Because they attach themselves to the "big thing" in tech right now. The "big thing" is a really big tide so the hope is that it will lift all boats. That's it, it doesn't have to actually do anything, just associating with this gives a good momentary boost. Used to be blockchain and NFTs until just recently.
But some companies are not boats, they're stones on the bottom barely keeping their peak above water. Others are not even in the water but on the side of a hill far away, like the Long Island Blockchain. They still wave the "big thing" flag so some of its image will rub off on them even.
Of course, unlike the the Long Island Blockchain these guys can actually generate content with AI but I don't think that's the endgame here. Doesn't take a genius to tell that hundreds of clones can now pop up and deliver perfect equivalence with little to no effort. So the goal is probably to pump and dump while there's still something to dump.
https://edition.cnn.com/2021/07/10/investing/blockchain-long...
"National Enquirer Now Legit, According to Pulitzer Prize Board" https://abcnews.go.com/Entertainment/national-enquirer-now-l...
"Shock! Horror! National Enquirer wins a Pulitzer. Not" https://www.theguardian.com/media/2010/apr/13/shock-horror-n...
Evidently, the Pulitzer has lost its way, much like the rest of Journalism.
Lips stick on a pig...is still a pig. And you can sprinkle magic pixie dust on a turd and make it something else.
If the likes of The NE can get a Pulitzer, then that prize has lost its meaning and significance.
Let's not be so naive, please.
If you think of the prize as being given for results rather than the process, it’s not that wild: they broke a major story that affected a presidential election. People occasionally find something valuable while digging through the trash, which seems like a pretty apt analogy.
Likewise, the Nobel prize for grapheme was effectively for scribbling with a pencil on Scotch Tape and repeatedly pulling it apart. I used to do this in elementary school, but does it cheapen the Nobel Prize? I’d argue no, the prize is——-in part——-for realizing what you’ve got.
Relying on PageRank / links (like they currently do as the super-dominating factor) doesn't work if larger sites will employ AI (like BuzzFeed).
Relying on their ability to tell apart AI and human generated texts? I'm not sure. I'm not an expert, but I fed in text from a German public broadcaster into one of the "identify ChatGPT" things and it said it looked like ChatGPT. I doubt these technologically conservative editors are using ChatGPT to write official reports.
Instead, they'd need to understand the topic at hand to judge whether a page is useful in answering the user's question. And at that point, they have an answering machine and no longer need to show search results.
It's probably that they can't. It's hard to evaluate content which you don't understand and when you have the whole world gaming against you. Likewise, it will be also interesting to see if AI really can change this, because the results so far were not gamed by SEO. At the end AI has the same problem to solve if it wants to replace search engines. The question is also how fast AIs can update their knowledge. Google Search can be blazing fast in delivering the newest results even for the most random topic.
> Relying on PageRank / links (like they currently do as the super-dominating factor) doesn't work if larger sites will employ AI (like BuzzFeed).
Aren't they doing more than this for some time now? Though, true that links alone are not very useful against Manipulations. Maybe letting AI curate the results will be the trick? Another problem might be that Google want's to stay simple and vague. Which is of course legit, as more information will give SEO more attack-surface. But maybe add a more complex interaction-mode for people who don't just search 2-3 words. A built-in ai-prompt which switches to some ChatGPT-Like response is an obvious google-style solution to battle against the new wave of competitors.
> Relying on their ability to tell apart AI and human generated texts? I'm not sure.
Is this necessary? At the end, it's a matter of quality. If AI-Texts are high quality, delivering exactly what you want, why not? After all, the problem with spam today is the quality. Though, for this of course Google would need a kind of quality-check, which their own AI can deliver, maybe.
> Relying on their ability to tell apart AI and human generated texts? I'm not sure. I'm not an expert, but I fed in text from a German public broadcaster into one of the "identify ChatGPT" things and it said it looked like ChatGPT. I doubt these technologically conservative editors are using ChatGPT to write official reports.
But they usually use a very formalistic template-based style. Often it's even just copy & paste of paragraphs from different sources. So yes, news are already very AI-like today, even without AI involved. The human usually only brings in their skill in evaluating the content and composing them with matching parts.
> Instead, they'd need to understand the topic at hand to judge whether a page is useful in answering the user's question. And at that point, they have an answering machine and no longer need to show search results.
