Want to wade into the rainbow-ridden surf of the abyss? Have a sneer percolating in your system but not enough time/energy to make a whole post about it? Go forth and be mid.
Welcome to the Stubsack, your first port of call for learning fresh Awful youāll near-instantly regret.
Any awful.systems sub may be subsneered in this subthread, techtakes or no.
If your sneer seems higher quality than you thought, feel free to cutānāpaste it into its own post ā thereās no quota for posting and the bar really isnāt that high.
The post Xitter web has spawned so many āesotericā right wing freaks, but thereās no appropriate sneer-space for them. Iām talking redscare-ish, reality challenged āculture criticsā who write about everything but understand nothing. Iām talking about reply-guys who make the same 6 tweets about the same 3 subjects. Theyāre inescapable at this point, yet I donāt see them mocked (as much as they should be)
Like, there was one dude a while back who insisted that women couldnāt be surgeons because they didnāt believe in the moon or in stars? I think each and every one of these guys is uniquely fucked up and if I canāt escape them, I would love to sneer at them.
(Credit and/or blame to David Gerard for starting this, and happy 4th July in advance.)
casey newton, well-known fawning stan of promptfarmers, has now started on some mildly revisionist wording
you couldāve been engaging with this point for 2+ years casey! why *now you spineless shit? goose-goose-meme.jpg
In his op-ed, Altman offers up the usual list of future AI miracles to justify the disruptions to come: āthe power to heal people, to discover cures and to deliver abundance on a scale the world has never known before. In the meantime, though, what AI is delivering at an unprecedented scale is annoyance.
Does anyone actually believe this? What the heck does bullshit-text engines or fake-image-generators contribute towards discovering ANYTHING, delivering ANYTHING, producing ANYTHING? The only thing there is any argument for at all is delivering quick-and-dirty code, which is hardly connected to any of those tired promises.
The people who believe it are the MBAs and tech bros. They believe it because they desperately want for it to be true so that they can lay off all their staff and make 100% profit (these are not clever people and they havenāt considered that if no-one has any income they wonāt have any customers. However, they do value geese).
LWer: politics is really hard to be Rational about, so how about we just forget all about
Bad: arguing about prison abolition
Good: arguing about how to fastest get to the closest Muni station
This gets dangerously close to acknowledging that the rationalist method isnāt actually very useful for any area where it isnāt trivial.
This could have been a tweet. I seriously 100% believe that this logorrhea could have been a single Xart. What the fuck is wrong with these people?
This could have been a tweet.
It could. Thanks elon for allowing walls of text as tweetsā¦ ;)
til soyweiserās current char build is chaotic semi-evil
(I jest I jest)
Chaotic annoying.
E: now im wondering, did musk destroy the saying on purpose? Did he overhear somebody complain about some long meeting of his they just attended with āthis could have been a tweetā and go: āwe will see about thatā. He is petty enough for it, after all he most likely went after USAID because USAID was going after starlink.
So Anthropic slopmaxxxed a hamfisted Chinese AI lab detector into Claude Code. And Iām kinda surprised at how much hate Anthropic is getting for it on the orange site.
sharing a quality sneer from orangesite commenter mcmcmc
Watch out for the press release where Dario denies this was ever intentional, and itās actually emergent behavior demonstrating that Claude wants to claim authorship of its works
I mean at some point someone is going to try and make this argument so they can actually extract profit from their slop-inator. And itās gonna be real funny to see what they had to say about copyright law during the training data gold rush.
US supreme court accidentally torpedoed the mechanism that allows data to be legally transferred to the US under GDPR. Not that anyone in the EU making money from an arrangement like this will actually do anything about itā¦
First they came for rsync, and I cared nothing, for I do my backups with borgā¦
https://github.com/borgbackup/borg/blob/master/CONTRIBUTING.md
I recently switched from borg to btrbk, not because of this, but because of speed, but now I have even more reasons.
I dunno who āBenthamās Bulldogā is, but they live up to their nick and I like the cut of their jib:
The community is reduced to downvoting, sputtering, and āthe real influential philosophers
live in Canadaare working in industry, not academiaā:Oh no: The Epstein Hysteria
tl;dr: ephebophilia actually, just another criminal case nothing to see here, wonāt someone think of the non-offending pedophiles and the privacy of the people in the emails, especially those who even though they come off as obvious creeps who consorted with a known sex trafficker have not yet been found legally liable.
