ICYMI: This is one petition Canadians do NOT want to miss. This is not your average Change-dot-org petition. This is the kind that could lead to serious changes against paypro censorship www.ourcommons.ca/petitions/en...
An upcoming 'tiel 'tuber exploring some fresh air. Highly sussy, 18+ only
592 followers 1,153 following 1,121 posts
view profile on Bluesky
ICYMI: This is one petition Canadians do NOT want to miss. This is not your average Change-dot-org petition. This is the kind that could lead to serious changes against paypro censorship www.ourcommons.ca/petitions/en...
Trump trying to find a recharger for his body? I'm pretty sure that's the least of his health issues
That's probably an interesting enough but to be worthy of someone looking into the codebase and running a debugger to figure out just what is happening internally
What happened to make sudo take forever?
Is regulating the quantity of money that is permitted be used to do so a restriction?
They can't possibly watch us all right now
I cannot certify a design as idiot-proof, only as idiot-resistant
Even physics is not immune en.wikipedia.org/wiki/Action_...
Unfortunately the current chartered one is at best generously described as a complete and utter mess
reconstructing boobs from first principles (SFW female + male nipple)
> do not cover this hole
Please elaborate
Do you happen to have internal metrics about how long it takes for you to come up with responses?
it's not all fascists working on AI, it's just that if you work on AI and post here you gotta keep quiet about it I wouldn't be surprised if several of the big name trans posters who quietly work in tech actually do work on LLMs though, and just don't talk about it
Best I can do is 1 million triangles
Also yeah composable moderation sounds nice but when can I compose exceptions on top of a moderation list
*lives life, with weekly checks to see if godot has arrived yet*
"just make it exist first" you won't be mad if i use ai to do it, right? 🥺
One thing i think people are unaware of is just how much GPU is used in modern applications
Meanwhile, me munching french toast from behind the safety of being a mere bystander:
I understand a lot of people use reasoning like “I don’t trust big tech companies, how do I know that their models work the way they say they do” the answer to this is that, while the absolute current version of the big commercial chatbots is generally closed source, you can download recent models
attention-based transformer models were not developed with the goal of "we want to make an AI therapist" or "we want to make an AI girlfriend" or "we want to figure out how to supplant white-collar professionals" all of those things are attempts to find a market application for an algorithm
So... anyone happen to have a sexy butler outfit reference lying around?
Is this from you or the agent
Sometimes it's fine to forget
I'm pretty sure it's entirely feasible to criticize techbruhs without resorting to falsehoods, so why is the one that does the one that gets traction
I've noticed this too
📌
(no I do not have the money or the artist I want to comm lying around right now unfortunately)
So... anyone happen to have a sexy butler outfit reference lying around?
Please make more sexy clothes for guys that aren't just traditional femme clothes on put a male body. The cuts are never right and it doesn't look good, imho. Same with sexy poses. Though tbf I don't usually like traditional femme clothes on female bodies either so maybe my opinion doesn't count...
I personally like my founding documents with a nice bit of charring
You seem to certainly be AI assisted
Do the LLMs make any more sense of the docs than us humans?
"we have made up a slur to insult LLMs with" newspaper: "people make up slur for LLMs" "why are you calling it a slur"
Some google engineer: "Okay, so some users have gone out of their way to give feedback into the search bar. How can we make use of such feedback?"
I think we need grass widely available to make that widely applicable Probably should be native grasses to avoid undue maintenance butden
The Kant Car
Has anyone ever explained what the supposed negative externalities of text prediction or image generation existing that are even as comparable as the Internet existing.
u dont have to like ai or whatever but people always overreact. like this doesnt do anything unless you choose a chatbot (you don't have to! no one is forcing you to)
Definition of AI as inferred from usage: Computer operation that is insufficiently understood to be called an algorithm
Fun fact, leaving a comment boosts your engagement with that content, making it more likely for that creator to be recommended to you. I am also a menace to society
People mad at people “offloading their thinking to AI” are offloading their thinking to dumbass blocklists
The commonly cited AI water usage statistics are based on a methodology where, if your data center buys green energy from a hydroelectric dam (as lots do, including the hyperscalers), then “natural evaporation of the water from the lake behind the dam” is counted as that data center’s water use.
I've been added to two contradictory ones which makes me think the tripwires are just interacting with the wrong accounts on two sides of an onsite argument
How do I set up a caching proxy for the aur
Old Soviet joke for today: A man walks into a newsstand every day, looks around, and leaves. After a long time of this, the owner says “Can I help you find something?” “I’m looking for the obituaries.” “The obituaries are in the back of the newspaper, comrade.” “Not the one I’m looking for.”
I prefer checking the newsstands daily for front page obituaries
🤔
Also do the world models cover touch?
We still need ones for balance, and for the variety of internal monitoring senses such as bodily position and hunger
Do the different settings change the contents of the generated video?
