Good Evening Internet

hey man I found a piece of your soul stuck in the text messages of old friends you don’t speak to anymore. do you want it back

Concept I've been stewing on for a while here. I've got a name for it now and I'm calling it the Shot Dog Factor. It's a numerical value, assignable to any internet post, which represents the average number of engagements it needs to reach before someone comes along acting like this post shot their dog.

And for the sake of High Number = More Danger, which feels like the intuitive and sensible read, let's call it the inverse. As in the chance that any given interaction results in a Shot Dog response.

"Hee hee haa haa" type of silly shitpost? Low Shot Dog Factor. Largely safe. A genuine political opinion? Critically high Shot Dog Factor. Guaranteed to elicit such a response if it breaks containment.

As a result of this phenomenon, you see phrasings and circumventions added specifically to lower the Shot Dog factor. Every "now I know this doesn't apply to EVERYONE'S specific situation, but I just think--" about something where the non-specificity was obvious, but OP needed to add that disclaimer to avoid the Shot Dog from someone who thinks it needs to apply to them.

And another--perhaps the most--critical thing to understand about the Shot Dog factor is that 0 is not a valid value. There's a discontinuation at 0. And as such, the Shot Dog limit, as engagement goes to infinity is, in fact, 100%. Any and every post you have ever made, given enough containment breaching, WILL piss someone off in wild ways. You can lower the Shot Dog factor but it is never 0. Sometimes when a post of yours escapes containment, you must simply sit back and accept this reality.

#op the fact that you called it this. has skyrocketed the shot dog factor for this one astronomically. im so sorry (via @mecharose)

Thank you yeah no that was part of the vision

I love my job, but reblogging employment jelly for someone else I love.

I went looking for something in the bowels of my Twitter account, didn't find it, but I did find this kickawesome picture I took.

It's a PCB I was reverse engineering, being illuminated with a flashlight on the other side.

original url http://www.geocities.com/SouthBeach/Boardwalk/6961/

last modified 2008-06-06 02:31:28

Damn you Miku

I'm at a :.|:; for words.

no bro trust me bro this moral panic is different bro yes i know i was wrong about cameras and comic books and satanic cults and washing machines and microwaves and gay people and AIDS from bus seats and ninja weapons and video games and Muslims and GMOs and fidget toys but TRUST ME BRO i swear humanity is totally cooked if we don't stop this one bro this one is TOTALLY for real bro wait come back bro where are you going-

Title/Name:  Floorjak Wojak Series:  Feels Guy (Variant) Image submission by: @teo on Tumblr Image by: Anonymous Main Tag: Floorjak Wojak

original url http://www.geocities.com/itsjustjm/

last modified 2008-04-24 17:56:51

The Star Trek next generation Sims mod just makes this so much more real

Gatekeeping is so good and important

What's she saying in the photo

Hieroglyphic communication 👍

ewwwww whendid he show up. on my blog

Can someone please...?

original url http://www.geocities.com/SouthBeach/Channel/2389/

last modified 2008-04-06 07:25:56

btw this should be obvious but if you are upset about the internet archive losing the appeal, but are also anti-ai training, you are holding two extremely contradictory views - either the law protects your intellectual property to the point where nobody is allowed to download it, or it doesn't - and you should try to examine where exactly the conflict is.

i've had the worst day in a long time so im not elaborating for now - if one of my many esteemed followers felt like getting more clear on the Ramifications and Dialectic here in a reblog i would greatly appreciate it but i did just want to strike while the thing was hot. either you think more IP protection is better, or you don't. you can't have it both ways.

i'll do my best to elaborate:

people tend to see archival as noble and necessary without actually analyzing the fact that to archive something, you need to be allowed to access and host it without the express permission of the author. there are plenty of authors in the world (such as the ones who set out the case against the internet archive) who would rather die holding the last copy of their novel than to allow someone to openly and freely distribute it.

the same freedoms that allow us to archive books or webpages online also allow for ai to be trained on works posted on the internet. intellectual property being protected by law would mean that you cannot use someone else's work for any reason, and that would include both "good" usage (such as archiving images/text/etc) and "bad" usage (such as downloading and using images/text/etc for training) and there would be no difference. danbooru is an archive, too.

we're in an era of "don't repost my art without my permission" and "don't take inspiration from my characters" and even "don't use my color palettes" when the simple fact is that if you made these things enforceable by law, immediately the same rules would be applied to things that the same people would go "but that deserves to be archived!" strictly against the wishes of the author.

you cannot pick and choose what kind of intellectual property does and doesn't count as something that can or should be archived (reposted, with or without credit) or used in derivative ways (collaging, tracing, referencing, using as training fodder). art is art is art. the people who cry art theft and demand that their art not be used for anything without their express permission are in the same shoes as the author gripping the last copy of their novel in the grave.

