mfed1122 @ mfed1122 @discuss.tchncs.de

Posts

2

Comments

68

Joined

4 mo. ago

Yeah very true! It's just too bad that then it wouldn't be a core/universal feature, but I agree it makes the most sense on the client. I just wish it was possible to make it more universal, since this seems like a feature that would be useful to average users, but selecting clients based on these features seem more like a power-user level of concern. I suppose that would just be a matter of clients all copying useful features from each other if it gets popular.

In the same sense that some users might post only articles about ICE in California, or only articles about hurricanes in Florida, I still think that's not very strange. Some people are particularly invested in specific topics. Maybe the author is or is close to rape victims and is therefore especially interested in it. People dedicate their whole lives and careers to specific activist topics, so I don't think it's too strange for someone to dedicate most of their posting activity on one particular website to one. Anyways, I'm not sure what the ulterior motive would be here anyways - what do you think is the real reason for posting so many articles about rape?

But reasoning about it is intelligent, and the point of this study is to determine the extent to which these models are reasoning or not. Which again, has nothing to do with emotions. And furthermore, my initial question about whether or not pattern following should automatically be disqualified as intelligence, as the person summarizing this study (and notably not the study itself) claims, is the real question here.

Sorry, I can see why my original post was confusing, but I think you've misunderstood me. I'm not claiming that I know the way humans reason. In fact you and I are on total agreement that it is unscientific to assume hypotheses without evidence. This is exactly what I am saying is the mistake in the statement "AI doesn't actually reason, it just follows patterns". That is unscientific if we don't know whether or "actually reasoning" consists of following patterns, or something else. As far as I know, the jury is out on the fundamental nature of how human reasoning works. It's my personal, subjective feeling that human reasoning works by following patterns. But I'm not saying "AI does actually reason like humans because it follows patterns like we do". Again, I see how what I said could have come off that way. What I mean more precisely is:

It's not clear whether AI's pattern-following techniques are the same as human reasoning, because we aren't clear on how human reasoning works. My intuition tells me that humans doing pattern following seems equally as valid of an initial guess as humans not doing pattern following, so shouldn't we have studies to back up the direction we lean in one way or the other?

I think you and I are in agreement, we're upholding the same principle but in different directions.

But for something like solving a Towers of Hanoi puzzle, which is what this study is about, we're not looking for emotional judgements - we're trying to evaluate the logical reasoning capabilities. A sociopath would be equally capable of solving logic puzzles compared to a non-sociopath. In fact, simple computer programs do a great job of solving these puzzles, and they certainly have nothing like emotions. So I'm not sure that emotions have much relevance to the topic of AI or human reasoning and problem solving, at least not this particular aspect of it.

As for analogizing LLMs to sociopaths, I think that's a bit odd too. The reason why we (stereotypically) find sociopathy concerning is that a person has their own desires which, in combination with a disinterest in others' feelings, incentivizes them to be deceitful or harmful in some scenarios. But LLMs are largely designed specifically as servile, having no will or desires of their own. If people find it concerning that LLMs imitate emotions, then I think we're giving them far too much credit as sentient autonomous beings - and this is coming from someone who thinks they think in the same way we do! The think like we do, IMO, but they lack a lot of the other subsystems that are necessary for an entity to function in a way that can be considered as autonomous/having free will/desires of its own choosing, etc.

This sort of thing has been published a lot for awhile now, but why is it assumed that this isn't what human reasoning consists of? Isn't all our reasoning ultimately a form of pattern memorization? I sure feel like it is. So to me all these studies that prove they're "just" memorizing patterns don't prove anything other than that, unless coupled with research on the human brain to prove we do something different.

I do think the real world has some differences that make it more difficult. Mostly that whoever is coordinating the larger groups is very likely to have access to more power and resources and therefore is corruptible. And then that's one of the systems that brings about that Pareto distribution sort of imbalance among people. Some inequality in terms of power is not destructive, but too much is almost guaranteed to end badly. But online, the sort of power and resources that are accrued are ultimately just less likely to eventually reach a point of being able to exert full control over the smaller layers of the community. I mean sure, someone could start acting despotic with their own "fiefdom" as another commenter aptly put it, like has sometimes happened with open source repositories or forums, but it's hard for someone's website to get so popular that they're somehow able to directly force changes upon your website (not impossible, I know).

