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. Also, happy April Fool’s in advance.)
My University Keeps Sending Me Stupid Emails About AI, a continuing series:
From the email:
The debate will be chaired by Michael Pike. Speaking for the motion are Prof. Gregory O’Hare (TCD), Maeve Hynes, Prof. Gary Boyd and Chatgpt assisted by Student Curator Ruan McCabe, all UCD. Speaking against the motion are David Capener (UU), Lucy O’Connell, Peter Cody and Meabh O’Leary, all UCD.
Artificial Intelligence is the future of climate resilient design in the same way that asian pseudo-medicinal ED treatments are the future of the white rhinoceros.
Why even involve ChatGPT? The “against” side should have been allowed to bring a speak 'n spell.
Notwithstanding the subject matter, I feel like I’ve always gotten limited value from these Oxford-style university debates. KQED used to run a series called Intelligence Squared US that crammed it into an hour, and I shudder to think what that’s become in the era of Trump and AI. It seems like a format that was developed to be the intellectual equivalent of intramural sports, complete with a form of scoring. But that contrivance renders it devoid of nuance, and also means it can be used to platform and launder ugly bullshit, since each side has to be strictly pro- or anti-whatever.
Really, it strikes me as a forerunner of the false certainty and point-scoring inherent in Twitter-style short-form discourse. In some ways, the format was unconsciously pared down and plopped online, without any sort of inquiry into its weaknesses. I’d be interested to know if anyone feels any different.
I did two Intelligence Squared debates in the UK. The best bit was I got paid. The second best was the people running it were actually very nice and quite competent. Recommend to all.
As for content, the first one featured the guy behind Liberland, so lol
Competitive forensics is an intellectual exercise, not a way of addressing or solving social problems. It’s like translating Cæsar’s Commentaries — you’re not likely or expected to have any new and valuable insights that previous translators missed, but it builds and demonstrates certain kinds of abilities and readiness.
Seeing it any other way, using it any other way, is a conceptual error with the usual outcomes.
I have no knowledge or insight on the topic, but I used to get recommendations for “intelligence squared” videos on YouTube and I always thought it was a terrible, self-aggrandizing title for a series or event. Smart People Taking About Smart Things.
Intelligence2 didn’t seem half bad when Robert Anton Wilson was the one talking about it way back when, in retrospect all the libertarianism was a real time bomb.
Intelligence Time Cubed now, that’s the real deal.
New video from Jessie Gender, providing one long sneer at gen-AI: How AI is Destroying our Dreams
AI-Powered Wi-Fi 7 Versatile Outdoor/Indoor Mesh AP
we’re at the “the washing machine without a dateclock is marked millenium-bug safe in its marketing brochures” level of stupid
(that might seem like a stupid comparison but it’s one of the things I most viscerally remember seeing from that time (when I was still a youngin who was still largely years away from computertouching))
What if America built a Death Star, but it was 50% exhaust ports by surface area?
https://www.cnbc.com/2025/04/04/how-the-eu-could-target-us-big-tech-with-its-tariff-response.html
Also the big fuckoff wave motion laser cannon thing is aimed through the ship itself.
with a whole 2 Ts!
taking bets on whether that’s 2 terminators or whether it’s the amount of USD they want to burn to make it into a machine god
Llama 4 Maverick offers a best-in-class performance to cost ratio with an experimental chat version scoring ELO of 1417 on LMArena.
Everything on that page is silly nonsense, but this takes the cake. An Elo rating for chitchat!!!
coming up soon: tacticool arena prompt fights, where they get ranked on style and guile
(as was said on whose line is it anyway: “where everything is made up and the points don’t matter”)
I bet I could win a conversation against any LLM ant day.
Neither, actually. They were testing it by asking how many "T"s appeared in “Llama Four” and it kept saying “2” so they decided to roll with it.
if you, like me, were wondering what the point of that 25 hour non-filibuster filibuster by Booker was, here’s one potential answer.
Booker held a filibuster that wasn’t a filibuster - and he scheduled it to let him avoid attending his own committee’s probe of his Big Tech pals. […] Oh and Booker and Dems provided unanimous consent to advance a Trump nominee right after Booker’s speech.
Sorry im going to go offtrack here again, I mentioned it on bsky, and got no traction (not odd, as I think nobody cares about Booker his filibuster thing (stick a pin in that)), but I have some weird leading into conspiratorial questions about the whole thing.
First I heard about this thing is when people said it had gotten 200m likes on tiktok. Which seemed a bit high so I checked, and saw articles say it had gotten 300m. This seems impossibly high. For example, the global K-pop phenomenon ‘Gangam style’ has gotten 5.5B views and 30M likes in 12 years. I have a hard time believing that in 24hs this centrist political debate thing (which are not popular) has gotten 350m (the highest count I saw on a news site) likes.
Which makes me wonder a lot more if tiktok has simply given up on properly counting likes and just is winging it. Esp for larger events. Could be that people just like things instinctively on tiktok (I did check if you could like a thing multiple times, but nope, one account one like it seems). I found the whole thing weird.
