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 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 soo 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 thisā¦)
Finally a use case for āAIā: defrauding community colleges.
sigh
New piece from Tante: These are not the same, arguing that the infrastrucure left after AI bubbleās burst will leave infrastructure which will be actively damaging to a democratic society.
Its not the first piece Iāve seen about the bubbleās potential aftermath (that goes to MAIHT3K), but it does give another perspective on it.
I can see tanteās point. Besides AI datacenters being used for surveillance tech, I can also see LLM tech itself used nefariously post-bubble. Maybe maintaining an up-to-date LLM as a product is not viable, but a custom-trained model to snipe public online discourse around a crucial election could remain affordable for a wealthy fascist.
On the bright side, I am hoping for a brief period of powerful yet affordable gaming PCs thanks to retrofitted, slightly singed Blackwells.
I mean, what has crypto left over? All of that is only useful for sanctions busting and money laundering
deleted by creator
LessWronger puts in the work and determines that LLMs canāt spacially vizualize for shit, comments are like āwell youāre prompting it wrongā (paraphrased) as well as āwhy not pay experiences machinists to videotape what theyāre doing os their work can be automatedā
The article itself is worth reading for some insights in the challenges of using current āAIā (LLMs) to work in the real world.
āwhy not pay experiences machinists to videotape what theyāre doing os their work can be automatedā
none of these people had a job in manufacturing, or even possibly any real job, itās called CNC and there are way better tools for making it work than stochastic parrot
e: ok at least op is a machinist
Interestingly, many of these models also score at or above the level of some human experts on visual reasoning benchmarks like MMMU. That which is easy to measure often doesnāt correlate with real world usefulness.
benchmarks could be perhaps possibly fucked with? say it aint so!
My high level impression when reading the response is āsomeone who can parrot textbook knowledge but doesnāt know what theyāre talking aboutā.
with corpus consisting of toy problems only, and with solutions that depending on field can kill whoever uses these
The fact this commenter doesnāt mention the absolute fuckton of machining videos on Youtube tells me theyāre just talking out of their ass.
(not that forcing those into LLM slop would help but a hypothetical AGI would learn a lot)
Also
Dario Amodei, CEO of Anthropic, recently worried about a world where only 30% of jobs become automated, leading to class tensions between the automated and non-automated. Instead, he predicts that nearly all jobs will be automated simultaneously, putting everyone āin the same boat.ā
Dario is delusional. We donāt even have self driving cars outside of a very small, very expensive demo in SF
@gerikson Iād like to see him automating bed-turning a frail 90 year old in a nursing home so she doesnāt get bed sores (ulcersāopen wounds from lying on a creased sheet or just in the same position for too long). A 90yo with cognitive impairment whoās scared of robots.
@cstross @gerikson This doesnāt sound promising: āA growing body of evidence is finding that robots tend to end up creating more work for caregiversā
https://www.technologyreview.com/2023/01/09/1065135/japan-automating-eldercare-robots/
amazement at chatbots comes from demographic that did not binge watch how itās made in middle school
donāt want to spam the stubsack with US politics but guys I am awfully scared
It really looks like itās on an awful trajectory.
In my teens I read about how Leo Szilard took a train out of Germany the day after the Nazis took power. Passed the border before border checks had time to come into force. Seemed obvious then, now I am all to aware of the problems of such a āsimpleā plan and the ties that binds you, not least family. And of course not knowing in advance how bad it will be, until after. And not knowing if you jump from the ashes and land in the fire, lots of countries are on the same trajectory but further back. Fascism is yet again the choice, the owners choice in the face of climate change.
Iām rambling and itās late. Sympathies and solidarity.
Yeah, it is fucking scary. My sympathies and solidarity. Small thing, perhaps a good reminder for everybody to check their opsec (esp if you are an administrator of things, check which data you do not need, or should not fall into the current (or future) us/other fascist administrations hands. And remember while spying on Americans by the various orgs is illegal, trading for information on them with other countries is not. Donāt forget backups). It is horrible that it has come to this. One small point of light is that they are fools and canāt shut up or be subtle. At least that has a chance to motivate more people to do something, which if it gets to the worst (and it is getting close) more people will actually resist (sadly a lot of people will have to realize that the point becomes not to win, but to impose costs/friction, and any wins are a bonus, which is a horrible realization in itself). Really hope this doesnāt make things worse mentally btw.
I can only offer my solidarity and sympathies. If I were still living in the US, Iād probably have started taking firearms training and bought some by now.
I hear you! If like me you have ever worried that you are overreacting; know that you are not. I have no idea how so many of my coworkers can just treat this all as politics as usual.
My heart breaks for the civilians rounded up and treated as terrorists without due process.
