@PupBiru@kbin.social avatar

PupBiru

@PupBiru@kbin.social

This profile is from a federated server and may be incomplete. Browse more on the original instance.

PupBiru, (edited )
@PupBiru@kbin.social avatar

kilobyte (KB) is 1000, kibibyte (KiB) is 1024

at least according the the IEC, and id tend to go with them… SI units say that kilo means 1000

PupBiru,
@PupBiru@kbin.social avatar

which is why we have kibi, mebi, gibi, etc

PupBiru, (edited )
@PupBiru@kbin.social avatar

not exactly because of pairs unless you’re talking about 1 and 0 being a pair… it’s because the maximum number you can count in binary doubles with each additional bit you add:

with 1 bit, you can either have 0 or 1… which is, unsurprisingly perhaps, 0 and 1 respectively - 2 numbers

with 2 bits you can have 00, 01, 10, 11… which is 0, 1, 2, 3 - 4 numbers

with 3 bits you can have 000, 001, 010, 011, 100, 101, 110, 111… which is 0 to 7- 8 numbers

so you see the pattern: add a bit, double the number you can count to… this is the “2 to the power of” that you might see: with 8 bits (a byte) you can count from 0 to 255 - that’s 2 (because binary has 2 possible states per digit) to the power of 8 (because 8 digits); 8^2

the same is true of decimal, but instead of to the 2 to the power, it’s 10 to the power: with each additional digit, you can count 10 x as many numbers - 0-9 for 1 digit, 00-99 for 2 digits, 000-999 for 3 digits - 10^1, 10^2, 10^3 respectively

and that’s the reason we use hexadecimal sometimes too! we group bits into groups of 8 and call it a byte… hexadecimal is base 16, so nicely lets us represent a byte with just 2 characters - 16^2 = 256 = 2^8

PupBiru,
@PupBiru@kbin.social avatar

i didn’t downvote you, and i went to school before a bunch of things but technology evolves and either we evolve with it or we end up being just straight up wrong in a modern context

PupBiru,
@PupBiru@kbin.social avatar

let’s not go too far though… the holders of h264/h265 did put a lot of money and effort into developing the codec: a new actual thing… they are not patent trolls, who by definition produce nothing new other than legal mess

PupBiru,
@PupBiru@kbin.social avatar

afaik activitypub/fediverse doesn’t have to be fully open… there’s private messages and followers only profiles on mastodon… sure, any server admins of your followed would be able to see anything you post (and thus in this case for threads for example, if you accept any follower from threads then meta can see your stuff) but this also doesn’t grant them a license to use the content

also, bluesky will eventually be the same: it only doesn’t have those issues now because they haven’t opened up their software… it’ll have federation in the future, which means it has to be somewhat programmatically open to others

PupBiru,
@PupBiru@kbin.social avatar

i feel like i need to preface this comment with the fact that this is undeniably a bad thing and no amount of “but on the flip side” will change that, but it’s interesting to express regardless…

this could lead to a few interesting situations:

  • more ubiquitous ML could lead to enforcement of laws more evenly… ML doesn’t make “oh sorry sir i didn’t know who you were” decisions, and if that’s coupled with transparency then maybe we will be left in less of a “laws for thee and not for me” situation as it becomes more difficult to break laws for people in power
  • more ubiquitous ML, as long as it’s fairly openly available, will absolutely be used by media to piece together complex structures and do investigative journalism. it could help to hold people to account
  • more ML in tax could mean less tax evasion? or setting it to task on suggesting fixes for tax loop holes if it can see a lot more invasive data?
PupBiru,
@PupBiru@kbin.social avatar

and experience is ongoing learning, so if an LLM were training on things after the pretraining period then that’d allow it to be creative in your definition?

but in that case, what’s the difference between doing that all at once, and doing it over a period of time?

experience is just tweaking your neurons to make new/different connections

PupBiru,
@PupBiru@kbin.social avatar

it’s only qualitative because we don’t understand it

when an LLM “experiences” new data via training, that’s subjective too: it works its way through the network in a manner that’s different depending on what came before it… if different training data came before it, the network would look differently and the data would change the network as a whole in a different way

PupBiru,
@PupBiru@kbin.social avatar

that’s a lack of understanding of concepts though, rather than a lack of creativity… curation requires that you understand the concept that you’re trying to curate: this looks more like a dog than this; this is a more attractive sunset than this

current LLMs and ML don’t understand concepts, which is their main issue

id argue that it kind of does “think about its own thoughts” to some degree: modern ML is layered, and each layer of the net feeds into the next… one layer of the net “thinks about” the “thoughts” of the previous layer. now, it doesn’t do this as a whole but neither do we: memories and neural connections are lossy; heck even creating a creative work isn’t going to turn out exactly like you thought it in your head (your muscle memory and skill level will effect the translation from brain to paper/canvas/screen)

but even we hallucinate in the same way. don’t look at a bike, and then try and draw a bike… you’ll get general things like pedals, wheels, seat, handlebars, but it’ll be all connected wrong. this is a common example people use to show how our brains aren’t as precise and we might like to think… drawing a bike requires a lot of very specific things to be in very specific places and that’s not how our brain remembers the concept of “bike”

PupBiru,
@PupBiru@kbin.social avatar

i don’t agree with that definition of creative… there’s lots of engineering work that’s creative: writing code and designing systems can be a very creative process, but doesn’t involve feeling… it’s problem solving, and thats a creative process. you’re narrowly defining creativity as artistic expression of emotion, however there’s lots of ways to be creative

now, i think thats a bit of a strawman (so i’ll elaborate on the broader point), but i think its important to define terms

i agree we should be skeptical of marketing hype for sure: the type of creativity that i believe ML is currently capable of is directionless. it doesn’t understand what it’s creating… but the truth lies somewhere in the middle

