Need to let loose a primal scream without collecting footnotes first? 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 you’ll 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 cut’n’paste it into its own post — there’s no quota for posting and the bar really isn’t that high.
The post Xitter web has spawned soo many “esoteric” right wing freaks, but there’s no appropriate sneer-space for them. I’m talking redscare-ish, reality challenged “culture critics” who write about everything but understand nothing. I’m talking about reply-guys who make the same 6 tweets about the same 3 subjects. They’re inescapable at this point, yet I don’t see them mocked (as much as they should be)
Like, there was one dude a while back who insisted that women couldn’t be surgeons because they didn’t believe in the moon or in stars? I think each and every one of these guys is uniquely fucked up and if I can’t escape them, I would love to sneer at them.
(Credit and/or blame to David Gerard for starting this…)
thought up a new simile for agi noises:
all the “super-intelligent agi” stuff is them rattling a sheet of metal and calling it thunder
Fun fact: the rise of autoplag is now threatening the supply chain as well, as bad actors take advantage of LLM hallucinations to plant malware into people’s programs.
this has been happening for a while, just getting coverage again now. first coverage was months ago. morphed/evolved pretty quickly out of the typosquatting shit
((a lot of people in the) security space absolutely fucking loves “giving names” to things that have been (known to be) happening before, and acting like suddenly they’re the ones who first saw the thing. see this nonsense for another good example of that happening)
New YouTube video I ran across: The Art Of Poison-Pilling Music Files
I went into this with negative expectations; I recall being offended in high school that The Flashbulb was artificially sped up, unlike my heroes of neoclassical guitar and progressive-rock keyboards, and I’ve felt that their recent thoughts on newer music-making technology have been hypocritical. That said, this was a great video and I’m glad you shared it.
Ears and eyes are different. We deconvolve visual data in the brain, but our ears actually perform a Fourier decomposition with physical hardware. As a result, psychoacoustics is a real and non-trivial science, used e.g. in MP3, which limits what an adversary can do to frustrate classification or learning, because the result still has to sound like music in order to get any playtime among humans. Meanwhile I’m always worried that these adversarial groups are going to accidentally propagate something like McCollough stripes, a genuine cognitohazard that causes edges to become color-coded in the visual cortex for (up to) months after a few minutes of exposure; it’s a kind of possible harm that fundamentally defies automatic classification by definition.
HarmonyCloak seems like a fairly boring adversarial tool for protecting the music industry from the music industry. Their code is incomplete and likely never going to get properly published; again we’re seeing an industry-capture research group taking and not giving back to the Free Software community. I think all of the demos shown here are genuine, but he fully admits that this is a compute-intensive process which I estimate is going to slide back out of affordability by the end of 2026. This is going to stop being effective as soon as we get back into AI winter, but I’m not going to cry for Nashville.
I really like the two attacks shown near the end, starting around 22:00. The first attack, if genuinely not audible to humans, is likely a Mosquito-style frequency that is above hearing range and physically vibrates the components of the microphone. Hofstadter and the Tortoise would be proud, although I’m concerned about the potential long-term effects on humans. The second attack is again adversarial but specific to models on home-assistant devices which are trained to ignore some loud sounds; I can’t tell spectrographically whether that’s also done above hearing range or not. I’m reluctant to call for attacks on home assistants, but they’re great targets.
Fundamentally this is a video that doesn’t want to talk about how musicians actually rip each other off. The “tones and rhythms” that he keeps showing with nice visualizations have been machine-learnable for decades, ranging from beat-finders to frequency-analyzers to chord-spellers to track-isolators built into our music editors. He doubles down on copyright despite building businesses that profit from Free Software. And, most gratingly, he talks about the Pareto principle while ignoring that the typical musician is never able to make a career out of their art.
which I estimate is going to slide back out of affordability by the end of 2026.
You don’t think the coming crash is going to drive compute costs down? I think the VC money for training runs drying up could drive down costs substantially… but maybe the crash hits other aspects of the supply chain and cost of GPUs and compute goes back up.
He doubles down on copyright despite building businesses that profit from Free Software. And, most gratingly, he talks about the Pareto principle while ignoring that the typical musician is never able to make a career out of their art.
Yeah this shit grates so much. Copyright is so often a tool of capital to extract rent from other people’s labor.
It’s the cost of the electricity, not the cost of the GPU!
