Wow, so unexpected. Who could have seen this coming? 🙄
At least Google had the decency to write “sponsored” on the sponsored results, but with this it’s not even an option.
I’ve been saying this for a while. It’s gonna be particularly difficult to get valuable insight. With sponsored search results you can keep scrolling. Sponsored content will “pollute” the context of the conversation even if only some responses push sponsors.
I’d wager anything this won’t even slow down the “artists” that use it.
Is it still enshittification if it was already shit?
So … now we have plausible gibberish … also known as Autocorrect on Steroids … that includes corporate sponsorship… seems like we’re moving closer to the true meaning of advertising with every iteration.
Next we’ll be asked to pay for this feature … oh wait.
I can’t wait until the Assumed Intelligence bubble finally bursts and takes with it some of the largest companies in the world … perhaps this is how we finally address climate change.
The fact that ads will be injected in LLM was clear. I just thought Google would do it first in their AI responses.
Lol imagine using AI.
I’d post one of Facebook’s “Imagine” images as a joke but it sort of defeats the purpose. So imagine I did it. Feel free to leave the tap on while you do if you really wanna get the full experience.
I accept my load of hallucinations and disguised approximations in exchange of relatively adfree neutral answers. That’s the only reason why I don’t go back on Google/DuckDuck for now. But as soon as I’ll see corporate bullshit forced into my chat, that’ll mark the end of my chat bot use
You think we’ll still have Google and DuckDuckGo to go back to?
“Hey, you know that chat bot that’s churning through billions in money and drinkable water every year to guess the answers to questions, lie, hallucinate, and gaslight false info? You know, the one that will never be profitable, but every giant company has a hard-on for, despite almost no one liking it…? What if we made it even more unreliable and inaccurate, by forcing in Corporate Answers™. And yeah, the customer still has to pay for it…”
Omg, even the thing that is enshittifying everything is now being enshittified itself.
If you were trying to link to an article, you just linked to an image.
There’s a link for me in the post?
It’s not visible via lemmy.today’s Web UI:
https://lemmy.today/post/44629301
Checking whether it’s just lemmy.today somehow mangling things, it doesn’t look like it. Here’s beehaw.org’s Web UI directly:
https://beehaw.org/post/23981271
As of the moment, both are a link to an image (though lemmy.today is proxying the link to that image):
https://fedia.io/media/93/77/937761715da35c5c9fb1267e65b4ea54c2b649c2eebbf8ce26d2b4cba20097bf.jpgSame behavior seen when browsing lemmy.today in Eternity.
EDIT: Fedia.io is running mbin. You submitted it via fedia.io. This community is on beehaw.org. I wonder if there was some kind of recent mbin/lemmy compatibility bug? I’d check fedia.io’s Web UI, but it currently disallows anonymous browsing (probably because of all the AI web scrapers hammering Threadiverse instances that allow anonymous access).
Might be due to OP being on fedia.io? I can’t access the original post (it wants a login), but the link shows up fine for me on Piefed.
Great minds think alike. Yeah, just followed up with a similar comment. So probably a bug, but not sure whether it’s an mbin or Lemmy bug. Like you, I can also see the intended URL on PieFed (olio.cafe).
EDIT: Same problem on lemmy.ml, and they’re the dev instance, run the latest version of Lemmy:
https://lemmy.ml/post/41016280
EDIT2: I submitted a post to !lemmy_support@lemmy.ml at https://lemmy.ml/post/41019532 to make the devs aware of the interaction, if they weren’t already.
There’s not for me but it’s probably the app formatting, using Summit rn, I’ll check with Thunder
Working with Blorp.
Blorp dev here. Looks like it might be an instance thing. On PieFed.social I’m seeing the article, but on Lemmy.zip I’m only seeing an image. PieFed.social is returning separate urls for post url and post thumbnail, but Lemmy.zip returns the same image url for post url and post thumbnail. I’m not sure if it’s a PieFed/Lemmy issue or just happens to be that the two instances I tested are working that way.
…and despite that people are upvoting, because who cares about facts, let’s just hate a thing that I don’t like
Facts like: they fixed the link to the factual article before your comment?
It’s not fixed, I also get this problem atm
That is straight-up nightmarish.
Like, people, it predicts the next “token” — referring to that as intelligence is way overstating it. I, too, think “AI” is impressive, but to say companies are overstating its capabilities is putting it lightly.
We really need yet another advertising vector. Not.
I think I heard about some dystopian story when I was a kid, and the premise was ads on the inside of your eyelids while you slept. Now, that’d be called “innovation.” 🤦♂️
Then it starts hallucinating and gives an ad for a non existing product with a dead link. Advertisers get mad and threaten to sue.
Getting harder to afford the setup, but there’s very compelling reasons to use local models instead













