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. Also celebrating my birthday on Friday)
OT: I just turned 26. Iām taking a trip to a local steakhouse to celebrate.
This policy seeks to ensure that our courses focus on requisite cognitive skills by default. It provides students with the opportunity to develop the skills they need to conceptualize, outline, draft, revise, and edit their work by forbidding the use of AI for these purposes in connection with work submitted for credit. It also forbids using AI to translate work for credit, thus providing students with the opportunity to develop and exercise their own fluency with legal English. [ā¦] The use of AI is prohibited for aid in conceptualizing, outlining, drafting, revising, translating, or editing any work submitted for credit. AI use is prohibited for any use for any purpose in any exam situation. Students may not upload course materialsāincluding assignments, readings, slides, class recordings, or other class contentāinto generative AI systems.
Scott Alexander has told many stories about why he is pseudonymous. In 2013, when he was applying for medical residency, he told a simple story which begins with an interviewer saying he would have to take his LiveJournal down if he were accepted:
I originally deliberately linked this blog and some of my other writings to my real name in order to bury Google evidence of certain stuff I did online when I was a dumb teenager, but that mission is pretty thorougly (sic) accomplished and now I agree with my interviewer that even if I donāt get hired (cue serious of elaborate ritual hand gestures to ward off unthinkable misfortune) it is to my benefit to switch back to my normal totally un-Google-able pseudonym of Scott Alexander (well, actually just my first and middle name).
FYI, this is why its a red flag that someone has public profiles on dozens of sites. Its a standard technique to bury negative posts in Google results, PR firms will write the profiles and get you a few press releases and talks to help. I hope this is the last thing I have to say about this particular creep.
What did he do, leave written evidence of white supremacist leanings?
if you held a gun to my head I would guess it would be his racism (he thinks that different groups have natural places in society driven by their inborn hereditary differences), but I donāt see any point in trying to track it down. He is clever but uses his brains for very common, very bad ideas and the things he has done since 2013 are much worse than the things he could have posted as a young man.
Annie Altman, Sam Altmanās sister who has accused him of abuse has recently posted:
If any harm ever comes to me āmysteriouslyā - you know who it was.
I love my life.
https://xcancel.com/anniealtman108/status/2089964553207775529
She also added " I am not suicidal, nor on any kind of drug I could overdose on".
As usual this sort of message actually makes me worried about what she might do to herself. :(
Iāve never heard of abuse allegations against Altman, so I looked them up and holy shit wow.
Came across this on the orange place: dontpastetheai.com
I was expecting like nohello.net but instead we get a decent intro explaining nuance as to why people ask other people questions followed by how to use AI and hide that you used it.
Is it me, or do attempts to make movies with AI tools
https://www.lesswrong.com/posts/24RKHEkwgZ6Hm6ygY/can-an-llm-make-a-feature-length-movie-on-its-own
Remind me of SCPs
Quoth the blog post:
I have a YouTube channel where I mostly get LLMs to write albums and make them into music videos.
ā¦ the fuck did I just read?
If you wonder how you can look at any of this without viscerally realizing that itās absolute dogshit, I found the
authorsperpetrators conclusion very insightful.Where we go from here
I expect these issues to all be fixed by the end of 2027.
I expect no-longer-obviously silly movies to be doable within two years, right now itās going slower than I expected in January 2026 because the frontier labs are no longer competing over having the best video generator, like they were when we had Sora2 and Veo 3.1.
And WHY are they no longer competing over this? Tell me. Could it be that its incredibly expensive to make somthing that nobody wants except fraudsters, and that further improvement via the kind of ML we do just produces exponentially diminishing returns for ever increasing huge amounts of effort and they realized itās not worth it?
Just like they will never understand that the openAI āpauseā for āsecurity and alignmentā is a convenient excuse for the fact that they have no goddamned money and are on a treadmill to oblivion.
Oh boy, this is the most straightforwardly delulu lesswrong post Iāve ever seen.