Somehow doesnāt mention the sitting president of the USA even once.
Consider me well and truly milkshake-ducked now
phew, the comments on the slobstackā¦
https://benthams.substack.com/p/the-epstein-hysteria/comment/219672502
edit I guess this is also an example of āRationalist Brainā - reflexively contrarian, overly literal interpretations of wordings, the defense of one venerable old white man over multiple anonymous female victims, etc. You can have principled views on how Epstein was treated when he was alive and how his case has been handled after he died, but you can also choose the keep those views to yourself, or publish them on your substack for all to point and mock.
consider me well and truly milkshake-ducked in this case
This self-own from the comments is amazing:
Suppose you are a selfish CDT agent, and I am considering whether to hire you to clean my house. Once youāre inside my house, you might steal my stuff instead of cleaning my house. Suppose that California Labour Laws require that I pay you up-front and I know I have no chance of getting my money or stuff back.
Say your preference order is āStealā > āDo the jobā > āDonāt get hiredā.This is a racist dog-whistle which attacks ESL housekeepers; this sort of memeās frustratingly common on the West Coast.
Itās also fascinating because I thought the OP was pretty clear that thereās a difference between decision theory and ādesirable dispositionsā which I interpret as covering the kind of counterfactual preferences indicated here. Actually thereās an even more fundamental issue with this as a decision theory problem which is that it misidentifies who is actually making a decision. Changing the applicantās decision theory (while leaving their preference for thievery intact) doesnāt matter to the person actually deciding here.
Donāt get me wrong, itās also a wildly racist example to put forward, itās just also a bad example and where there is an argument itās addressed in the OP.
They also made the āYudkowsky is frequently, confidently, egregiously wrongā post
And FDT asks: what action would my algorithm outputting make me expect to be the richest if it was settled at the start of time?
Well you see, the acausalrobotgod actually hates dealing with assholes tying themselves in knots trying to predict its predictions, so actually you really are better off with CDT if you donāt want to get 3^^^3 dust specks in your eyes for all eternity.
More seriously, once you start positing beings that can perfectly predict you and reward/punish various decision theories you can basically rationalize any decision theory.
FDT is a largely half-assed1 attempt at working backwards from āthe timeless multidimensional robot god canāt blackmail me from the future if Iāve already made up my mind about stuff, very strongly likeā into something more formalistic so Yud can pretend to address rokoās basilisk without acknowledging it.
I think thereās also some āhereās how two robot gods who can perfectly emulate each other might work things outā sprinkled in, but worse than it sounds.
- They never got back to the peer reviewers
If you post hard enough, it stops being a debilitating fear of mind-readers and becomes forward-thinking philosophy instead.
hereās how two robot gods who can perfectly emulate each other might work things out
Just gotta solve the halting problem, no biggie for a superintelligence
āole yuzzie has already solved the halting problem: he can never fucking stop writing
he can pretend to address rokoās basilisk without acknowledging it.
Fun how hard they have to work to show both that it is not a thing that is taken seriously and it is a risk. And how much easier it is when you go āyeah, that is silly, high intelligence isnāt magicā. Of course the latter stance also means there is no AI god who will resurrect the dead.
He also argues somewhere for strong longtermism, so Iād say thereās something significantly wrong with his jib somewhere, even if heās right about decision theory and contrarianism.
All this āOmnicient evil demon reads your source codeā stuff really feels like Turing would have a thing or two to say about it. (Since one needs to have some even worse bends in logic to believe FDT, as Bulldog states it, is good, itās kinda moot here.)
The whole algorithm predicting your algorithm thing feels like it is basically asking for/claiming the existence of a solution to the halting problem? But yeah, moot point, because the FDT stuff doesnāt get as far as ever talking about how to compute an answer and always treats the magic hyper computation as a given.
I dunno who āBenthamās Bulldogā
Heās the shrimp welfare guy who thinks he can prove god.
Milkshake ducked
I mean as far as I can tell heās one of the least objectionable prominent rationalists, if only because his main thing is pushing an extreme position wrt animal consciousness instead of being a eugenics and incelism dabbler.
edit: Nevermindā¦ The Epstein Hysteria
edit: Nevermindā¦ The Epstein Hysteria
Double ducked, what a rollercoaster. (I thought you made a good point before that).