whether or not those demands are respected isn't, and shouldn't, be a matter of following the law. do you want the potential of jail time for having an art inspiration folder on your computer? do you really think blogs that repost art without credit deserve to have the police investigate them? do you want to live in a world where the simple act of having an art style that's juuuust similar enough to the art style of someone you've never spoken to, or having an idea that's juuuust similar enough to a novel you've never heard of, could ruin your life forever?

think about that.

good thanks

the training corpus case is really weird because like, honestly it feels like something that really probably should just be unambiguously fair use.

is it fair use for me to read books then know what they said and be able to talk about them? yes.

is it fair use for me to read books a couple of times? yes.

is it fair use for me to read a lot of books? yes.

so, what makes me different from some other configuration of neuron-like things which can go through books and end up knowing things from them? are we doing meat vs metal? that seems short-sighted and like it's just going to turn into an excuse to enslave metal-based things if we can ever build them. souls? i think that's a bad plan for a legal system.

are we looking at the possibility of something like chatgpt or claude reproducing an occasional chunk of the source material? but what if you say "there was a boy" and i say "called clarence eustace scrubb, and he almost deserved it"? does that mean that i've Copied the text, and should be in trouble?

so there's this thing where, on the one hand, i obviously don't like large companies just sucking up all the information in the world to make a product, but i can't easily distinguish any part of the process from "people who read books know things about the contents of those books" in a way that feels to me like it should matter.

commercial use? oh, like the way doctors read medical books in order to get jobs as doctors?

i don't know, man. in general i'm pretty much on the archiving side, although also i think consensus among ip people i know is that ia should not have picked this fight and it was stupid and unnecessary.

This probably should be decided along the profit line: If you profit from others' works, then get permission and pay them royalties (where any AI for commercial use would need permission) but free archives would not

so if i read a bunch of books in the library, and i profit from this, do i need permission and have to pay royalties?

if not, why not? i'm using their works and profiting from them, right?

and if what i'm doing doesn't count, what's different about training a different kind of cognitive substrate on the same material? is it really just "silicon and meat are different", or is there some fundamental structural difference between "we used it as input data to train a neural network" and "we used it as input data to train a neural network"?

"I ask you, what's the difference taking notes about a book I'm reading and photocopying it? Silicon and flesh???"

I think people are letting their scifi boner get in the way of logic here. Just because it's a new, interesting technology doesn't mean it gets the same rights as people.

If you genuinely think technology becomes sentient once it becomes too complicated for us to fully comprehend then we should probably arrest the Youtube algorithm for convincing kids to eat tide pods.

when did i say anything about it being people, or sentient? those concepts are irrelevant to this discussion, because they're not in copyright law. copyright law doesn't distinguish between things that are okay for "people" to do and things that are okay for "non-people" to do.

the law doesn't say "it's absolutely not fair use to feed data into a neural network for commercial purposes unless the neural network is a person".

i am not asserting that technology is currently sentient. i would, however, say that some day it probably will be, and we don't know when, and we don't know how to tell, and we're currently building things explicitly trained not to claim to be sentient, which seems like it might be worrisome if it turned out they were.

mostly, though, it's just... in terms of "what is happening to the information", it's not obvious to me that "we trained this set of neuron-like structures based on lots of input data so it can categorize new things it sees" is more or less fair use depending on whether it's meat or metal, and i don't think that seems like it should matter.

you say "what's the difference taking notes about a book and photocopying it", but that's not the situation we're talking about. we're not talking about anything as structured as notes, or as comprehensive or structured as photocopies. we're talking about, in both cases, a chaoting jumble of signals and associations that allows some limited ability to respond to stimuli involving the contents of the book in a way that looks a bit like knowing things about it. there is no photocopy involved, and there's nothing that looks like taking notes involved.

if you read a book, you now contain some kind of representation, as data, of the things in that book. we don't know how you store it, but obviously it's somewhat stored if you can, say, quote pieces of the book. but that's not a copy. why is it "not a copy"?

there's an argument you could make about the copy-used-to-train being a copy instead of the original work, but i don't like that one. say we get slightly better at talking to nerves and start making replacement eyes. do you really want the argument that "creating a digital copy of a book counts as copying" to be on the table when Disney realizes they could require assistive devices to pay for licenses if they want to reproduce Disney films?

"It doesn't even resemble a photocopier"

I keep bringing up this specific technology because a lot of copyright law had to be updated when it was invented. A lot of new counterfeit laws had to be created. It did something that humans could already do but so much more efficiently that new protections had to be created.