Being transgender I am acutely aware of how close I am to the top of the US autocracyās public enemy list. This is self serving but since January Iāve been working on getting while the getting is good (plus, I hate that my taxes go to the current federal government).
Me and my three siblings (a librarian, a researcher, and a med student) are all worried for our futures for different reasons; which is statistically just kind of impressive!
Solidarity from TN
Noted sex pest Andrew Cuomo wants to run for Mayor in New York, and he might actually have a chance given the incumbent Democratic candidate is corruption magnet Eric Adams. Housing is a big issue in New York, so whatās Cuomoās plan? Well, his plan is to use chat gpt to write his plan.
Angry New York Democrats were using the slogan DREAM meaning Donāt Rank Eric Adams for Mayor. It was then amended to Donāt Rank Eric or Andrew for Mayor, and amended again to Donāt Rank Evil Andrew for Mayor.
Update: another acronym.
eric adams should change his platform to āIām not cuomoā
I dunno, howās the āIām not Donald Trumpā platform been working out so far?
Well. I donāt want adams to win
COMMERCIAL NOTICE: Pivot to AI now has T-shirts! yeah, I sorted out Redbubble. Also does mugs and stuff. All proceeds go to a worthy cause, i.e. me.
New piece from Ed Zitron - a 44-minute read about OpenAIās threat to the entire tech industry
If OpenAIās funding round with Softbank goes as planned, itāll raise the equivalent of the entire GDP of Estonia ā a fairly wealthy country itself, and one thatās also a member of Nato and the European Union. That alone should give you a sense of the truly insane scale of this.
so not only saltman is capable of burning enough power to be comparable to middle-sized euro country, he can also burn small euro country gdp
they are not serious people. damn if i only started grifting instead of getting socially useful skillset
Link to vote Ed for eternal podcast president:
eek
I was reminded of Risky Business and how some of the cybersecurity is sometimes relevant to this sub, which reminded me they talked about this Cryptocurrency people being actually rubber hosed
Today in relevant skeets:
transcript
Skeet: If you can clock who this is meant to be instantly you are on the computer the perfect amount. Youāre doing fine donāt even worry about it.
Quoted skeet: āWhy are high fertility people always so weird?ā A weekend with the pronatalists
Image: Egghead Jr. and Miss Prissy from Looney Tunes Foghorn Leghorn shorts.
Word of advice to the press: Stop š giving š natalists š free š platforms š
Every 3 months! Every 3 months fucking hell. (I know because last time they came in the news I opened a tab with them in it, and their faces was the last I saw before my monitor exploded, like literally suddenly I sat in darkness.
All Iāve learned from the most recent round of publicity is that herself has a new hat. It looks stupid
A new hat let me check almost clicks wait a minute, I canāt afford a new monitor, you are trying to trick me!
Just close your eyes and imagine a mediaeval yokel who has accused her sisterās girlfriend of being a witch so she can take over her turnip patch. No need for a new monitor
Infamous Dr Who ābig name fanā Ian Levine is using generative AI to recreate lost episodes of the show. This got some mainstream press coverage and fans seem miffed.
christ. canāt he go back to disco, he was good at that
I think Ian Levine loves Doctor Who more than he loves disco. Thatās the only explanation for Doctor In Distress.
I got a spam message with a phishing linkā¦ Via Github? Seriously? Are we really doing this?
Not a completely unusual commentā¦ From the URL it was very obvious that this was a phishing link though. Curiosity got the better of me. The site shows you a ācloudflareā captcha. OK, letās click the checkbox. The usual loading animation starts, then this is shown:
Yeah ok, rightā¦
Iām actually a bit impressed with this, these captchas are so common, I didnāt even really think about checking the box. But of course, that interaction means the browser will allow the site to add something to your clipboard.
But likeā¦ Why distribute it via Github? I cannot think of a worse audience to try and con into āpaste something random into your windows consoleā. Am I just being naive here? Is this something common I somehow never experienced before?
Yeah that copy paste commands that give control over your pc shit is pretty nasty. Also over a year old at least. So it isnt specifically an attack at vibe coders. Not sure what the initial targets were but these usually try to target people not that knowledgeable in computer matters so lol that this now includes vibe coders.
Ah, I had been wondering if this is new, or I had just never noticed it. So itās the latter, good to know.
Itās kind of genius as wellā¦ A single comment will get the attention of potentially dozens of people, sent to a valid email address without having to guess/buy lists, and thereās an air of ātrustā around the (completely legit) mail you then get from github, containing the link.
I donāt know how new it is, but it first dropped on my radar about a year ago due to listening to the Risky Business cybersecurity podcast, not to be confused with the recent (and baffingly named (*)) podcast called āRisky Business with Nate Silver and Maria Konnikovaā (**) by dweeb Nate Silver. So I donāt know how long it was going on the wild, and im talking about the windows button + r attack method and not the github comments, no idea how long they used comments as a vector. And yes that part is also good, like the addition of trust of github + quite an effective attack is clever. Shouldnāt work on Real Nerds however.
