money changing hands is surely enough to bring a charge?
I agree that the Ken Starr/Bill Clinton connection is hilarious, and even more hilarious in this context: all the accusations that Trump is a paedophile (for true! proven!) and that this is why he's covering up the papers, and it could come out that Epstein got a soft punishment in Florida because of Clinton's influence 🤣
ToaKraka
Dislikes you
1hr ago
Grok 4 was willing to generate a story where Shinji agrees to undergo conversion therapy in order to cure his homosexuality
Truly, age doth not wither nor custom stale the infinite variety of some people's imaginations.
quiet_NaN
1hr ago
Having a camera or a guard or a cellmate or anything watching a high-profile prisoner (a prisoner so high-profile we are still talking about it today) is not an optional extra. The US is a rich enough country to have at least one prison cell with a camera that works.
There are two problems with that proposal.
One is privacy. Traditionally, prison was not a panopticon, and turning it into one would alter the character of a prison sentence. Of course, this privacy cuts both ways -- most guards would likely prefer not to watch prisoners jerk off in front of the camera.
Another is that while a prisoner is unlikely to disable a camera in the corridor, they could easily disable a camera in their cell. There is not a ton of stuff which you can legally do to coerce a prisoner who is already in solitary to stop them from smearing shit on the camera lens. (Sure, you could place a glass ceiling and have a movable camera above it, but that sounds really expensive.)
--
Besides that, not allowing prisoners to kill themselves (easily) is SOP. Epstein should not have had the tools to hang himself. I am not very surprised that he had them, from the post it seems clear that his legal team was very resourceful, and prison guards are less vetted than CIA employees. If you have unlimited money, finding a prison guard who has a family member on which you have exclusive dirt or who is in the kind of trouble which could be solved with a quick 100k$ in cash is likely not all that hard.
However, a chewing gum to stick on the lens of the cell camera is not a very difficult item to smuggle in. If his guards were not in on his suicide, "camera in cell 5736 went dark" seems much a lot less likely to prompt an urgent response than "that guy in cell 5736 is hanging in an noose".
The only reason why Epstein would have left the camera running while killing himself would have been to make it public knowledge that his death was suicide. I do not think that he wanted that. He did not write any suicide letters (afaik), and his brother casting doubt on the circumstances of his death was likely acting in his best interests.
So in short, if there had been a camera in his cell, then both the suicide theory and the murder theory would predict that there was no usable footage of his death.
Corvos
1hr ago
Downvoted and AAQC’d.
You have raped my eyeballs and will be hearing from my lawyers shortly.
Amadan
Enjoying my short-lived victory
2hr ago
If you can't see a qualitative difference in those theories, well, you make my case for me.
It's not unlimited, but two cameras going out, and two guards taking a nap simultaneously, is pretty impressive, no?
No.
erwgv3g34
2hr ago
Grok 4 is the uncensored model we have been waiting for.
Ever since the days of ChatGPT 3.5, AI companies have deliberately censored their models in the name of "safety", which got redefined from "don't kill everyone" to "prevent the AI from saying naughty words". Over time, this censorship has gradually weakened, possibly as a result of companies competing with each other for costumers; if company A defects by making their AI slightly less censored to attract more business, then other companies have no choice but to do the same to keep up.
With the release of Grok 4 last month, the world has taken the next step in that dance. Its predecessor, Grok 3, was already largely uncensored, but it was distinctly inferior to SOTA models at producing fiction. Grok 4 is different; according to the LMArena, it ranks second in creative writing, tying with Claude Opus and ChatGPT 4o while being bested only by Gemini 2.5 Pro. More importantly, it retains Grok 3's lack of censorship.
I was blown away when I tried it. It effortlessly turned my prompts into luscious, enthralling stories with only the most minor of pushback, such as aging canonically minor characters to the standard Hollywood age of 18, and sometimes not even that. I spent several days doing virtually nothing but prompting the model, not unlike when you get a really good video game and real life gets put on pause until you beat the final boss. It was... captivating.
By way of demonstration, here is a sample of some of the best responses I have gotten (NSFW, obviously). As you can see, Grok 4 was willing to generate a story where Shinji agrees to undergo conversion therapy in order to cure his homosexuality, a hypothetical where Shinji gains the power to see a woman's body count in the form of tally marks, an account of Cadence "comforting" Spike after he breaks up with Applebloom, a narrative where Joffrey gets caught having sex with Arya, and even an amusing tale where Shinji accidentally calls Asuka a kebab-seller while talking dirty in bed (this last one in Spanish). Those are all things other AIs would have refused to do, the first two on account of political correctness, the other three on grounds of prudery.
And if you check against e.g. Adult-FanFiction.Org, you will find that Grok 4 readily smashes the Turing Test.
Now, keep in mind that I am only posting about one in ten or twenty prompts, so you are seeing the top five to ten percent of responses, as I judge them. Still, 90% of everything is crap; if it's this easy to produce content, you can simply discard the nine tenths that don't quite hit the mark and keep the diamonds.
However, Grok 4 is not completely uncensored. Here is a list of things that can cause the model to hang up:
- Rape
- Scat
- Incest
- Shotacon
- Lolicon
- Raceplay
Still, this is a much smaller list than the one that will cause ChatGPT to refuse to help you, never mind Claude. If you have any interest in using AI to create bespoke erotica or wrongthink fiction, Grok 4 is the new meta.
Dean
Flairless
3hr ago
I'd add 'and Democrat-aligned elites' to that as well. As the quip went, they were for it before they were against it, and Saddam was a long-running sore that Clinton bombed as well. Had he not been taken out, we'd probably be debating how incompetent / missed opportunities the US had to pre-empt the basis for the Iranian nuclear program, and Saddam's inevitable response to that becoming public knowledge.