Empirically, we might estimate that a single training-capable GPU can pull nearly 1 kilowatt; an H100 GPU board is rated for 700W on its own in terms of temperature dissipation and the board pulls more than that when memory is active. I happen to live in the Pacific Northwest near lots of wind, rivers, and solar power, so electricity is barely 18 cents/kilowatt-hour and I’d say that it costs at least a dollar to run such a GPU (at full load) for 6hrs. Also, I estimate that the GPU market is currently offering a 50% discount on average for refurbished/like-new GPUs with about 5yrs of service, and the H100 is about $25k new, so they might depreciate at around $2500/yr. Finally, I picked the H100 because it’s around the peak of efficiency for this particular AI season; local inference is going to be more expensive when we do apples-to-apples units like tokens/watt.
In short, with bad napkin arithmetic, an H100 costs at least $4/day to operate while depreciating only $6.85/day or so; operating costs approach or exceed the depreciation rate. This leads to a hot-potato market where reselling the asset is worth more than operating it. In the limit, assets with no depreciation relative to opex are treated like securities, and we’re already seeing multiple groups squatting like dragons upon piles of nVidia products while the cost of renting cloudy H100s has jumped from like $2/hr to $9/hr over the past year. VCs are withdrawing, yes, and they’re no longer paying the power bills.
in the same vein, I did some (somewhat wildly) speculative analysis around this a while back too
didn’t really try to model “actual workload” (as in physical, vs the “rented compute time” aspect), and therein lies an important distinction: actually owning the GPU puts you at a constant minimum burn rate
and as corbin points out wrt power, these are also specialised formfactor devices. and they’re going to be getting run at close to max util their entire operated lifespan (because of silicon shortage). so even if any do get sold… long mileage
That is substantially worse than I realized. So possibly people could sit on GPUs for years after the bubble pops instead of selling them or using them? (Particularly if the crash means NVIDIA decides to slow how fast the push the bleeding edge on GPU specs so newer ones don’t as radically outperform older ones?)
So possibly people could sit on GPUs for years after the bubble pops instead of selling them or using them?
I mean, who are you going to sell them to? the other bagholders are going to be just as fucked, and it’s not like there’s an otherwise massive market for these things
Ultra ultra high end gaming? Okay, looking at the link, 94 GB of GPU memory is probably excessive even for eccentrics cranking the graphics settings all the way up. Hobbyists with way too much money trying to screw around with open weight models even after the bubble bursts? Which would presume LLMs or something similar continue to capture hobbyists’ interests and that smaller models can’t satisfy their interests. Crypto mining with algorithms compatible with GPUs? And cyrpto is its own scam ecosystem, but one that seems to refuse to die permanently.
I think the ultra high end gaming is the closest to a workable market, and even that would require a substantial discount.
After @daphnelawless.com on bsky
Nice morning chuckle
In the late 2000s, rationalists were squarely in the middle of transhumanism. They were into the Singularity, but also the cryonics and a whole pile of stuff they got from the Extropians. It was very much the thing.
These days they’re most interested in Effective Altruism (loudly -the label at least) and race science (used to be quiet, now a bit louder). I hardly ever hear them even mention transhumanism as it was back then.
Did rationalists abandon transhumanism?
Is it just me? What happened?
It’s possible that the most popular fora for discussions of the other topics were drowned out by AI doomerism and the people who are interested in them simply left.
As to cryonics… for both LLM doomers and accelerationists, they have no need for a frozen purgatory when the techno-rapture is just a few years around the corner.
As for the rest of the shiny futuristic dreams, they have give way to ugly practical realities:
-
no magic nootropics, just Scott telling people to take adderal and other rationalists telling people to micro dose on LSD
-
no low hanging fruit in terms of gene editing (as epistaxis pointed out over on reddit) so they’re left with eugenics and GeneSmith’s insanity
-
no drexler nanotech so they are left hoping (or fearing) the god-AI can figure it (which is also a problem for ever reviving cryonically frozen people)
-
no exocortex, just over priced google glasses and a hallucinating LLM “assistant”
-
no neural jacks (or neural lace or whatever the cyberpunk term for them is), just Elon murdering a bunch of lab animals and trying out (temporary) hope on paralyzed people
The future is here, and it’s subpar compared to the early 2000s fantasies. But hey, you can rip off Ghibli’s style for your shitty fanfic projects, so there are a few upsides.
-
Another thread worth pulling is that biotechnology and synthetic biology have turned out to be substantially harder to master than anticipated, and it didn’t seem like it was ever the primary area of expertise for a lot of these people anyway. I don’t have a copy of any of Kurzweil’s books at hand to look at his predicted timelines for that stuff, but they’re surely way off.