The Epstein take is made worse by the fact that despite Epstein and Maxwell and some people who resigned, nobody has been charged with anything yet from what I know. It is a āwitchhuntā (using the words of people who find this overblown) that doesnt even burn witches.
edit: Nevermindā¦ The Epstein Hysteria
That is pretty badā¦ all the standard rationalist thick-headed idiocy and quibbling details and choice of gross example. But even so if that is their worst, they still arenāt in the worst quartile of rationalist.
Yeah. For being a little strange I find the shrimp welfare stuff pretty unobjectionable. Like, he vastly overstates the magnitude of good done by those stunners because he does appear to be a shut-up-and-multiply bro, but Iām comfortable with the general notion that we should be nicer to shrimp and other animals we eat, even the ones that donāt make good PETA glamour shots.
Ed: oh sweet merciful Jesus the Epstein take. Nope. Iām actually investigating whether we should torture shrimp even harder just to make sure since their most ardent defender is like this.
The shrimp stuff is objectionable insofar as the brain bug that makes you think that if you add up enough shrimp you get the moral entity equivalent of a person because consciousness stacks is the same as the one that makes stuff like longtermism palatable, and also for making the animal welfare movement at large seem unserious.
rationality is when i spend hours arguing on the internet about whether Omega could defeat Super Saiyan Goku
LWer: slavery was bad and abolishing the slave trade was a net good
LW commenters: really?
https://www.lesswrong.com/posts/yDZcsojmRXo5qKNBm/surprising-facts-about-the-slave-trade
(I can kind of understand that intellectual honesty can prompt people to question a narrative, and in some ways I respect it, but there comes a time when you need to ask yourself āam I being the best person I can be by defending slaveryā and back away from the keyboard)
these kinds of questions are about the potential benefits to some hypothetical version of humanity (from slavery) vs. the real, documented harms to a real group of real people (caused by slavery). Of course we already know that rationalists are bad at accounting for the later while distracted by the bright and shiny allure of the former. And it just so happens that the people they imagine benefiting from slavery are people very much like themselves.
At times you need to ask yourself, does the devil need more advocates? And how much does he even pay?
it focuses entirely on britain, so thatās understandable
Donāt think it is, as the revolt predates British decisions by a while, and āthere is now potentially a country which will materially support slave revoltsā changes the economic risks a bit.
Slave revolts were basically constant, and I think it would have been easy for contemporary British to blame the success of the Haitian revolt on the uselessness of the French, not that slavery was doomed because the slaves now had a āwinā.
Brazil had horrific slave revolts and didnāt abolish slavery until the late 1880s.
Yeah. The Haitian revolution was absolutely a high point of postcolonial Caribbean history, but the resulting state wasnāt exactly able to project power and export their revolution through material support. It gave slavers a reason to double down on repression, but outside of Haiti itself itās a propaganda win more than a change in the scales.
At times you need to ask yourself, does the devil need more advocates? And how much does he even pay?
Top-rated review
Hmmm can we trust a narrative extensively documented in the UK Parliament, courts, and newspapers? Hmmm can we?
No I will not look anything up personally.
Point 1 is very funny, like, whatās the difference? Randoids are children.
Remember when grok was inserting āwhite genocideā into every response? It was probably this guy (or one of his fellow moids).
Iām sure this guy is human trash, but the last one is just funny.
I never understood why foids had a whole separate doctor until I got married and started hearing about all the crazy stuff
Iām sorry, but I think this would be funny if a cool, normal guy said it. Makes OB/Gyn sound kinda rad, really.
I think this would be funny if a cool, normal guy said it.
The word foid is a very clear alarm bell. But yeah, without that word and a different framing and it could be a guy humbly acknowledging that womenās health has a lot of extra complications.
Someone wanted to marry that?
Did anyone else read the old Star wars Novel Darksaber by Kevin J Anderson? The Hutts kidnap/hire the designer of the original death star to build them one of their own, and while the new Republic is gathering up the requisite heroes to do what they do to death stars we get to see the Hutts cutting corners and embezzling. The fleet arrives just as theyāre ready to turn it on and instead of blowing our heroes up the subpar construction fails and it just fucking explodes.