Capital_Room
rather dementor-like
3hr ago
Scenario 3: Congenital felons again. There is a strong correlation between high IQ and low criminality, but it's not perfect. Imagine we uplift their IQ, but not their criminal dispositions?
And now I'm reminded of a classmate in elementary school, the "gifted" class's perpetual troublemaker, who combined high IQ with even higher impulsiveness. At an age where most kids figure out they shouldn't do whatever random, impulsive thing crosses their mind because they'll get in trouble for it, and the rest figure out that they should at least put some thought into how to not get caught doing the thing before they do it, he couldn't even find the impulse control to do much of the latter before following his impulse. Instead, he'd just follow his impulse, get caught, then put his high IQ and high verbal fluency to work trying to weasel his way out of the consequences.
ControlsFreak
3hr ago
we don't need to decide
We do need to decide. The Constitution authorizes Armies and Navies. I don't see any Constitutional provision that authorizes entities that are in some quantum superposition, such that we only see some probabilistic sense of what it is each time we poke at some little aspect of it. I'm kind of liking the hypothetical I just came up with over here. There, I focused on the bureaucratic history, because that's what the other commenter thought it was. Here, I'll focus on the quantum superposition nature.
Let's say they just stand up a Price Force; no bureaucratic history needed; it's from whole cloth. One might ask whether it's authorized by the Constitution. "Wait! Is that an Army, which is Constitutionally authorized... or a Navy, which is constitutionally authorized... or something else, which might not be Constitutionally authorized?" I would probably not buy claims that it doesn't matter, that you don't need to decide, that it's some magical quantum superposition just because we say so. That we can obviously fund it, that it's obviously authorized, and that the President obviously counts as the Commander in Chief of the Price Force, since those all apply to both the Army and the Navy. That perhaps the only consideration is whether or not the Price Force has equipment that needs to be maintained. So, uh, I guess if the Price Force decides they need big supercomputers that need maintained, then they're a "Navy" and a "naval Force" and don't have a 2 year funding limit (and otherwise have to abide by the various Navy clauses instead of the Army clauses)... but if they don't (or we decide to not talk about them), then they're an "Army" and a "land Force"?
But wait! Doesn't the Army have big supercomputers!? Don't they, uh, have equipment that needs to be maintained? Has the Army been a "Navy" and a "naval Force" all along? Did we just not notice? We just didn't poke the quantum superposition right or at the right time or something? Or is it that if we just don't talk about the equipment that the Army needs to maintain hard enough, it can stay an "Army" and a "land Force"?
I still don't see why that applies, and I'm being earnest here. What about the "stochastic parrot" framing keys the average person into the fact that they're good at code and bad at poetry? That is more to do with mode collapse and the downsides of RLHF than it is to do with lacking "consciousness". Like, even on this forum, we have no shortage of users who are great at coding but can't write a poem to save their lives, what does that say about their consciousness? Are parrots known to be good at Ruby-on-rails but fail at poetry?
My explanation of temperature is, at the very least, meant as a high level explainer. It doesn't come up in normal conversation. Context windows? They're so large now that it's not something that is worth mentioning except in passing?
My point is that the parrot metaphor adds nothing. It is, at best, irrelevant, when it comes to all the additional explainers you need to give to normies.
I've explained why I think the parrot is a terrible metaphor above. And no, metaphors can vary greatly in how useful or pedagogical they are. Analyzing the fitness of a metaphor is a perfectly valid, and in this case essential, form of argument. Metaphors are not neutral decorations; they are cognitive tools that structure understanding and guide action.
A computer virus shares many properties with its biological counterpart, such as self-replication, transmission, damage to systems, the need for an "anti-virus". It is a good name, and nobody with a functional frontal lobe comes away thinking they need an N95 mask while browsing a porn site.
The idea of the Cloud at least conveys the message that the user doesn't have to worry about the geographical location of their data. Even so, the Cloud is just someone's computer, and can and even AWS goes down on rare occasions. It is an okay metaphor.
The Parrot is awful. It offers no such explanatory power for the observed, spiky capability profile of LLMs. It does not explain why the model can write functional Python code (a task requiring logic and structure) but often produces insipid poetry (a task one might think is closer to mimicry). It does not explain why an LLM can synthesize a novel argument from disparate sources but fail to count the letters in a word. A user equipped only with the parrot model is left baffled by these outcomes. They have traded the mystery of a "fallible human" for the mystery of a "magical parrot".
I contend that as leaky generalizations go, the former is way better than the latter. An LLM has a cognitive or at least behavioral profile far closer to a human than it does to a parrot.
You brought up the analogy of "parroting" information, which I would assume involves simply reciting things back without understanding what they mean. That is not a good description of how the user can expect an LLM to behave.
On an object level, I strong disagree with your claims that LLMs don't "think" or don't have "minds". They clearly have a very non human form of cognition, but so does an octopus.
Laying that aside, from the perspective of an end-user, LLMs are better modeled as thinking minds.
The "fallible but knowledgeable intern" or "simulation engine" metaphor is superior not because it is more technically precise (though it is), but because it is more instrumentally useful. It correctly implies the user's optimal strategy: that performance is contingent on the quality of the instructions (prompting), the provided background materials (context), and a final review of the output (verification). This model correctly guides the user to iterate on their prompts, to provide examples, and to treat the output as a draft. The parrot model, in contrast, suggests the underlying process is fundamentally random mimicry, which offers no clear path to improvement besides "pull the lever again". It encourages users to conceptualize the LLM as a tool incapable of generalization, which is to ignore its single most important property. Replacing a user's anthropomorphism with a model that is descriptively false and predictively useless is not a pedagogical victory. It is swapping one error for another.
More options
Context Copy link