If you see issues in the YouTube video, rest assured that those issues survived even more than two generation attempts.
lol
AI slop is very good at falling into the uncanny valley, whether that paragraphs of pure nothingness from a text extruder, or strange almost-human looking imagery that looks and feels inexplicably wrong. Its a great way to produce some accidental nightmare fuel, especially considering the AI bros glazing such shit are incapable of recognising its accidentally disturbing nature.
As a āfunā sidenote, there was an SCP article on the site which acted like a slop generator (SCP-1004, Factory Porn). Its writer, DrBright, was permabanned in 2022 after a lengthy history of sexual harassment was uncovered (and the article was deleted altogether in February of this year). Feels oddly fitting, in the worst way possible.
Iāve had the same thought about Factory Porn, itās is basically what Image slop generators are, right down to the addictive nature, even when it isnāt porn being generated.
It still cant keep scenes consistent for more than 5 seconds amazing.
I got to 47 seconds and thereās a guy driving a tent stake with a mallet that isnāt quite there, only to have the mallet disappear.
āSo you printed out a piece of the internet for us to throw away.ā
My favourite error, and it appears repeatedly, is that instead of an object passing from one character to another, a copy of the object suddenly finds itself in the second personās hands. It is shockingly bad at portraying any kind of action! Itās never going to get better!
I am mildly fascinated by the types of issues that appear and the contrasts with other things that can locally in space and time look right.
An object doing nothing will have a consistent surface and show perspective relative to the viewpoint. The systems are able to have representations of surfaces of different types and how they can fit together within an object. I am of course talking about within a given generation, not between generations where its utterly unsurprising that consistency is very difficult or impossible.
But multiple objects in interaction with each other do very strange things. Their relative sizes change as if they are in different positions relative to the camera. They snap between individually plausible relationships, without going through intermediate states. Doors open on the hinge side when the other side is not visible. The relative size of and distance to the background can suddenly change, as the foreground suddenly interacts with something that should be far in the background.
Objects that change also do so in bizarre ways. Living things morph between different archetypes. Flames in particular change wildly between types and sizes and respond to the facial expressions of humans, smoke and water effects blend together. Sudden movements with no cause occur, sudden morphings of one object into another when the context around them changes and something else makes sense, especially when held in a hand. Time-reversed motions occur mixed in with time-forward motions, and slow-mo with regular time. Debris suddenly appears from an object but when the dust clears the original fails to have been eroded away into the fallen debris.
On multiple occasions, a carried candle keeps moving with a characcter rising and falling with their footsteps hovering in front of them when both hands become occupied with other tasks. This is fascinating and indicates that the relationship between the two objects motion is represented separately from the idea of something being ācarriedā. (This is positively Lovecraftian.) Candles also indicate something else, with flowing wax changing wildly in timescales that do not make sense, with the system apparently understanding that there are different patterns but having no idea how they come about or change. There is no generality here, just an endlessly compounding list of rules of thumb.
Excessive correlations between objects across the frame are rampant. Footsteps preferentially synchronized. People in the background lipsyncing with foreground characters, faces changing expression in unison. Textures changing across multiple objects at once.
I have said it before, and I will say it again, the relationship between the outputs of these systems and what they mimic are precisely the relationship between a stick insect and a stick, or a social-parasite-beetle and a baby ant. Not just in form, but because that is also precisely the same forces that drove both things into existence - superficial resemblance to something else with a very different internal set of causes, that can fool to a first inspection by the inspection applied but just isnt doing the same thing. And again, SCP-2030 feels the same.
Wow, I watched 15 minutes from 45:00 and itās just likeā¦ wtf. The guy is old, younger, mustache, no mustache, sets a gun down and then pulls another gun out of thin air, sometimes the gun itās a double barrel side x side, sometimes itās an over-under, sometimes a muzzle loaderā¦ The candles, wtf is up with the candles? The dog is remarkably consistent but the mouth doesnāt move right and itās just a horror show. The cave setting changesā¦
Yeah, itās just so fucked up.
