Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Wikipedia says traffic is falling due to AI search summaries and social video

techcrunch.com

Post reply on HN
Print view View original

441–450 of 464 posts

Re: Wikipedia says traffic is falling due to AI search summaries and social video

#441

Post by int_19h »

dns_snek wrote: post #350
bawolff wrote: post #187

Earlier quoted context omitted.

No, it was a fork from like a decade ago. https://www.conservapedia.com/Main_Page It started out crazy but it only got crazier as time went on. The biggest problem with wikipedia competitors, is the only people who tend to put in effort to make one are usually crazy.

The page on the Second law of thermodynamics has a chapter on "Liberal abuse of second law of thermodynamics". I wasn't sure if I was reading an imitation of The Onion that tries to take itself too seriously.

It's founded by the son of https://en.wikipedia.org/wiki/Phyllis_Schlafly, so that brand of crazy is basically vintage.

That said, it is indeed so insane that trolls sometimes become contributors just to see how far they can push it. There were several known cases of trolls going far enough undetected to get various admin rights (to block etc).

Re: Wikipedia says traffic is falling due to AI search summaries and social video

#442

Post by int_19h »

imtringued wrote: post #379
adastra22 wrote: post #358

Earlier quoted context omitted.

I think we have a fundamental disagreement here. There is no bias in the statement "2 + 2 = 4". Nor is there bias in stating that "any two masses experience an attraction force proportional to the masses and inverse-square proportional to the distance separating them." Not less bias -- no bias.

There is a pretty strong bias in the statement "2 + 2 = 4" because math doesn't care about the encoding used to describe a sum. For example, in an alternative society the expression could be written with flipped number symbols, right to left or in a vertical layout. Axioms are also a form of bias.

Axioms would only be a form of bias if they were taken at random. It's not biased to take the axiom that repeatedly holds true in empiric testing, over others that do not.

Re: Wikipedia says traffic is falling due to AI search summaries and social video

#443

Post by int_19h »

internetter wrote: post #271
boringg wrote: post #260

Earlier quoted context omitted.

Wow they could easily maintain that for infinity at a low cost (less inflation and you know upkeep).

This is only the size of the text at one moment in time. The edit history is dozens of terabytes and the media is hundreds. All of this must be served to billions of requests a day. Almost every single request involves executing arbitrary lua and/or template syntax which was transcluded into the page. The media, particularly, introduces huge copyright problems. Ect ect. The WMF is undoubtably bloated, but it is not a…

I would be extremely surprised if most articles that are regularly viewed aren't precached and thus served directly without involving any code execution.

Re: Wikipedia says traffic is falling due to AI search summaries and social video

#444

Post by int_19h »

al_borland wrote: post #51
RickJWagner wrote: post #16

Earlier quoted context omitted.

Not for me. Wikipedia had its day, in between print encyclopedias and quick query AI. Its place in history is now set. Something else will come along soon enough.

LLMs are useless without source material. AI companies should be donating large sums of money to Wikipedia and other such sites to keep them healthy. Without good sources, we’re going to have AI training off AI slop.

They should be, yes, but they won't. We already know this from the way those same companies have treated open source projects that they depend on.

One thing that I would really like to see is some kind of hefty tax on any kind of income derived from models trained on Wikipedia. Basically, make it legal to train, to share weights etc freely, and hosting them locally. But the moment you start charging people for subscription, the society should start charging you to maintain the commons that you are profiting from.

(This likely goes for more than Wikipedia, but that case is especially simple since there's a single legal entity that could be given the money.)

Re: Wikipedia says traffic is falling due to AI search summaries and social video

#445

Post by scuff3d »

mmooss wrote: post #425
scuff3d wrote: post #412

Earlier quoted context omitted.

To think that I'm saying all humans intend to produce truth you'd have to intentionally misread my comment. Wikipedia obviously has to deal with bad actors and vandalism, and they have processes in place for that. My point is that the intent matters. Calculators aren't a useful analogy for LLMs. They produce a deterministic output based on a (relatively) narrow range of inputs. The calculations to produce those outpu…

> Wikipedia obviously has to deal with bad actors and vandalism People produce false information for many reasons besides malice, and even when they intend to be honest. Look how much bad information was out there before the advent of scientific method.

You're aware that if you don't have anything useful or interesting to contribute to the conversation you can just not comment right?

Re: Wikipedia says traffic is falling due to AI search summaries and social video

#447

Post by dvsfish »

tines wrote: post #430
dvsfish wrote: post #254

Earlier quoted context omitted.

I get the sentiment of what you're saying, but I don’t think that’s totally fair to the parent. Biases aren’t automatically "bad" or "failings". they’re akin to heuristics, and it’s practically impossible to eliminate them entirely. For example, we’re all here talking about how we should treat Wikipedia with skepticism. That's a sort of "neutral" bias that doesn't conjure a strong emotion, and is perhaps more accepta…

> In fact, the claim that “bias can be avoided and should be absolutely”, that is implicit in your resposne reflects a bias of its own: a bias toward moral or intellectual purity, as if the parent recognizing bias is equivalent to endorsing it. Nonsense. Your definition of the word "bias" includes any assertion whatsoever. Bias is distortion from reality and truth. Saying that we can avoid distortion is not itself a…

The parent’s point was plainly that bias is unavoidable and making it overt is realistically all we can do. It’s a pragmatic take, albeit overly terse. They could have additionally included something like “and try to minimize its effects” sure, but that to me was implicit. I may be reading that fairly charitably, but perhaps that is just my own bias. I certainly have a bias to judge it more charitably than I do someone who leaps straight to moral outrage and judgement this early in the interaction.

In this case, I simply felt your judgment of the parent wasn’t fair, and showed a moral bias. Maybe they weren’t perfectly clear, and too absolute, but your response wasn’t proportionate either. It condemned more than it understood. I interpreted it as an epistemic observation and you interpreted it as an offense. The very fact that we came away with two completely different readings of the same short sentence rather proves the point.

Thank you for putting words in my mouth regarding my definition of the word bias, but lets use your own: "Bias is a distortion from reality and truth." If that is the case, we can never hope to avoid it, because we will never have perfect information. Using that definition, we are quite literally constantly in a state of bias. Your very own definition is far more broadly supportive of the notion that bias can't be avoided and consequently suggests bias is effectively ubiquitous. This to me is the primary point the parent was making.

I was perhaps too charitable, and you not enough. We are both biased, and going by the advice of the parent, I'm pointing it out. I don't think there is much more I can do.

Re: Wikipedia says traffic is falling due to AI search summaries and social video

#450

Post by epolanski »

int_19h wrote: post #437
epolanski wrote: post #411

Earlier quoted context omitted.

In countless ways, there's endless wars among admins. An example though is that several historically relevant facts are edited out to favor some narrative. E.g. Crimea, the Ukrainian autonomous republic that seceded/was annexed by Russia, tried to split 3 times from Ukraine, once even during the Soviet Union, but those events have been edited out on multiple pages to favor some western-centric narratives.

It literally has an article on this exact thing you're complaining about. https://en.wikipedia.org/wiki/Republic_of_Crimea_(1992%E2%80...

> once even during the Soviet Union

That's in 1992, there was another occurrence in the previous decade.

Post reply on HN