Yes, I like it smaller! Ideally you have a sort of fractal structure of a bunch of smaller, tighter communities, which are also bound up in larger but looser communities. Then you can get the benefits of broad exposure and resource sharing from large communities, as well as the benefits of meaningful individual engagement and respectful kinship from smaller communities. I think that personal sites along with forums and the rest of the Internet really can do a great job of bringing this about.

As with many things, the responsibility ultimately lies on the individual to protect themselves and resist falling into bad patterns. Most primarily, maintaining your small community takes effort, and it's much easier to just be a passive part of a very large community that subsists on infrequent uninvested involvement from many people. It's even easier to be part of a "community as a service" like Facebook, Instagram, Tiktok, etc. where all the incentives behind community building responsibilities have been supplemented with real income or fame. But of course then the people making posts, suggesting ideas, steering trends, managing communities, etc. are all in it for reasons that are not necessarily aligned with the well-being of community members. Hence the platform becomes a facade of a healthy community. Really good community upkeep seems to need to be done out of a love for the community, and any income you collect is to support that, rather than the other way around. But love for a community is often not sufficient fuel to power someone to serve huge groups out of the goodness of their heart, when they don't even know 99% of the members. Not to mention that even if someone really is that altruistic and empathetic, the time and resources become unfeasible. So ultimately, a fractal model or an interleaved model seems to be the only one that could work.

Don't get me wrong. Large communities are awesome in their own ways and have their own benefits. They have more challenges, though. Ultimately the best way to build a good large community is by building a good small community.

Everyone lamenting this needs to check out neocities, or even get into publishing your own website. Even if it's on a "big evil" service like GoDaddy or AWS, whatever. As long as it's easy for you. Or learn to self host a site. The internet infrastructure itself is the same, but now we have faster speeds, which means your personal sites can be bigger and less optimized (easier for novices and amateurs to create). People still run webrings, people still have affiliate buttons, there's other ways to find things than search engines, and there's other search engines than the big ones anyways.

There are active communities out there that are keeping a lot of the old Internet alive, while also pushing it forward in new ways. A lot of neocities sites are very progressive. If you have an itch for discussion, then publish pages on your website in response to other people's writings, link them, sign their guestbook.

Email still exists. I have a personal protonmail that I use only for actually writing back and forth to people, I don't sign up for services with it aside from fediverse ones. People do still run phpbb style forums, too. You'll find some if you poke around the small web enough.

A lot of these things are not lost or dead. They just aren't the default Internet experience, they're hard to find by accident. But they are out there! And it's very inspiring and comforting.

I'm not sure about this, but does defederation or blocking work both ways? I.e can an instance stop federating their own content to another instance? I would think so. It seems like corporate instances be pretty effectively boycotted like that.

Sorry, my examples maybe didn't make clear what my issue with the post is. The fact that public support for Israel in Western Europe is at the lowest point ever recorded, is not really a "YSK", it's not a piece of advice or tip that I can use in my daily life. It's good information, but it belongs under News, or Politics. It's not, as the sidebar says "things that can make your life easier", unless you went to argue that it psychologically makes my life easier, in which case then I can fit just about anything into this community, in which case why do I even have the community? If everything belongs in the community, then the community may as well not exist.

Just think of how much better and more honest this post would have been if it had been made in a news community with a title that was just the title of the article and then a link to the article. But by being posted here in this manner, it comes across as engagement bait - and yes, the title is definitely contributing to that. Is it really news to anyone that people don't like genocidal murderous bastards? Is that really something "I should know"?

Technically anything that's news could also be posted here, if we take the definition of the community at its most literal level. But if that's the case, why should we have a separate news community and a ysk community? Clearly, there should be some sort of distinction between things that belong in ysk versus in the various news communities.

