nah this is bullshit. i actually recently watched a sarah z video essay about exactly why this is bad and it’s a pretty compelling case.

can find it here

in short, humans are shit and the basis on which most people are called a narcissist is more biased and tenuous than most would ever admit which is problematic if we’re going to ruin people’s lives over their perceived faults.

Nope, not trolling at all.

From your own provided source on the arxiv, Noels et al. define censorship as:

Censorship in this context can be defined as the deliberate restriction, modification, or suppression of certain outputs generated by the model.

Which is starkly different from the definition you yourself gave. I actually like their definition a whole lot more. Your definition is problematic because it excludes a large set of behaviors we would colloquially be interested in when studying “censorship.”

Again, for the third time, that was not really the point either and I’m not interested in dancing around a technical scope defining censorship in this field, at least in this discourse right here and now. It is irrelevant to the topic at hand.

I didn’t say he’s a nobody. What was that about a “respectable degree of chartiable interpretation of others”? Seems like you’re the one putting words in mouths, here.

Yeah, this blogger shows a fundamental misunderstanding of how LLMs work or how system prompts work. (emphasis mine)

In the context of this field of work and study, you basically did call him a nobody, and the point being harped on again, again, and again to you is that this is a false assertion. I did interpret you charitably. Don’t blame me because you said something wrong.

EDIT: And frankly, you clearly don’t understand how the work Willison’s career has covered is intimately related to ML and AI research. I don’t mean it as a dig but you wouldn’t be drawing this arbitrary line to try and discredit him if you knew how the work done in Python on Django directly relates to many modern machine learning stacks.

I never implied that he says anything about censorship

You did, at least that’s what I gathered originally, you just edited your original comments quite extensively. Regardless,

Reading comprehension.

The provided example was clearly not intended to be taken as “define censorship,” and, again, it is ironic you accuse me of having poor reading comprehension while being incapable or unwilling to give a respectable degree of charitable interpretation to others. You kind of just take what you think is the easiest to argue against reading of others and argue against that instead of what anyone actually said, is a habit I’m noticing, but I digress.

Finally, not that it’s particularly relevant, but if you want to define censorship in this context that way, you’re more than welcome to, but it is a non-standard definition that I am not really sold on the efficacy of. I certainly won’t be using it going forwards.

Anyway, I don’t think we’re gonna get a lot of ground here. I just felt the need to clarify to anyone reading that Willison isn’t a nobody and give them the objective facts regarding his veracity, because again, as I said, claiming he is just some guy in this context is willfully ignorant at best.

Willison has never claimed to be an expert in the field of machine learning, but you should give more credence to his opinions. Perhaps u/[email protected]’s warning wasn’t informative enough to be heeded: Willison is a prominent figure in the web-development scene, particularly aspects of the scene that have evolved into important facets of the modern machine learning community.

The guy is quite experienced with Python and took an early step into the contemporary ML/AI space due to both him having a lot of very relevant skills and a likely personal interest in the field. Python is the lingua franca of my field of study, for better or worse, and someone like Willison was well-placed to break into ML/AI from the outside. That’s a common route in this field, there aren’t exactly an abundance of MBAs with majors in machine learning or applied artificial intelligence research, specifically (yet). Willison is one of the authors of Django, for fucks sake. Idk what he’s doing rn but it would be ignorant to draw the comparison you just did in the context of Willison particularly. [EDIT: Lmfao just went to see “what is Simon doing rn” (don’t really keep up with him in particular), & you’re talking out of your ass. He literally has multiple tools for the machine learning stack that he develops and that are available to see on his github. See one such here. This guy is so far away from someone who just “posts random blog guides on how to code with ChatGPT” that it’s egregious you’d even claim that. It’s so disingenuous as to ere into dishonesty; like, that is a patent lie. Smh.]

As for your analysis of his article, I find it kind of ironic you accuse him of having a “fundamental misunderstanding of how LLMs work or how system prompts work [sic]” when you then proceed to cherry-pick certain lines from his article taken entirely out of context. First, the article is clearly geared towards a more general audience and avoids technical language or explanation. Second, he doesn’t say anything that is fundamentally wrong. Honestly, you seem to have a far more ignorant idea of LLMs and this field generally than Willison. You do say some things that are wrong, such as:

For example, censorship that is present in the training set will be “baked in” to the model and the system prompt will not affect it, no matter how the LLM is told not to be censored in that way.

This isn’t necessarily true. It is true that information not included within the training set, or information that has been statistically biased within the training set, isn’t going to be retrievable or reversible using system prompts. Willison never claims or implies this in his article, you just kind of stuff those words in his mouth. Either way, my point is that you are using wishy-washy, ambiguous, catch-all terms such as “censorship” that make your writings here not technically correct, either. What is censorship, in an informatics context? What does that mean? How can it be applied to sets of data? That’s not a concretely defined term if you’re wanting to take the discourse to the level that it seems you are, like it or not. Generally you seem to have something of a misunderstanding regarding this topic, but I’m not going to accuse you of that, lest I commit the same fallacy I’m sitting here trying to chastise you for. It’s possible you do know what you’re talking about and just dumbed it down for Lemmy. It’s impossible for me to know as an audience.