In the same cut, the guy then materializes a hat in his hand and puts it on, replacing the hat he was already wearing.
45 minutes is even better
I picked a bit randomly where a walking stick changes shape, disappears, appears, the guys hair changes color etc.
SlateScott first floated his bright idea about āmedical ethics are for foreigners not a good-hearted genius like meā in 2010.
Hereās what happens. In some far off corner of the world, some psycho doctor does something horrible like deliberately inject a patient with flesh-eating bacteria. The media hears about it and panics the public, who become suitably enraged and make demands of their politicians. Their politicians go to leading medical administrators and demand that doctors be more ethical. The leading medical administrators obviously donāt control the actions of every single psycho doctor, so they think quick and say āOkay, weāll make all medical students take a class on Ethics and Professionalism,ā hand the implementation off to professors, and then raise their own salaries for thinking of such a brilliant idea.
I guess California and Alberta and Sweden and Japan were far off from Cork, Ireland. How did he pass that class?
Hippocratic oath is government-imposed oppression apparently
There is no way that the class did not cover how MDs and nurses have been central to state atrocities for 150 years, from Bushās torture program to MK-ULTRA to Unit 731 to the original eugenicists (who kept committing wild sterilizations into the 1970s). There is no way that it did not cover how authorities often hint at what they want done without being so vulgar as to say it where it could come up at their trial.
But didnāt he get that authoritarian followers are deterred if an authority figure like a professor says āthe following acts are unacceptable?ā
OpenAI is slowing development because of uh safety https://www.theguardian.com/technology/2026/aug/18/open-ai-pause-hack
Definitely not because improvement is stagnating and theyāre running out of money
It just occurred to me that this makes Anthropic look bad, actually
if OpenAI is slowing down because they concerned about
moneyAI safety, then it makes Anthropic look like fools for publishing that blog where they waffle on about keeping up the race because of ābad actorsāIf OpenAI is slowing down, an Anthropic that actually is concerned about safety would slow down too. But they wonāt, because
moneyChinaābad actorsāWell Iād like it if they actually stuck to this, since it would finally lessen the flow of AI slop. Wouldnāt be surprised if this gets quietly ignored because OpenAI
Update on the slop dentist situation: second opinion says that loved one has zero cavities. I suspect that her only problem was having a non-median mouth that wasnāt well-represented in Pearl AIās training data.
It was called fucking Pearl AI? god these people suck so fucking bad
ugh, right?
otherSneerClub has a post where some kind of statistics/science VTuber (??) analyzes aellaās twitter polls and resulting studies. itās fun. Iāll link to the substack directly but the reddit post should be easy to find. Link
this blog post is so badass
the woman herself shows up in the substack comments with the most pitiful display of āIām just a smol self-taught bean why are you being so mean to me?ā
Erika rightly responds by calling out Aellaās own attitude towards the people who know how to apply the level of rigor to the data that she (apparently admittedly?) doesnāt. Like, you canāt simultaneously be this dismissive of the norms and processes of the scientific establishment and also claim ignorance and inability when your inability or refusal to follow those norms leads you to crap conclusions.
I really like the self perpetuating optimism loop thing in the reply to Aella, itās like a whole new cognitive bias I hadnāt thought applied to rationalists.
I also still believe academia is a deeply broken culture. I believe this in large part because most academics have told me this.
You seem to believe uncritically what they tell you here. The truth is, everyone LOVES to complain but if you tell them āTHEN LEAVE??ā theyād rather keep their position. And itās very normal, actually. most systems (idk, liek law, healthcare, environmental protection) are ādeeply broken.ā
Maybe you think that only systems worth engaging with are those with self-perpetuating optimism loops? where everyone is so happy and satisfied with the community and activities? If so, it matches really well the intellectual environment I think you are in.