But in all practicality, every Lemmy user already knows about Israeli genocidal behavior in Gaza. If every community just becomes format-differentiated reposts of the same stuff, all of Lemmy becomes one big content-blob.

Even if I totally agree that, for example, Elon Musk is obnoxious, and I want to hear some news that he got punched in the face - I don't want to open Lemmy and see:

You should know Elon Musk got punched in the face

Mildly interesting: Elon Musk got punched in the face

Mildly infuriating: Whoever punched Elon Musk in the face didn't punch him hard enough

Map porn: Countries where Elon Musk has been punched in the face

Gaming: Would you play a Punch Elon Musk In The Face Simulator?

Am I the asshole: for thinking Elon Musk deserved to be punched in the face?

Programmer Humor: if(isElonMusk){punchedInFace = True;}

Privacy: If it's illegal to punch Elon in the face why is it legal to punch my privacy in the face with tracking?

LinuxMemes: sudo punch Elon Musk in face

Uplifting news: Elon Musk punched in face

Depressing news: Elon Musk not punched twice in face

Television: Just watched this character get punched in the face. Remind you of anyone?

Classic Rock: "Facepunch" - 1982

Piracy: Links to movies where billionaires get punched in the face?

I love this comment so much. One of the biggest things that destroyed the quality of Reddit, although this is almost never talked about, was the trend of shoehorning the same topic into every subreddit, no matter how niche. Then to make matters worse, people will insist on leaving the post in an unsuitable community just because they like the sentiment of the post. But over time this means that the purpose of communities completely breaks down, and the whole site just becomes "different formats for us all to express the same take on the same current event". Absolutely insidious. Entire purpose of communities is so that people can customize their experience and see different types of content depending on what they're interested in. Forcing the same topic into every community not only makes the service insufferable, but it also means there's no point to joining small communities or contributing to them. You devolve to everyone just looking at the top most popular stuff, because all they would see anywhere else is just cutesy forced variants on that same thing anyways. Do not force topics into every community.

Again: Do not force topics into every community.

Yeah, that was a critical quote. Undeniably he is saying that black people should be sterilized.

University course lectures are still the best and you can find tons of them in YouTube in their entirety.

The problem is the location of the steepness makes the difference between whether this means it's easy first and slow progress later, or slow progress first and easy later. Is it like, x^1.5, or is it like ln(x)? Both are very steep at some point.

Yeah, I don't think the phrase "learning curve" has any built-in suggestion, even culturally, to imply that the reasonable default assumption is one way or the other. I only ever heard learning curve to refer to something getting easier after awhile, which is indeed a valid curve

Yeah this is a common misunderstanding I've had to clarify to people as well, even people who work in tech. I support only using "Law" for things that are scientifically actually laws. I don't even like to use it as a joke (Murphy's Law) because, unbelievably, some people really do take that to be a law of the universe too.

Hm, this is interesting. I only have a passing understanding of control theory, but couldn't a positive feedback loop indeed be good when the output is always desirable in increased quantities? A positive feedback loop doesn't necessarily lead to instability, like you said. So maybe this is just me actually-ing your actually, lol.

As for "more optimal", oof, I say that a lot so maybe I'm biased. When I say that I'm thinking like a percentage. If optimal is X, then 80% of X is indeed more of the optimal amount than 20% of X. Yes, optimality is a point, but "more optimal" just seems like shorthand for "closer to optimal". Or maybe I should just start saying that?

This reminds me of a professor I had who hates when people say something is "growing exponentially", since he argued the exponent could be 1, or fractional, or negative. It's a technically correct distinction, but the thing is that people who use that term to describe something growing like x^2, are not even wrong that it's exponential. I feel like when it comes to this type of phrasing, it's fine not to deal with edge cases, because being specific actually makes what is said more confusing.

"I'm in a negative feedback loop with respect to my laziness which will soon stabilize with me continually going to the gym daily, which is closer to optimal than before. As a result, my energy levels are going to increase exponentially, where the value of the exponent is greater than 1!"

Hmm. Now that I say it that doesn't seem that crazy. Although I do still think some common "default settings" don't do any harm.