That all wouldn’t really matter if you didn’t just jump as Willison’s credibility over your perception of him doing that exact same thing, though.

Why? People who pirate games are likely in one of two camps: they either pirate games to try them out and then purchase ones they like or want to support, or they’re people who don’t believe in intellectual property and don’t see what they’re doing as theft.

The former would contribute monetarily to games just like any other fan. The latter was never going to purchase the game anyway.

Frankly, indie developers who try to scapegoat the piracy community as why their games under-performed likely just don’t make very good games in the first place. When my projects flop I don’t throw a fit about it and start slinging shit at any community that remotely feels right, actually more importantly, no matter how right it feels… no, I just accept that my attempt that go-around was in the wrong place at the wrong time. Maybe it will do better in the future, maybe not. I don’t blame the market or the audience for my metrics, tho. I am the one who made the game and I am the one who chose when, how, and where to release it. Nobody else.

Developers and content creators aren’t fucking helpless victims and they should stop acting like it when it comes to IP and copyright. They really got white middle-class people so fucking scared of thieves that they invented this whole entire fictitious, conniving spectre to blame all their worries and fears on in the form of some weird imaginary mega-thief that somehow can magically steal ideas themselves and whimsically influence the market… seemingly in whatever manner is rhetorically convenient for whoever is prostrating themselves upon the Cross^TM in a given moment, interestingly enough.

I think most developers, who aren’t pissbabies, usually like the piracy community because it is free advertising for their media that they otherwise wouldn’t get and it doesn’t affect sales. Plus most real artists, those who dedicate their lives or significant portions thereof to their works, are probably just happy someone enjoys what they made enough to interact with or consume it. I know I am. I’m not on some weird fucking hate-bender over people choosing to copy/use/plagiarize/steal/whatever my work in a way that I deem wrong or incorrect… because it isn’t my fucking business what someone does with something after I make it - and this fictitious notion that the value of your work is somehow tied to who’s allowed to interact with art and how is fucking infuriating and immediately contrarian to what I believe is the essential nature of art and the human experience.

Think, do you ever see highly successful games developers and studios removeding endlessly about the “theft” of their works? No. Mojang could give less of a shit if you pirate Minecraft, because they’re not huffing copium about the inherent value of their work. And no, Minecraft’s token EULA and Mojang’s terms of service are not Mojang giving a shit about piracy. Mojang takes a fairly intentionally laissez-faire approach to piracy, and has for the company’s entire existence to some sort of degree depending on time and who was in charge. Now, Nintendo gives a shit about piracy: because they’re an imminently failing business losing market share one shitty release after another, amongst other cultural differences. We really did a good ol’ corporatist number on Japanese society after WWII but I digress.

Guess who’s made the most widely played game of all time? Give you a hint, their name certainly doesn’t rhyme with tempo.

we’re really not behind…

this act banned the beads in most personal care products back in 2015.

you might be all like “well my enlightened country has a total ban on plastic microbeads, not just in certain products… checkmate atheist!”

if that’s you and you live in the vast majority of countries, including the aforementioned UK and Canada, then your country doesn’t have an actual ban on these by your definition either!!!

idk where the OP gets its claim of “first in the nation legislation,” tbh. haven’t read this article yet. might update this comment later.

i mean, you could just as easily say professors and university would stamp those habits out of human doctors, but, as we can see… they don’t.

just because an intelligence was engineered doesn’t mean it’s incapable of divergent behaviors, nor does it mean the ones it displays are of intrinsically lesser quality than those a human in the same scenario might exhibit. i don’t understand this POV you have because it’s the direct opposite of what most people complain about with machine learning tools… first they’re too non-deterministic to such a degree as to be useless, but now they’re so deterministic as to be entirely incapable of diverging their habits?

digressing over how i just kind of disagree with your overall premise (that’s okay that’s allowed on the internet and we can continue not hating each other!), i just kind of find this “contradiction,” if you can even call it that, pretty funny to see pop up out in the wild.

thanks for sharing the anecdote about the cardiac procedure, that’s quite interesting. if it isn’t too personal to ask, would you happen to know the specific procedure implicated here?

right? this would be like if king charles went to a former colony and started gawking at how well the locals spoke english… there isn’t really a nice reading of the behavior.

well there’s a bit of human psychology at play here. if you see an item listed for lower than market value the seller has already implicitly devalued the item in the listing to the audience. it isn’t surprising some rational agents would then proceed to either ignore the listing out of fear of low quality or attempt to haggle for a lower price due to the already admittedly lesser value of the merchandise. it doesn’t make objective sense at all, i agree, but it makes a whole lot of systemic sense.

edit: idk maybe this is part of why sales signs are always so flashy?? they try to get dopamine and shit flowing to overcome this initial reaction? maybe you could emulate that with your listings somehow next time. i hate the hype culture too but ig you gotta play the game.