That is didn’t do jack shit (apart from giving people hopium about Booker, while it seemingly being his way of avoiding responsibilities) is the cherry on top.
I’m noticing my confusion.It is fucking weird.not odd, as I think nobody cares about Booker his filibuster thing (stick a pin in that)
I agree, or at least anyone thinking critically. I think that anyone would agree that the speech was, as you said, hopium. He’s giving braindead dem voters what they want: a nice, tall, liberal man who looks like he is resisting the reds. Expect him to run for the democratic nomination in 2028, assuming the trump presidency lets an election happen.
RE: view counts. Most charitably, maybe it’s 300m views aggregated across different sources. Neutrally, I mean I wouldn’t be surprised if tiktok, or any other social media platform was manipulating view counts. Least charitably, someone probably asked an LLM for the view counts and just took the answer because people are fucking stupid
No it was 350m likes, not views. That is why im so confused. even at an high rate of 1 in 10 people liking it, it is just so large I’m confused.
“How about a source senator?” another source on hopium, not noticing that in no world do those numbers align properly. “since amassed over 700,000 followers. … It garnered over 350 million likes, with over 150,000 active viewers at the tail end …”
Are you telling me it’s improbable that the equivalent of every single American and then some liked that video?
I mentioned this on bsky and somebody went ‘well it was very popular all over the world’, and I, the weird european who focuses too much on the US politics had not even heard of it. So I just had a few alarm bells going off. But yeah, lets say 1 in 5 people like it, that is a casual 1.7 billion viewers. Large part of the worlds population joined in.
As a Canadian I knew it was happening, but I never tuned in or anything like that.
OH. so two things.
- I obviously misread your comment, as my brain filtered “likes” into “views” as I guess it subconsciously thought that was more plausible.
- I believe on tiktok live specifically, you can like something multiple times, and it is counted. I occasionally watch a stream that floats at about 100 viewers, and sometimes this hits 100k likes over the course of two hours. So to hit 350m likes over the course of 25 hours, you might need like 280 viewers on average, which seems doable.*
*Please fact check this arithmetic. I have run out of motivation, in general
oh, cool, I didn’t think I’d learn a new way to hate the “likes” clusterfuck that plagues the internet, but here we are!
That would explain a lot and would remove all my confusion about it. (also makes the number useless and lol at everybody running with it even more then).
Gamers Nexus put out their April Fool’s joke for this year, and became the first journalistic outlet to break Betteridge’s law of headlines in the process.
they should’ve had an overlay tracking counter in that opening montage, goddamn
(just got to 4min:) “oh! it’s not just opening montage” ahahaha
Came across this fuckin disaster on Ye Olde LinkedIn by ‘Caroline Jeanmaire at AI Governance at The Future Society’
"I’ve just reviewed what might be the most important AI forecast of the year: a meticulously researched scenario mapping potential paths to AGI by 2027. Authored by Daniel Kokotajlo (>lel) (OpenAI whistleblower), Scott Alexander (>LMAOU), Thomas Larsen, Eli Lifland, and Romeo Dean, it’s a quantitatively rigorous analysis beginning with the emergence of true AI agents in mid-2025.
What makes this forecast exceptionally credible:
-
One author (Daniel) correctly predicted chain-of-thought reasoning, inference scaling, and sweeping chip export controls one year BEFORE ChatGPT existed
-
The report received feedback from ~100 AI experts (myself included) and earned endorsement from Yoshua Bengio
-
It makes concrete, testable predictions rather than vague statements that cannot be evaluated
The scenario details a transformation potentially more significant than the Industrial Revolution, compressed into just a few years. It maps specific pathways and decision points to help us make better choices when the time comes.
As the authors state: “It would be a grave mistake to dismiss this as mere hype.”
For anyone working in AI policy, technical safety, corporate governance, or national security: I consider this essential reading for understanding how your current work connects to potentially transformative near-term developments."
Bruh what is the fuckin y axis on this bad boi?? christ on a bike, someone pull up that picture of the 10 trillion pound baby. Let’s at least take a look inside for some of their deep quantitative reasoning…
…hmmmm…
O_O
The answer may surprise you!
First graph reminds me of that ‘human progress stunted by dark ages/Catholic church’ image which makes historians so mad.
E: also the idea of these LLM based AGIs hiding and evading detection is quite funny. They have quite the power/gpu/storage footprint. But sure the elphant has a few levels in sneak and now it can just go by unnoticed. In an era where we previously had other threats which were looking to abuse similar resources. The reaction to ‘wow all our gpus suddenly maxed out’ will just go from ‘ah cryptominer’ to ‘ah, a cryptominer or somebody is messing with an LLM’. I’m sure they will give the AGI some magical abilities to get around this.
Fucking anti-khaganate propaganda. Real ones know this is the true version of that progress image.
slabs table
This is is sea peoples erasure and you know it! Damn liar!
First graph reminds me of that ‘human progress stunted by dark ages/Catholic church’ image which makes historians so mad.