ML is definitively creating something new that didn’t exist before (in fact i’d say that its trouble with hallucinations of language are a good example of that: it certainly didn’t copy those characters/words from anywhere!)… this fits the easiest definition of creative: marked by the ability or power to create

the far more difficult definition is: having the quality of something created rather than imitated

the key here being “rather than imitated” which is a really hard thing to prove, even for humans! which is why our copyright laws basically say that if you have evidence that you created something first, you pretty much win: we don’t really try to decide whether something was created or imitated

with things like transformative works or things that are similar, it’s a bit more of a grey area… but the argument isn’t about whether something is an imitation; rather it’s argued about how different the work is from the original

PupBiru,
@PupBiru@kbin.social avatar

from what i understand, solid state batteries are legitimately about as revolutionary as lithium ion were because they are all of those things, and by their very nature they have a huge number of charge cycles

… whether this specific announcement results in a mass-production-capable battery is another story

PupBiru,
@PupBiru@kbin.social avatar

in fact it’s literally the basis for the “tools” functionality in the new openai/chatgpt stuff!

that “browse the web”, “execute code”, etc is all the LLM formatting things in a specific way

PupBiru,
@PupBiru@kbin.social avatar

lemmy doesn’t really have a TOS afaik (and even if the software had something in its license, it’s the fediverse: people can interact with you on lemmy without using lemmy!). your instance might, but other instances might have totally different rules… there’s nothing stopping a just-ads-lemmy.com instance, other than it’d probably get defederated pretty quick

PupBiru,
@PupBiru@kbin.social avatar

as a linux professional, congrats you’re a junior and have a lot to learn about the world

AI content generation is on fast track to kill the whole porn industry.

If you think about it, there are already tools that can do pretty convincing face swaps with few clicks and just from one low res photo at whatever angle. Long are gone the times you had to train the models for hours and had to have few hundred photos to just get “okay” fake. Once deepfake video creation becomes this simple,...

PupBiru,
@PupBiru@kbin.social avatar

i make porn

i’ve spent thousands of dollars on camera equipment

i post it online, in 4k, for free, and there’s no way to pay me for the content

i do it because i like the exhibitionism (among other things)

i do it because in the sex-positive communities i prefer to associate with, it’s like posting your day to facebook

i do it because the camera pushes me to imagine new and interesting things to do

i do it because it helps me meet people with similar tastes that aren’t flakey and are equally sex positive and non-judgemental

there’s an enormous amateur community (that often becomes professional) that makes porn because they want to… money is a nice thing to help buy new gear for porn, but it’s not the primary concern

PupBiru,
@PupBiru@kbin.social avatar

nothing is forcing you to pay for every single one, correct… but actively managing subscriptions is a pain, and you don’t get to watch things as they come out so you don’t have that topic of conversation with people (sometimes)

meanwhile, piracy automatically downloads new episodes often mere hours after it’s released, for free, you don’t have to actively manage subscriptions, you don’t have to remember which service has what, there’s no concern that your favourite show is suddenly going to get removed from the platform

i’m happy to pay a subscription fee or 2, however i will absolutely not pay an exorbitant amount for an objectively worse experience… and when it hits that tipping point, well pirate all the things is the go

PupBiru,
@PupBiru@kbin.social avatar

entertainment is mental health and mental health is just as important as physical health… don’t be so damn condescending!

PupBiru,
@PupBiru@kbin.social avatar

mental health doesn’t mean you’re suicidal… maintaining mental health is about your whole life and is something you do every day

PupBiru,
@PupBiru@kbin.social avatar

i wouldn’t say minimal, but it certainly just works with no intervention once you set it up… not torrents; start on usenet and you’ll never look back!

PupBiru,
@PupBiru@kbin.social avatar

couldn’t agree more!

PupBiru,
@PupBiru@kbin.social avatar

we output nothing other than what we’re trained on; the only difference is that we’re allowed to roam the world freely and consume whatever information we stumble on

what you say would be true if the LLM were only trained on content by the author seeking to say that their works had been infringed, however these LLMs include a lot of other data from public domain sources

one could consider these public domain sources and our experience of the world to be synonymous (and if you don’t i’d love to hear the distinction), in which case there’s some kind of a line that you seem to be drawing, and again i’d love to hear where you think that line is

is it just ratio? there’s precedent to that for sure: current law has fair use rules which stipulate things like “amount and substantiality”. in that case the question becomes one of defining the ratio. certainly the ratio of content that the author is referring to vs the content not trained by the author is minuscule

PupBiru,
@PupBiru@kbin.social avatar

it’s nothing like that at all… if someone bought a book and produced a big table of words and the likelihood that the next word would be followed by another word, that’s what we’re talking about: it’s abstract statistics

actually, that’s not even what we’re talking about… we then take that word table and then combine it with hundreds of thousands of other tables until the original is so far from the original as to be completely untraceable back to the original work

PupBiru,
@PupBiru@kbin.social avatar

i think the distinction that either side is seeing here is that you think humans are inherently different to a neural network, where i think that the only difference is in the complexity: that if we had a neural network at the same scale as the human brain, that there’s nothing stopping those electronic neurons from connecting and responding in a way that’s indistinguishable from a human

the fact that we’re not there yet i don’t see as particularly relevant, because we’re talking about concepts rather than specifics… of course a LLM doesn’t display the same characteristics as a human: it’s not of the same scale, and the training is different but functionally there’s nothing different between chemical neurons firing and neurons made of transistors firing

we learn in the same way: by reinforcing connections between our neurons

  • All
  • Subscribed
  • Moderated
  • Favorites
  • All magazines