So last night I finished editing the latest odium symposium episode (āezra klein is a disgusting worm,ā https://www.patreon.com/OdiumSymposium/posts/1-ezra-klein-is-166891080 ), and i went to upload it to our rss host, zencastr. friends. zencastr is so fucked now. the only interface, as far as i can tell, is an llm, and it tormented me. im just gonna dump the screenshots. this is a fragment of an exhausting hour of back and forth.
note that i apparently have 10k credits, which i assume are temporary and there as a welcome to the new service. a single credit a few bucks so in total the bill just for the conversation comes out to about 6-7 dollars
it also tried to get me to use this absolutely horrible episode description
i asked it to schedule the episode for 9 am aug 18th and it said it did and it turns out it went out this morning. cool! if you listen by rss itās out now i guess.
i posted on reddit asking how to switch hosts. the ceo found that post today and offered his hand, which i bit
i keep thinking about that bit in a philip k dick novel where the protagonistās door tries to chisel tips out of him before opening, and threatens to sue him when he bypasses the lock. i hate the future/present so much
holy fuck thatās infuriating
imagine the worst possible CLI and then make it even worse and then make it an asshole
real rich of their CEO (presumably? it looks like a burner account which still checks out if itās their CEO or someone who works for him doing damage control and trying to take this shit to DMs) to both anthropomorphize their chatbot and assert that the problem is you need uhh training on the system that doesnāt do what you want it to do and keeps trying to charge you for shit that makes your workflow worse
honestly the thing that makes me think it is actually their CEO is Iāve watched these fuckers use āletās take this to DMs/emailā as an attempted power move. they seriously think we want to talk to them privately about, like, anything at all. no fucker Iām here to call you out in public because your product is shit and youāre shit too.
I havenāt been listening to your podcast (no shade, itās just not my thing) but Iāll have to listen to this because my God that man frustrates me.
He basically exists to funnel run the:
intellectually curious libs -> smug superiority complex with status quo bias -> reactionary
pipeline.
hell yeah, get his ass! o7
What were they thinking?!
PS:
So last night I finished editing the latest odium symposium episode (āezra klein is a disgusting worm,)ā¦
A+ hook
The episode dropped and Iām like 2/3 in and have already passed my previous record for the number of times Iāve had to pause an episode of OS to make a little angry noise.
Well it took longer than I expected, but the AI 2027 guys finally posted a second update
This table is essentially the TLDR
They also finally explained the ā2025 was 75% of AI 2027ās progressā thing and it seems like that number was mostly skewed by revenue (edit: valuation) numbers (Edit 2: apparently this is wrong?? the 75% figure is supposed to represent 2026 progress. So 2025 was still 65%, I think. These people are confusing)
For some reason their website lists the date of ASI as 12/2028 even though their actual graph only had a 32% chance tops of ASI happening on that date
<edited> Didnt I hear them say a while back that its not even revenue numbers? That its valuation of the companies being taken as proxy of value produced.
I find myself caring so little what these people think about anything
That is even worse than using the ARRs, which Ed Zitron has extensively explained are wildly distorted numbers done by shuffling sales and costs around to get really good 30-day to 1-month stretches.
Hereās that side-by-side comparison
Bioman has already pointed out the āEconomic Valueā numbers they are using for these tables are probably bullshit (based on self reported run-rate extrapolations that are deliberate distortions at best, based on VC valuation at worst). To add to thisā¦ the compute values are also probably bullshit. They are likely based on data center announcements and not confirmed totally complete data centers (Ed Zitron has ripped into how much bs there is in data center announcements). āCoding Time Horizonā is probably METR, which, while some of the best numbers for estimating actual AI improvement for practical purposes, are still really bad in several key ways. (They donāt have enough human task performers for the longer duration tasks even if everything else was right, because they arenāt, and there are several ways systematic bias could have leaked in and compelted distorted the constructed measure of task duration.)
āAI Software R&D Upliftā is the single most important category to their scenario of recursive self improvementā¦ and they have it at a small fraction of what they estimated.