Yes, but people have not always questions to answer. And it's also a kind of trust-issue whether I want to see the original source, or some more context to the answer, or whether I blindly trust the answer from Google. I mean they already today have common Q&As in their search results, and you can see also where they are coming from. But often they don't quite satisfy my demand. Sometimes they are to short, sometimes they are plain out wrong. Sometimes they are just not what I search. Traditional search still has its place even with AI, until we reach the age of completely trustable strong AIs.
You're right that AI-generated wouldn't matter if the AI-text was of high quality. So far, Google says that AI-generated is against the guidelines (read: you can/will be penalized if you do, like you might be for cloaking). Maybe they'll change their stance if AI provides high quality texts. But if it can, will anyone still need Google, or will they go to the AI directly? And again, to judge the quality, Google would probably need to be able to generate the result themselves, in which case they'd have the trustable AI.
Search engine companies learnt long ago that parsing all that content was time-consuming and costly, so they simply use the 'meta' tags instead.[0]
ex: some of my sites are effectively invisible on search engines unless the engine is "filtering" based tightly on my geographic location because the wider "content network" is simply flooded with noise, and that problem is worsening already.
You can already pay people to do things for you, using their "online identity". This will just continue that.
Maybe the internet sends you a DNA sequence and you have to inject it, translate it and then image the folded protein (computers are ok at protein folding but not great yet)
Deciphering a captcha or staring at a camera each time you want to post are real annoyances and such a site probably wouldn't get much traction. But if the right infrastructure were in place, a lot of us would be happy to participate in a community where we're confident that everyone there is an actual person, not a spambot or a sock puppet.
The right infrastructure here could a tall order, but I don't think this is an impossible problem.
* The iPhone stores your biometric data in its Secure Enclave
* The Secure Enclave also holds the cryptographic private keys
* Every action you do against the server requires your phone to add a cryptographic signature, which can only be written when the iPhone verifies you via FaceID
Apple could act as the Certificate Authority letting websites know that this signature aligns with an iPhone user. Apple may not need to let the website know exactly who performed the action, but just say that Apple verifies that this is a real person making the action.
Eventually? Are you saying Buzzfeed content is currently NOT worthless?
"Am I so out of touch? No, it's the children who are wrong."
What does the online future look like? Drowning in mountains of AI generated garbage and AI based search engines who won't be able to differentiate between fake AI generated stuff and the truth, so we'll get garbage answers?
SEO-driven marketing content, however, is worse than worthless: it produces negative value to the mankind. It pollutes Google searches, it distills knowledge even further, it is written by either robots or underpaid freelancers, and it drowns out the work written by professionals in the area.
And maybe thas is a good thing. Maybe as a civilization, we need to come out of slumber, of drugging ourselves out day and night with endless tv series, super hero movies and so called "games" (about 90% of which are garbage and/or cookie clickers at their core).
So it’s fascinating : our best shot at capitalism created all this technological innovation, but also so much parasitic, essentially pointless "work" where people toil away creating "content" and "entertainment" such as games, movies etc the majority of which is consumed in our "time off" to unwind and forget about "working" (I’m thinking in particular of the movie Kooyanisqatsi which illustrates this madness).
We couldn’t imagine a different way of life - but the funny thing is, AI will help us come out of this slumber by making the majority of so called "entertainment" - which is garbage let’s face it (80/20 rule and all that) - pretty much obsolete as well as valueless.
So this seems paradoxical but I believe the best thing about AI is it will free us and drive us towards real value and authenticity in our day to day relationships and expression.
Gonna be a long road though…
Talk about the trading economy being a zero-sum game where people are moving tokens from one hand to another? Wait until you see the AI economy.
The AI will need a corpus of _your content_ to draw upon if it's to write in your style. One could feed it one's work inbox (presuming one's employer allows this) but new grads won't have the luxury of having once written corporate emails by hand.
Take a ChatGPT algorithm and train it on funny texts v. train it on enraging texts, for example.
The goal will be to train an AI that writes the most engaging and addictive content for the target audience.
A good CEO provides irreplaceable value, but there is a significant fraction of enterprises that would probably do just well if they were run by ChatGPT.
on the content side of things i don't think people will notice
i don't think this is widely adopted though, most people i know are still married to the idea of having an app than using their mobile browser, it will take some time (or a stellar marketing campaign for the google and bing apps)