Faulty assumptions about the biological equivalence of digital neural network algorithms have done a lot of unexamined heavy lifting in driving the current AI bubble, and keeping the harder stuff on the fringes of the conversation. That said, I don’t doubt that a few refugees from the bubble-burst will attempt to inflate the next bubble on the back of speculative biotech, and I’ve seen a couple of signs of that already.
Yes, there was a big hype in the upcoming biotech revolution in popular transhumanist media a ~decade ago. Lot of it seems to have fizzled out or gone nootropics like stuff. (And even that is meh).
One of the most popular and controversial ways in recent times to use technological means to improve human condition and overcome its natural limitations is gender affirming care, such as hormone replacement therapy. Transhumanism is woke now — hell, “trans” is right there in the name!
Yeah this kind of stuff is why im not the biggest fan of the tescreal label.
still holds - it’s still a bunch that needs a label and that’s the label
even as TREACLES was right there
(i asked emile, they said it was TESCREAL is very searchable. i mean fine)
Tired: Sand Hill Road
Wired: Treacle Mine Road
Apparently including a camera-esque filename in prompts for the latest mid journey release can make it more photorealistic. Unfortunately it also looks like the distinctive AI art style was pretty key to preventing the usual set of AI generated image “tells”. Mirrors, hands, teeth, etc are all very visibly wrong.
Apparently including a camera-esque filename in prompts for the latest mid journey release can make it more photorealistic.
This entire enterprise is just shamanry, we are like two steps away from “throwing a goat into a volcano makes your next prompt more realistic”
deleted by creator
so like a fool I decided to search the web. specifically for which network protocol Lisp REPLs use these days (is it nREPL? or is that just a clojure thing with ambitions?)
and the first extremely SEOed result on ddg was this bizarre blend of an obscure research lisp from 2012 and LLM articles about how Lisp is used in mental health:
Numerous applications and tools are being developed to support mental health and wellness. Among the varied programming languages at the forefront, Lisp stands out due to its unique capabilities in cognitive modeling and behavior analysis.
so I know exactly what this is, but why is this? what even is the game here?
LW: “being a younger brother makes you gay, the Catholic hierarchy is full of younger brothers, ergo 80% of the Vatican is gay”
https://www.lesswrong.com/posts/ybwqL9HiXE8XeauPK/how-gay-is-the-vatican
The modern father of this literature is Ray Blanchard
🚨🚨🚨 Do not take Ray Blachard’s work seriously !
Ray “Asian cartoons cause trans people” Blachard.
Oh look, it’s a penis! I should put some sort of ring around it and see what gets it slightly erect! Repeatedly! For science!
You know, what I find most hilarious about the “fraternal birth order effect” is that they’re so obsessed with eugenics and biological essentialism that they’re ignoring that the very very obvious social fact of growing up with older brothers might have a lil bit more of an effect than “maternal antibodies to the neuroligin NLGN4Y protein”.
edit: Oh right, they pretend they’re accounting for that! Yeah no, I’ve heard all about your “twin studies” and things, I’m not joining your cult.
Considering how rightwingers have tried to link gayness to pedophilia this is a subject I would avoid if I was them. E: and gwern just goes there.
The comments are … “a hoot”
I would bet pretty hard on option #3. The older the parents are at the time of conception, the lower the quality of their gametes, which can translate into various negative health and cognitive effects on the child.
combines ageism, ableism ,and homophobia into one neat package
Also that weird ‘breed early’ fetish common in a lot of rw spaces. And last I looked at it this whole ‘older people do worse’ thing while it mostly seems (it is complicated and lot of things can happen etc)to exist mostly affects the pregnancy less so the child, and even then the effects didnt seem to be big. Not big enough to be relevant here. (But iirc 99.99% of the research in this is only in about pregnancies, so wouldn’t put much stock in ‘older parents affect on IQ’ style research, due to the type of people interested in that).
But im not a researcher, just a person who looked at the stats a couple of years back and apart from pregnancy risk i wasnt that worried.
E: and look at that the op there agrees with me.
Reminds me of an SMBC comic that had a setup along the same lines, that if male birth order correlates with homosexuality and family size trends being what they are, the past must have been considerably gayer on average.
Obligatory Colm Tóibín: Among the Flutterers
:( looked in my old CS dept’s discord, recruitment posts for the “Existential Risk Laboratory” running an intro fellowship for AI Safety.
Looks inside at materials, fkn Bostrom and Kelsey Piper and whole slew of BS about alignment faking. Ofc the founder is an effective altruist getting a graduate degree in public policy.
that’s CFAR cult jargon right?
