xpi blocking AI elephants or just a css-filter?

[palidaluna]

is there any extension like "bye bye AI" for PM?

[Moonchild]

More information about what you are referring to, including links, would be helpful.
Help us help you.

[Mæstro]

I presume you are asking about how to block out such ‘AI overviews’ as Google or DuckDuckGo present, forged results in image searches etc. This is impossible with Google, but DuckDuckGo maintains a subdomain with uncontaminated results. Moreover, a blacklist for µBlock Origin exists.

[Gemmaugr]

I couldn't find any greasemonkey userscripts, but ironically enough, you could probably use https://www.yeschat.ai/gpts-9t557kHi5CA ... -Generator to generate one.

[palidaluna]

Yes, was about blocking "Hallucinating AI" mainly Google´s AI Overview.

Link for a Chrome extension "Bye Bye, Google AI"
https://chromewebstore.google.com/detai ... clpppchggc

Link for a Firefox extension: "byebye AI"
https://addons.mozilla.org/en-US/firefo ... /byebyeai/

The idea of using a LLM to block it per script is also worth a look.

[Mæstro]

Now your original title makes more sense. For the uninitiated, seeing pink elephants is an English idiom for hallucination.

Off-topic:
Gah, Google has already included these for over a year! How time passes…

[palidaluna]

Adding this line to µBlock Origin solved my issue. Let´s see how long.

google. com##. hdzaWe

[Mæstro]

My understanding, which inspired my use of ‘impossible’ above, is that blocks like these in Google only hide the LLM output from the user, but internally, Google still wastes the energy to make the stochastic parrot vomit something, and this habit (I cannot call it a feature) cannot be disabled. Because you seem neutral to LLM use in principle, from how you were willing to consider using one to produce a user script, this might not bother you, but i should leave this note, pending any correction, for anybody else who opposes LLM use in principle.

[Michaell]

I'm sure I oppose LLM but I don't know what it is except something relating to AI.

[Gemmaugr]

I'm sure I oppose LLM but I don't know what it is except something relating to AI.

To my understanding, LLM stands for Large Language Model, and is the gathered and scraped and bought and telemetry data from people's online conversations and articles. Which is the basis for how the AI algorithm generates "human" responses.

[palidaluna]

Failure using Yeschat.AI and same using the filter for µBlockOrigin, worked just till restart.

I´ll filter google Overview AI using metasearch engines like Duck, searxNG or Startpage.

[Moonchild]

Or, just stop using search engines that harm the overall internet and ecosystem. Sure, maybe you need to go to page two of results sometimes. But is that really that much of an inconvenience you're willing to cause so many consequences for?

[Mæstro]

To my understanding, LLM stands for Large Language Model, and is the gathered and scraped and bought and telemetry data from people's online conversations and articles. Which is the basis for how the AI algorithm generates "human" responses.

This is correct. I refuse to refer to this software as ‘AI’, for it has neither anything to do with traditional artificial intelligence (rules-based chatbots of the kind I played with as a child, CPU-controlled characters in a video game…) on one side, nor fictional artificial intelligence (such as humanoid robots) on the other.

[Gemmaugr]

To my understanding, LLM stands for Large Language Model, and is the gathered and scraped and bought and telemetry data from people's online conversations and articles. Which is the basis for how the AI algorithm generates "human" responses.

This is correct. I refuse to refer to this software as ‘AI’, for it has neither anything to do with traditional artificial intelligence (rules-based chatbots of the kind I played with as a child, CPU-controlled characters in a video game…) on one side, nor fictional artificial intelligence (such as humanoid robots) on the other.

It would be more accurate to call them A.I.M, Artificial Intelligence Mimicry, as they're not even Artificial Sentience, never-mind Artificial Sapience (The former two are not Turing-compliant, while the latter would be).

[palidaluna]

The LLM approach can´t achieve intelligence.These things do not think. Good as entertainment, bad for science, deadly as medicine.

[Mavericks_10_9]

Or, just stop using search engines that harm the overall internet and ecosystem. Sure, maybe you need to go to page two of results sometimes. But is that really that much of an inconvenience you're willing to cause so many consequences for?

I find the "AI" search summary text paragraph(s) that Brave Search provides to be quite useful and time-saving. So many web pages now have un-blockable pop-up videos, that just waste my time. Sports web pages are especially bad.

https://search.brave.com/