I took a quick look over the article again, this spreadsheet contains how they measure their metrics. āComputeā seems to be based off number of chips, which is probably in part based off data centres anyways
And apparently they have ditched METR and are now using ācoding uplift (i.e., how much of a speedup AIs are providing to software engineers at AGI companies) and revenue.ā
Ed Zitron has also explained his suspicion that lots of GPUs are sitting around in warehouses waiting to be installed, in some cases sold (to juice NVIDIAās revenue) but not even shipped yet.
And Iām really skeptical speedup from AIs claimed by the LLM companies is in anyway related to reality.
According to one of the writers apparently 17% is supposed to be āprogress is 17% slower than expectedā, so the 34% figure here implies progress is 34% slower than expected, so thatās actually worse
That wouldnāt make sense, because 1.07 would mean progress is negative. I think they meant it as a multiplier? So 1.07 is almost exactly what the predicted, .17 is only 17% of what they expected.
Anyway, it doesnāt really matter, because so much of the input numbers to these calcs are garbage.
Yeah thatās what tripped me up, because if that was true then itād make no sense why the percent of progress went up while the uplfit number went down.
The percentage isnāt total progress to some key point of AI 2027, it is progress relative to their timeline. A constant 1.0 would be staying on track with their predictions, numbers less than that would be falling behind. So they are admitting the real numbers are falling behind their predictions more and more (while still not acknowledging their entire timelines was bs in the first place).
I read their spreadsheet and their % progress numbers in terms of AI capabilities is MASSIVELY skewed by Claude Mythos. The rest of it is below the mark if you exclude that
Alternative title: Adam No! Bonk
Word is that both Adam Conover and Carol Queen have backed out.
can we kick aella and her friends out of perversion
Perverts (shower-positive)
Every time a new person learns who Aella is, an angel loses its wings.
Nina Hartley coshing a Nazi with a dildo was not on my 2026 Bingo card but I can hope
Edit / Adam Conover of Adam Ruins Everything seems like a mainstream āMythbustersā type TV presenter, if he likes older women and she needs a dildo caddy this has potential. We talked about CJ the X here.
Some things are classics for a reason.
Regarding Grettaās job title āorgy organizerā:
besides the point but orgynizer was right there
https://bsky.app/profile/mecha-lenin-3000.bsky.social/post/3mtcx7bixkc2u
Whoever CJ TheX is, he looks like he strolled straight out of Neil Straussā book. Is it valid to call somebody an āunreconstructed pickup artistā if there was never any attempt at (or even possibility of) reconstruction?
so maybe this is a bit too obvious to state at this point but: conoverās very obviously tied up in aellaās plzdontkillus EA influencer scheme, his apparent anti-AI stance is purely performative, and the apology he did for doing marketing for worldcoin on his channel isnāt accepted. he knew what he was doing.
also, I really donāt need the Adam Ruins Everything guy in the same part of my brain as sex, thanks
countdown to Comedian Sex Pest Scandal in
cj noooā¦ someone come get my favorite parasocial essay-they out of there
A 404 Media investigation was able to reveal Amazonās book buying operation, which hasnāt been previously reported, by placing a tracking device in a rare book we suspected would be acquired by an AI company for training data, and following it around the country to its final destination.
That final destination was an Amazon warehouse in Las Vegas, Nevada. Amazon employees who work at this location say all they do is receive massive shipments of printed books which they then cut the bindings off in order to scan the books more quickly. The printed book is destroyed in the process. The logo of the Amazon team that works at this warehouse, called VGT3, is a dinosaur, brandishing its teeth and with a book in its hands.
I think Iāve ranted before about how gallingly unnecessary the destruction is. Google books invested the time and money to improve their nondestructive scanners and apparently thanks to the incoherent application of copyright law that was not only unnecessary but actively counterproductive and opened them up to more legal risk. Is there a legal theory that allows me to smack Judge Alsup with a rolled-up newspaper? Might make me feel better.
Greta D. says she met Yudkowsky at the wedding of his ex and Duncan S. (cw: rationalists propositioning each other) āMawwage, is a dweam within a dweam ā¦ā