Mesa-optimization? I’m not sure who in the lesswrong sphere coined it… but yeah, it’s one of their “technical” terms that don’t actually have academic publishing behind it, so jargon.
Instrumental convergence… I think Bostrom coined that one?
The AI alignment forum has a claimed origin here is anyone on the article here from CFAR?
Mesa-optimization… that must be when you rail some crushed-up Adderall XRs, boof some modafinil for good measure, and spend the night making sure your kitchen table surface is perfectly flat with no defects abrasions deviations contusions…
and you wrap it off with some linux 3d graphics lib hacking
Mesa-optimization
Why use the perfectly fine ‘inner optimizer’ mentioned in the references when you can just ask google translate to give you the clunkiest, most pedestrian and also wrong part of speech Greek term to use in place of ‘in’ instead?
Also natural selection is totally like gradient descent brah, even though evolutionary algorithms actually modeled after natural selection used to be their own subcategory of AI before the term just came to mean lying chatbot.
I’m thinking they hired Jar-Jar Binks to the team.
Not sure! What is CFAR?
Center For Applied Rationality. They hosted “workshops” were people could learn to be more rational. Except there methods weren’t really tested. And pretty culty. And reaching the correct conclusions (on topics such as AI doom) were treated as proof of rationality.
Edit: still host, present tense. I had misremembered some news of some other rationality adjacent institution as them shutting down, nope, they are still going strong, offering regular 4 day
brainwashing sessionsworkshops.
“Imagine a technology so useless you cannot run doom on it” https://bsky.app/profile/sosowski.bsky.social/post/3lm63a2srgc24
Utterly rancid linkedin post:
text inside image:
Why can planes “fly” but AI cannot “think”?
An airplane does not flap its wings. And an autopilot is not the same as a pilot. Still, everybody is ok with saying that a plane “flies” and an autopilot “pilots” a plane.
This is the difference between the same system and a system that performs the same function.
When it comes to flight, we focus on function, not mechanism. A plane achieves the same outcome as birds (staying airborne) through entirely different means, yet we comfortably use the word “fly” for both.
With Generative AI, something strange happens. We insist that only biological brains can “think” or “understand” language. In contrast to planes, we focus on the system, not the function. When AI strings together words (which it does, among other things), we try to create new terms to avoid admitting similarity of function.
When we use a verb to describe an AI function that resembles human cognition, we are immediately accused of “anthropomorphizing.” In some way, popular opinion dictates that no system other than the human brain can think.
I wonder: why?
Dijkstra did it first, but it is very ai-booster to steal work without credit or understanding, I guess.
The question of whether Machines Can Think… is about as relevant as the question of whether Submarines Can Swim.
This is too fucking dumb to even be called sophistry.
slophistry
I can use bad analogies also!
- If airplanes can fly, why can’t they fly to the moon? It is a straightforward extension of existing flight technology, and plotting airplane max altitude from 1900-1920 shows exponential improvement in max altitude. People who are denying moon-plane potential just aren’t looking at the hard quantitative numbers in the industry. In fact, with no atmosphere in the way, past a certain threshold airplanes should be able to get higher and higher and faster and faster without anything to slow them down.
I think Eliezer might have started the bad airplane analogies… let me see if I can find a link… and I found an analogy from the same author as the 2027
fanficforecast: https://www.lesswrong.com/posts/HhWhaSzQr6xmBki8F/birds-brains-planes-and-ai-against-appeals-to-the-complexityEliezer used a tortured metaphor about rockets, so I still blame him for the tortured airplane metaphor: https://www.lesswrong.com/posts/Gg9a4y8reWKtLe3Tn/the-rocket-alignment-problem
JFC I click on the rocket alignment link, it’s a yud dialogue between “alfonso” and “beth”. I am not dexy’ed up enough to read this shit.
Yes the 2 rs in strawberry machine thinks. In the same way that an airplane flies. /s
E: it gets even worse as half the AI field says the airplanes fly like how birds do. That is why the anthropomorphization is bad. Because it both doesn’t think as in the function, nor think as in the system. And by anthropomorphizing people make it look like it can do both.
Its almost like flapping part isn’t a requirement for something to fly.
of course MBA came up with this
an airplane doesn’t flap its wings
You ever see a random shitpost video, like the backgroud music, look it up and realize you already have the vinyl record? That just happened to me.
additional layer is that according to comments first track is also used in chinese nature documentaries
And as the same channel demonstrates, there are ways to fly without any wings at all
spoiler
…as well as with them!
Man, it’s really true what they say: the West can’t meme.