My mind jumped to the idea of some people who stubbornly stick it out in an area despite it being already more than a meter deep below sea level and beyond the current coastline, and regularly battered by storms
To elaborate: The root cause of > 99% the cases of "ChatGPT madness" are that ChatGPT was extremely willing to entertain delusions and exacerbate existing issues. Just not being that way would have prevented those. (By the way, OpenAI has already deployed such adjustments.)
People trying to come up with all these safety checks to prevent issues, when all you really need is for the model to have it drilled into its weights that the user might be a numbnut
(the bar for being less bad is very much underground at this point)
It'll be bad, but a lesser bad than the government activity it holds back
📌
If i didnt say AI, but i told you there was a system that could read 100 pages of a pdf and pull out structured output 3 years ago, and it takes seconds to set up. people wouldve lost their mind, but its so wrapped up in identity now
Hmm... is there a way to apply line search when you don't have the actual function (or a good approximation) close at hand?
Line search time? If the new point is worse than the old one, back up until it isn't worse
Fair, though it'll probably take more than a few stove-burned hands before it gets resolved
For exorcise?
honestly its the pro-copyright RIAA shilling that pisses me off i don't actually care about any of the AI companies. i think the big companies have their own stupidity but defend copyright and you are my enemy for life copyright should not only be abolished but treated as a stain on human history
The human body already has an integrated system for self monitoring, even if it's flawed (just ask me and my response to an allergen 8 hours after I ate a tiny bit of it)
That one seems like it has an obvious response: Mere sensors that don't protrude are allowed everywhere, but body augmentations require case-by-case preapprovals
Bios update done
why do people still think AI detection—in software or in your head—is actually real and reliable? People are absolutely sure they can identify AI based on one type of sentence or another and that just doesn’t seem true at all.
Also I'm pretty sure this runs right into a violation of EU regulations
"Linux is free if you don't value your time." I dealt with a USB issue for two years that persisted through Windows reinstalls. My USB devices would conk out, get disconnected, come back ten seconds later. 1-2 times a day. Nothing was ever in the system logs and sifting through useless bot-written
There has got to be a case where someone has a hoarding disorder involving guns
What would they be looking for, anyways? "Okay, fine, I'm paying for the book that I pirated a few years ago. Good luck with fighting for anything more than that in court."
most thorough water use in ai article so far
chatgpt ≠ ai AI/LLMs as a technology can be very good, but specific companies and productized versions of this technology can do great harm to individuals and society anti-AI proponents are foolishly mirroring the spirit/energy behind the recent age verification and internet censorship laws
People trying to come up with all these safety checks to prevent issues, when all you really need is for the model to have it drilled into its weights that the user might be a numbnut
there is, unfortunately, not a clean dividing line between the AI that does stuff we hate and the load-bearing infrastructure of the modern internet
one of the wildest bits of discourse, is the idea that we can just ban ai we cannot ban ai. code is speech and you can run these models on a 4080. there's no putting this genie back in the bottle.
Wire cutters are a lot more reliable at shutting down misbehaving electronics
Took all day, but I finally got a systemd socket activated server working
Reason #893 why it's funny when people try to delineate between the "Good AI" (traditional classification, which necessarily includes facial recognition) and "Bad AI" (LLMs, LMMs).
Comments are skeptical this is possible. But it's already happening: many orgs are trying to undermine other players' market position by releasing something open-weight. Some are fully open. There are also ways to *shape* the way the tech develops without training a 70B model from scratch.
they *already* can't tell you how to kys, you have to push the conversation meta by making it about a story you're writing. this is just an inherent vulnerability in language use. fucking human beings are vulnerable to it! any language powerful enough to be useful will always have rules bypasses
a) companies should not produce sycophantic AI slaves that validate your every eccentricity and maladaption. b) people demand sycophantic AI slaves that validate their every eccentricity and maldaption. b) regulating LLM personality or wokeness is a minefield that will usually end in censorship.
Here's my best attempt at designing a DIY michelson-morley experiment: Mount two laser pointers onto a large turntable at 90 degrees, so that they converge at a rough surface. Place a video camera between them pointing at this point. Rotate the turntable gently. Repeat 3 months later.
Because "Skype for Business" didn't sound corporate enough
@deertexting.bsky.social
Hard to argue with his dad at this point.
How much intent to be blocked was behind that post Was it "wait wtf", "whatever, to be expected", or "good (grumpy cat impression)."
This is straight up just not true. Companies which only serve models are profitable because there's more demand for models than capacity. Companies like Groq, Fireworks and Together (which don't train models) make a profit from customer revenue. sacra.com/c/groq/
Probably because Newsom is just plain louder and more annoying about his resistance
I think he just gives himself permission without asking anyone else
@dollspace.gay I remember you have a prompt that you designed for a similar purpose
For concrete comparisons of how much cheaper models are: Nowadays, you can get a model with about the same number of parameters and level of performance as GPT-2 in a manner of minutes, and will cost you around a dollar in on-demand GPU compute. github.com/KellerJordan...
He'd probably try to figure out how to get a 3d printer or cnc machine to make them You know, means of production and all
Professional 3d rendering, video processing, scientific simulations:
*Year 8 of being dismayed that d.io is not a jojoke website*