But muh religon bad
Bruh what is the fuckin y axis on this bad boi??
words that the AI-O-sphere are gonna vomit
shout out to the “legions of CCP spies”, if you’re listening
你好 同志
(I really hope they don’t enshittify google translate, my ability to make jokes like this would be destroyed, a personal 9/11 if you will).
-
New(ish) piece from Gary Marcus: AI has (sort of) passed the Turing Test; here’s why that hardly matters
Ended up reading it a couple times, thinking of turning my thoughts into a full-length post.
The fact that the turing test is still the go-to example of a machine intelligence test goes to show that the AI field needs more haters in it.
Sorry you are wrong, it is very important that the AI field has a ‘can it imitate a woman’ test. They should base their field on this idea.
(For the people who do not know, the OG Turing test involves faking being a woman).
These fuckin nerds don’t care about the imitation game, they only want the imitation gams
My favorite meme message board on the internet keeps putting out bangers, but this one is extra good and relevant http://www.b3ta.com/board/11416629
NaNoWriMo? Na, No Mo’. Does this have anything to do with their bungled AI policy? Maybe, maybe not, but hey, the news article that I saw this announcement in thought that it was pertinent to mention.
Why does NaNoWriMo need a nonprofit anyway? What’s next, No Nut November LLC?
As funny as that is, I am sure that there are nonprofits that are aiming to stop fapping.
however they are exactly canceled out by nonprofits aiming to promote fapping
the Effective Onanism movement
Refractorive Altruism
There’s a fucking for profit that does this and it’s used by the god damn USA Speaker of the House so that his son can monitor his fondling
I am familiar. Not a problem!
considering their endless history of scandal, good riddance
Not dating your press releases on your own site is a choice. (Lot more places do this and it is driving me slowly mad) So, did they drop this yday? Some weird schrodingers april fools joke deflection attempt?
Their dumb contraction should have doomed the project from the start…
It is a CIA psyop to prevent revolutionary thought from arising in written media, in that it’s a self-inflicted PIP that makes you feel inadequate in your creative writing ability. CIA funding (obviously funneled through NGOs via CIA front USAID) is how it’s survived so long with such a stupid name. Of course, now that government spending on the CIA is going down, so is NaNoWriMo. All that’s left is for it to be supplanted by some benevolent Chinese competitor that people will say steals your data, or something.
/s /s /s
I demand you prove your assertion! preferably using wallspaghetti and alphabet soup, of course. so’s I know it’s totes legit and original!
(/s ofc, and link too (incase anyone didn’t see that))
[This is what tankies actually believe!!2!one!!!]
I am fairly certain that their stupid AI stance at least played a role. I’m part of a largish writing community, where at least several hundred people did NaNo each year, a good part of them donated too. Last year, no one partook, instead we did our own internal thing.
If this happened across other communities too, and I wouldn’t be surprised if it did, they must have felt it financially.
I’ll believe it! I didn’t want to make that claim since I had no evidence.
AI Overturns Centuries of Forensic Fingerprinting Practice?
Published in Science… Advances
Probably a fair bit to sneer at in the actual study that I’m missing, and the article I first found it in is peak AI Hype. (Big Forensics is trying to keep you from knowing the Truth as found by an undergrad with a GPU) But the part that I found most concerning is that even the whole paper doesn’t appear to break down their 77% accuracy index and provide the specific result ratios that go into it. In a field where each false positive represents a step on the road to innocent people being convicted of major crimes I would really like to know that number specifically.
At a <30s glance, I’m going to guess the model is finding correlations in the artefacts of fingerprint taking, digitalisation, presentation and so on instead of fingerprints proper, like every damn time this sort of story comes up.
OK hear me out: they make a new season of Hell Girl, but she only gets people who use gen ai
OpenNutrition –
a datasetan LLM that allows you to play “vibe nutritionist”https://news.ycombinator.com/item?id=43569190
First response is good quality:
This is not a dataset. This is an insult to the very idea of data. This is the most anti-scientific post I have ever seen voted to the top of HN. Truth about the world is not derived from three LLMs stacked on top of each other in a trenchcoat.
Dang… The author realized an actual problem / inconvenience people had, and somehow went on to think “I know! I’ll have a random nonsense generator make up the data! That’s a great solution!”
Just had a video labeled “auto-dubbed” pop up in my YouTube feed for the first time. Not sure if it was chosen by the author or not. Too bad, it looks like a fascinating problem to see explained, but I don’t think I’m going to trust an AI feature that I just saw for the first time to explain it. (And perhaps more crucially, I’m a bit afraid of what anime fans will have to say about this.)
Given the apparent state of the art for autogenerated captions (and by extension the initial challenge of speech recognition) being firmly in the “good enough” range I would not trust the chain of speech recognition -> translation -> text-to-speech. That’s a lot of room for errors to chain, multiply, and obscure themselves through GIGO even if the latter two steps did work as expected.
That image needs a content warning for YouTube Face. Jesus H. Fuck.