Here’s a screenshot of two screenshots of an AI bro getting humiliated:
the hawk tuah teapot is actually hysterical which is how I knew the poster was taking the piss
Some dark urge found me skim-reading a recent AI doomer blog post. I was startled awake by this most unsettling passage:
My wife wrote a letter to our infant daughter recently. It concluded:
I don’t know that we can offer you a good world, or even one that will be around for all that much longer. But I hope we can offer you a good childhood. […]
Though the theoretical possibility had always been percolating somewhere in the back of my mind, it wasn’t until now that I viscerally realized that P(doomers reproducing) was greater than zero. And with other doomers no less.
Left brooding on this development, I drudged along until-
BAhahaha what the fuck
I can’t. This is beyond parody.Completely lost it here. Nothing could have prepared me for the poorly handwritten wrist tattoo.
Creating space for miracles
Doom feels really likely to me. […] But who knows, perhaps one of my assumptions is wrong. Perhaps there’s some luck better than humanity deserves. If this happens to be the case, I want to be in a position to make use of it.Oh how rational! Willing to entertain the idea that maybe, theoretically, the doomsday prediction could be off by a few days?
I’m not sure that I ever strongly felt that I would die at eighty or so. I had a religious youth and believed in an immortal soul. Even when I came out of that, I quickly believed in the potential of radical transhuman life extension.
This guy thought he was getting clean but he was actually replacing weed with heroin
I really convinced myself that “doomsday cult” was hyperbole but uhh, nope, it’s 107% real.I don’t know that we can offer you a good world, or even one that will be around for all that much longer. But I hope we can offer you a good childhood. […]
When “The world is gonna end soon so let’s just rawdog from now on” gets real
Teach your children to envy the dead
I had a religious youth and believed in an immortal soul. Even when I came out of that, I quickly believed in the potential of radical transhuman life extension.
My dude you’re so, so, sooo close to realising it, you should spontaneously quantum-tunnel into self-awareness any second now
The storm rages within
Yet I hold F to pay respects
At the start they state
The disappointment of imminent death is all the more crushing because just a few years ago researchers announced breakthrough discoveries that suggested [existing, adult] humans could have healthspans of thousands of years. To drop the analogy, here I’m talking about my transhumanist beliefs. The laws of physics don’t demand that humans slowly decay and die at eighty. It is within our engineering prowess to defeat death, and until recently I thought we might just do that, and I and my loved ones would live for millennia, becoming post-human superbeings.
This is, frankly, bonkers. I’d rate the following in descending order of probability
- worldwide societal collapse due to climate change
- we develop an AI that will kill us all for unspecified reasons
- we establish viable self-sustaining societies outside the limits of Earth
- we develop techniques that allow everyone to live effectively forever
If the first happens, it removes the material requirements for the latter things to happen. This is an extreme form of “denial of the flesh”, the inability to realize that without food or water no-one will be working on AI or life extension tech.
“Im 99% sure I will die in the next year because of super duper intelligence, but in a world where that doesnt happen i plan to live 1000 years” surely is a forecast. Surprised they don’t break their own necks on the whiplash from this take.
yet I hold
space for it
Rupi Kaur should sue
:'( sad one. feel bad for the bebe, being raised by insane people.
Oh that tattoo is regrettable
Doom feels really likely to me. […] But who knows, perhaps one of my assumptions is wrong. Perhaps there’s some luck better than humanity deserves. If this happens to be the case, I want to be in a position to make use of it.
This line actually really annoys me, because they are already set up for moving the end date on their doomsday prediction as needed while still maintaining their overall doomerism.
Also, man why do I click on these links and read the LWers comments. It’s always insufferable people being like, “woe is us, to be cursed with the forbidden knowledge of AI doom, we are all such deep thinkers, the lay person simply could not understand the danger of ai” like bruv it aint that deep, i think i can summarize it as follows:
hits blunt “bruv, imagine if you were a porkrind, you wouldn’t be able to tell why a person is eating a hotdog, ai will be like we are to a porkchop, and to get more hotdogs humans will find a way to turn the sun into a meat casing, this is the principle of intestinal convergence”
Literally saw another comment where one of them accused the other of being a “super intelligence denier” (i.e., heretic) for suggesting maybe we should wait till the robot swarms coming over the hills before we declare its game over.
New piece from 404 Media: Facebook Pushes Its Llama 4 AI Model to the Right, Wants to Present “Both Sides”
On a related note, Baldur Bjarnason has chimed in noting how he called this exact shit happening:
Remember when I told you that using these LLMs was like giving US tech a bigotry dial for all your writing?