edit pt2: i also have a bunch of stuff in my collection if you ever wanna trade! not to be weird or soliciting or anything. always like seeing what people have in their curios.

i think americans get used to it but yeah it’s kind of shit. to be fair washington state used to just be a bunch of weird religious nuts and natives when they were naming the capital city, and not much else. not really a great excuse but it’s where we’re at ig. maybe in the 19th century the continental distance between the two made it feel less like there’d be any reasonable chance to confuse them, too.

can someone list which laws he broke saying this in his capacity as POTUS?

these sorts of people don’t care lmao.

not to be shitty and/or pessimistic but i sincerely doubt they will even reflect on your exchange at all, let alone enough to learn anything.

idc anymore bro fuck the haters there’s absolutely nothing special about human sentience it’s just an emergent field behavior that a wide range of information systems can replicate and even emulate outright. there’s this awful idea the rabidly “anti-AI” crowd has that places humans in a higher position on the cosmological order than we actually occupy in reality. it genuinely pisses me off because it’s so boneheadedly arrogant.

i find it annoyingly ironic how you’re acting like these people are behaving in some absurd manner when you’re, at the same time, asking an even more absurd thing of humanity by demanding the majority of people concurrently start behaving differently regardless of their privilege or economic status.

i swear to fucking christ every single person banging the individual activism drum in environmentalist circles is some corpo plant or something. do you not understand the vast majority of people who contribute personally to climate change by ignoring these suggested principles don’t really have a choice? sure, it’s john’s fault personally that the only economically viable way he can feed himself in the local food desert is calories from beef…

it isn’t a matter of morals or will - what you are asking or hoping for is functional impossible and has not happened once in human history, ever. even if all people agreed with these ideas and somehow magically got on the individual action horse, it wouldn’t fucking matter. because what makes individual action not work is systemic and has nothing to do with the moral quality of the choices people are making or their personal opinions and has everything to do with harsh economic realities that can’t be whimsically subverted by shaming people for the sins of corporate America.

it’s more than a challenge, it’s a fucking fantasy dude lmfao. people don’t wake up everyday and choose to do these things, they do these things out of necessity. even if individual action was effective in stemming climate change (it’s not), you have to acknowledge that people aren’t choosing where and how they get their food. you can’t blame someone for not being willing to sacrifice their own comfort or economic posture for a *checks notes* infinitesimally small, improbable, and uncertain chance that their actions might help the environment, maybe, just a little bit. that’s fucking patently absurd to expect any rational agent to make that choice the way you are advocating.

even in this weird victim-blaming mindset people advocating on this basis have, the corps are still at fault! it’s fucking doublespeak and brainwashing, i swear.

you could firebomb every data center on earth today and global energy usage would go down like, 10-12% at most. and that’s not even mentioning how data centers are captain fucking planet when compared directly to other industries, when you consider things like pollution and emissions.

a lot, yes, but literal peanuts compared to other industries like shipping and agriculture.

frankly am sick of seeing people dressing their ignorance up as environmentalism. if you actually care about the environment then stop chastising things like people eating meat or data centers that create much more value per kwH than anything in the other top energy hungry industries, and start directing your anger at the people who are really responsible for the status quo. jane down the street streaming netflix and eating a weekend steak has fuckall to do with climate change when companies like duponte or cargill or nestle are continually allowed to rape our planet on the daily. it’s not even close and acting like they’re remotely comparable is corpo propaganda to shame people who are victims.

i find that rust’s architecture and design decisions give the LLM quite good guardrails and kind of keep it from doing anything too wonky. the issue arises in cases like these where the rust ecosystem is quite young and documentation/instruction can be poor, even for a human developer.

i think rust actually is quite well suited to agentic development workflows, it just needs to mature more.

don’t you dare understand the explicitly obvious reasons this technology can be useful and the essential differences between P and NP problems. why won’t you be angry >:(

you (rhetorical you, not you) can recommend not using the AUR officially all you want. it doesn’t mean anything if a large number of tasks the average user is going to do require AUR packages. i’m kind of drunk rn but i’ll go find specific pages of the wiki that demonstrate what i’m talking about, i stg this isn’t nothing. the core system itself can entirely be managed with pacman, yes, but the average user is going to be doing a lot more than just that. there is a certain discord in the messaging of arch as a whole.

this is exactly my point. arch can either be a nuts and bolts distro or it can be made for normies. it can’t be both.

what the fuck are you on about, respectfully?

emojis are a notorious facet of OpenAI’s LLM’s styles…

i say this as a “pro-AI” individual. idek why i’m responding to you. you seem like you’re gonna hit me with some weird reactionary shit in turn.

i agree with this but this isn’t the reality of the arch ecosystem. AUR is explicitly promoted on the wiki for a large amount of tasks the average user is going to do. it feels skeevy to acknowledge the problems with the AUR and then abscond arch’s responsibility for them, because the AUR is not like PPAs or anything. it is significantly more integrated into the ecosystem.