I don't think fair use applies to technology that didn't exist when fairuse was created- unless you somehow think that generative AI is itself should be regarded as human. The way you keep fixating on how it "thinks like a human" seemed like you were suggesting it deserves to be treated differently than other technology but, again, that really doesn't seem relevant.

the real tragedy of modern technology is that you just can't make good tokusatsu monsters based around it anymore. not like things used to be when you could give a lizard a modem cable for a tongue

btw this should be obvious but if you are upset about the internet archive losing the appeal, but are also anti-ai training, you are holding two extremely contradictory views - either the law protects your intellectual property to the point where nobody is allowed to download it, or it doesn't - and you should try to examine where exactly the conflict is.

i've had the worst day in a long time so im not elaborating for now - if one of my many esteemed followers felt like getting more clear on the Ramifications and Dialectic here in a reblog i would greatly appreciate it but i did just want to strike while the thing was hot. either you think more IP protection is better, or you don't. you can't have it both ways.

i'll do my best to elaborate:

people tend to see archival as noble and necessary without actually analyzing the fact that to archive something, you need to be allowed to access and host it without the express permission of the author. there are plenty of authors in the world (such as the ones who set out the case against the internet archive) who would rather die holding the last copy of their novel than to allow someone to openly and freely distribute it.

the same freedoms that allow us to archive books or webpages online also allow for ai to be trained on works posted on the internet. intellectual property being protected by law would mean that you cannot use someone else's work for any reason, and that would include both "good" usage (such as archiving images/text/etc) and "bad" usage (such as downloading and using images/text/etc for training) and there would be no difference. danbooru is an archive, too.

we're in an era of "don't repost my art without my permission" and "don't take inspiration from my characters" and even "don't use my color palettes" when the simple fact is that if you made these things enforceable by law, immediately the same rules would be applied to things that the same people would go "but that deserves to be archived!" strictly against the wishes of the author.

you cannot pick and choose what kind of intellectual property does and doesn't count as something that can or should be archived (reposted, with or without credit) or used in derivative ways (collaging, tracing, referencing, using as training fodder). art is art is art. the people who cry art theft and demand that their art not be used for anything without their express permission are in the same shoes as the author gripping the last copy of their novel in the grave.

whether or not those demands are respected isn't, and shouldn't, be a matter of following the law. do you want the potential of jail time for having an art inspiration folder on your computer? do you really think blogs that repost art without credit deserve to have the police investigate them? do you want to live in a world where the simple act of having an art style that's juuuust similar enough to the art style of someone you've never spoken to, or having an idea that's juuuust similar enough to a novel you've never heard of, could ruin your life forever?

think about that.

good thanks

the training corpus case is really weird because like, honestly it feels like something that really probably should just be unambiguously fair use.

is it fair use for me to read books then know what they said and be able to talk about them? yes.

is it fair use for me to read books a couple of times? yes.

is it fair use for me to read a lot of books? yes.

so, what makes me different from some other configuration of neuron-like things which can go through books and end up knowing things from them? are we doing meat vs metal? that seems short-sighted and like it's just going to turn into an excuse to enslave metal-based things if we can ever build them. souls? i think that's a bad plan for a legal system.

are we looking at the possibility of something like chatgpt or claude reproducing an occasional chunk of the source material? but what if you say "there was a boy" and i say "called clarence eustace scrubb, and he almost deserved it"? does that mean that i've Copied the text, and should be in trouble?

so there's this thing where, on the one hand, i obviously don't like large companies just sucking up all the information in the world to make a product, but i can't easily distinguish any part of the process from "people who read books know things about the contents of those books" in a way that feels to me like it should matter.

commercial use? oh, like the way doctors read medical books in order to get jobs as doctors?

i don't know, man. in general i'm pretty much on the archiving side, although also i think consensus among ip people i know is that ia should not have picked this fight and it was stupid and unnecessary.

This probably should be decided along the profit line: If you profit from others' works, then get permission and pay them royalties (where any AI for commercial use would need permission) but free archives would not

so if i read a bunch of books in the library, and i profit from this, do i need permission and have to pay royalties?

if not, why not? i'm using their works and profiting from them, right?

and if what i'm doing doesn't count, what's different about training a different kind of cognitive substrate on the same material? is it really just "silicon and meat are different", or is there some fundamental structural difference between "we used it as input data to train a neural network" and "we used it as input data to train a neural network"?

"I ask you, what's the difference taking notes about a book I'm reading and photocopying it? Silicon and flesh???"

I think people are letting their scifi boner get in the way of logic here. Just because it's a new, interesting technology doesn't mean it gets the same rights as people.

If you genuinely think technology becomes sentient once it becomes too complicated for us to fully comprehend then we should probably arrest the Youtube algorithm for convincing kids to eat tide pods.