*: The name means that at least one of they didnāt [know|care|google] about the decades old cybersecurity podcast before naming their podcast that is true. Any of those is odd.
**: addition to above, the tagline of the podcast is āa weekly podcast about making better decisionsā Look inwards Nate, look inwards.
Getting in early on targeting the vibe coder demographic.
Oh god
Althoughā¦ Do you think VideCodersTM read github issues?
Github only catches strays, itās much more widely deployed
Yeah, that this happens elsewhere I have no doubt, I just never noticed on GH before.
āInference Magazine,ā a substack written by a young feller named āWisemanā who, in the most recent article, says ānuh-uh, youāre the parrotā to the work of Dr. Bender and co (and the rest of humanity, by extension).
https://inferencemagazine.substack.com/
The hackernews thread is a real Bad Philosophy turkey shoot:
These people need to sit through a college level class on linguistics or something like that. This is a demonstration of why STEM majors need general higher education.
Orange site really is out here reinventing hard behaviorism.
āWe canāt directly observe internal states beyond our own subjectivityā -> Letās try to ignore them and see what we get" -> āWeāve developed a model that doesnāt feature internal states as a meaningful element of cognitionā -> āthere are no internal statesā -> āI know Iām a stochastic parrot but what are you?ā
Damn well put!
some parrots are more stochastic than others
Serious question: what are peopleās specific predictions for the coming VC bubble popping/crash/AI winter? (Iāve seen that prediction here before, and overall I agree, but Iām not sure about specificsā¦)
For exampleā¦ Iāve seen speculation that giving up on the massive training runs could free up compute and cause costs to drop which the more streamlined and pragmatic GenAI companies could use to pivot to providing their āservicesā at sustainable rates (and the price of GPUs would drop to the relief of gamers everywhere). Alternatively, maybe the bubble bursting screws up the GPU producers and cloud service providers as well and the costs on compute and GPUs donāt actually drop that much if any?
Maybe the bubble bursting makes management stop pushing stuff like vibe codingā¦ but maybe enough programmers have gotten into the habit of using LLMs for boilerplate that it doesnāt go away, and LLM tools and plugins persist to make code shittery.
Iāve repeated this prediction a bajillion times, but I suspect this bubbleās discredited the idea of artificial intelligence, and expect it to quickly die once this bubble bursts.
Between the terabytes upon terabytes of digital mediocrity the slop-namiās given us, LLMsā countless and relentless failures in logic and reason, the large-scale enshittification of daily life their mere existence has enabled, and their power consumption singlehandedly accelerating the climate crisis, I feel that the publicās come to view computers as inherently incapable of humanlike cognition/creativity, no matter how many gigawatts they consume or oceans they boil.
Expanding on this somewhat, I suspect AI as a concept will likely also come to be seen as an inherently fascist concept.
With the current bubbleās link to esoteric fascism, the far-rightās open adoration of slop, basically everything about OpenAIās Studio Ghibli slopgen, and God-knows-what-else, the publicās got plenty of reason to treat use or support of AI as a severe indictment of someoneās character in and of itself - a ātech asshole signifierā, to quote Baldur Bjarnason.
And, of course, AI as a concept will probably come to be viewed as inherently anti-art/anti-artist as well - considering how badly the AI bubbleās shafted artists, and artists specifically, that kinda goes without saying.
I think you are much more optimistic than me about the general publicās ability to intellectually understand fascism or think about copyright or give artists their appropriate credit. To most people that know about image gen, itās a fun toy: throw in some words and rapidly get pictures. The most I hope for is that AI image generation becomes unacceptable to use in professional or serious settings and it is relegated to a similar status as clip art.
I think weāre going to see an ongoing level of AI-enabled crapification for coding and especially for spam. Iām guessing thereās going to be enough money from the spam markets to support a level of continued development to keep up to date with new languages and whatever paradigms are in vogue, so vibe coding is probably going to stick around on some level, but I doubt weāre going to see major pushes.
One thing that this has shown is how much of internet content ācreationā and ācommunicationā is done entirely for its own sake or to satisfy some kind of algorithm or metric. If nobody cares whether it actually gets read then it makes economic sense to automate the writing as much as possible, and apparently LLMs represent a āgood enoughā ability to do that for plausible deniability and staving off existential dread in the email mines.
Yeah I also worry the slop and spam is here to stay, itās easy enough to make, of as passable quality for the garbage uses people want from it, and if GPUs/compute go down in price, affordable enough for the spammers and account boosters and karma farmers and such to keep using it.