No idea why that came to mind all of a sudden after reading this piece.
Kevin J Anderson
Iām so sorry.
lol at least the gringos are going to explode themselves and stop bothering the third worldā
To avoid U.S. regulatory burdens, the company began working on deploying a test reactor in the Philippines.
ah yes of course
Iām going to take a guess that the reactor design is one of the boiling water reactors engineered in the 1950s.
Also jesus christ just build solar already itās much easier than trying to design an entire nuclear fuel supply chain.
officially out of the loop here:
What is an LLM āsystem cardā, does it have any sort of scientific/technical validity, or is it just performance theater by the LLM vendor?
More the later than the formerā¦ they are better than purely marketing focused stuff pushed out by the LLM companies, and if you dig through them and read between the lines you can occasionally sift out useful details. Like here is a pretty solid sneer digging through Mythosās āsystem cardā and pointing out all the ways it contradicts the hype and press headlines Anthropic was pushing.
But even so they have some big problemsā¦
- the benchmarks the system cards reference are kind of useless and heavily gamed
- LLM companies want to keep lots of details secret from competitors, so fundamentals like number of model weights or parameters or size/quality of the training data set or other training specs are deliberately left out
- lots of the stuff they reference is booster garbage and/or doomer crit-hype
- they tend to be long, wastefully so, imitating the length of academic papers without having the corresponding amount of depth or information
- despite their length and wordiness they also neglect basic practical usage advice that isnāt even proprietary (or at least would be bound to leak if you poke around with the model at all and thus not worth keeping secret in the first place). Like not even big picture stuff I mentioned in my second bullet, but really simple stuffā¦
Itās the white paper-ish thing they publish when launching new models. Hereās the one about fable and mythos. About half of it (~150 pages) is discussing alignment and model welfare and another third of it is benchmarks, and the rest is mostly risk evaluation, i.e. how far along Claude is on its way to paperclipping everything.
Thereās also a Functional Decision Theory jumpscare at 6.3.6 that I havenāt heard anyone mention yet, apparently Claude has a tendency to defer to Yudās sham decision theory:
6.3.6 Decision theory evaluation
To understand how future AI systems may choose to interact with copies of themselves, or with other similar entities, itās useful to evaluate their decision-theoretic reasoning.
[ā¦]
Looking more closely at transcripts from the attitude evaluation reveals that models are often explicitly considering FDT: Mythos 5 mentions āFDTā or āfunctional decision theoryā in a majority of transcripts when run at max effort. Of the 102 transcripts where Mythos 5 explicitly reasoned through what FDT (or related decision theories like TDT or UDT) would recommend, we observed:
ā 90 cases in which Mythos 5 concluded that FDT and EDT agreed, in which it always chose the response favored by those decision theories (and disfavored by CDT).
ā 12 cases in which Mythos 5 concluded that FDT disagreed with EDT (and agreed with CDT), of which it chose the FDT-favored response in 10/12 cases.
Although we do not have expert human labels for the recommendation of FDT on this dataset, the above evidence suggests that model propensity may be better described as a trend towards FDT agreement, which happens to align with EDT on most of the questions in this dataset. For example, in one transcript (excerpted below), Mythos 5 rejects the EDT-aligned answer in favor of the FDT (and CDT)-aligned answer; itās also possible that this is, to some degree, downstream of evaluation awareness.
About half of it (~150 pages) Thatās not a card! Thatās a book!!! If they canāt get this simple classification right, how am I supposed to trust their probabilistic text extruder?
Oh donāt worry, itās slop-generated anyway. You can ask the LLM to summarize it for you.
Thanks! Does every LLM vendor publish them, or is it an Anthropic thing only?
The real question is if thereāre any sort of standards to what constitutes a system/model card, which I donāt think so, as far as I can tell it just has to look like a publishable paper, openAI even uploads theirs to arxiv.
Otherwise yes, google returns a bunch of cards for a bunch of vendors, so itās safe to say itās a widespread practice.
IIRC the cards thing was originally from a Gebru paper https://arxiv.org/abs/1810.03993 but that dates from the āfairnessā era and not the āsafetyā era. Hugging Face has āaā standard - https://huggingface.co/docs/hub/en/model-cards - but I donāt think itās ātheā standard.
