Welcome to DU! The truly grassroots left-of-center political community where regular people, not algorithms, drive the discussions and set the standards. Join the community: Create a free account Support DU (and get rid of ads!): Become a Star Member Latest Breaking News Editorials & Other Articles General Discussion The DU Lounge All Forums Issue Forums Culture Forums Alliance Forums Region Forums Support Forums Help & Search

justaprogressive

(3,407 posts)
Thu Apr 24, 2025, 07:52 AM Yesterday

'You Can't Lick a Badger Twice': Google Failures Highlight a Fundamental AI Flaw - Wired

Here’s a nice little distraction from your workday: Head to Google, type in any made-up phrase, add the word “meaning,” and search. Behold! Google’s AI Overviews will not only confirm that your gibberish is a real saying, it will also tell you what it means and how it was derived.

This is genuinely fun, and you can find lots of examples on social media. In the world of AI Overviews, “a loose dog won't surf” is “a playful way of saying that something is not likely to happen or that something is not going to work out.” The invented phrase “wired is as wired does” is an idiom that means “someone's behavior or characteristics are a direct result of their inherent nature or ‘wiring,’ much like a computer's function is determined by its physical connections.”

It all sounds perfectly plausible, delivered with unwavering confidence. Google even provides reference links in some cases, giving the response an added sheen of authority. It’s also wrong, at least in the sense that the overview creates the impression that these are common phrases and not a bunch of random words thrown together. And while it’s silly that AI Overviews thinks “never throw a poodle at a pig” is a proverb with a biblical derivation, it’s also a tidy encapsulation of where generative AI still falls short.

As a disclaimer at the bottom of every AI Overview notes, Google uses “experimental” generative AI to power its results. Generative AI is a powerful tool with all kinds of legitimate practical applications. But two of its defining characteristics come into play when it explains these invented phrases. First is that it’s ultimately a probability machine; while it may seem like a large-language-model-based system has thoughts or even feelings, at a base level it’s simply placing one most-likely word after another, laying the track as the train chugs forward. That makes it very good at coming up with an explanation of what these phrases would mean if they meant anything, which again, they don’t.


https://www.wired.com/story/google-ai-overviews-meaning/

The not-ready-for-primetime AI scam!
8 replies = new reply since forum marked as read
Highlight: NoneDon't highlight anything 5 newestHighlight 5 most recent replies
'You Can't Lick a Badger Twice': Google Failures Highlight a Fundamental AI Flaw - Wired (Original Post) justaprogressive Yesterday OP
And yet, this is what is deciding which Americans get kidnapped off the streets and sent to concentration camps. Scrivener7 Yesterday #1
AI is misnomer al bupp Yesterday #2
I'm more inclined to consider AI just a collection of algorithms EYESORE 9001 Yesterday #4
I like the term Simulated Intelligence---cuz it's a pretend form (sort of like tsf's brain) Maeve Yesterday #6
Yes, in the sense... al bupp 23 hrs ago #8
Boy, I got a doozy mimitabby Yesterday #3
Oh, drat! EYESORE 9001 Yesterday #5
So the other day I was reading up on Sam Jones the moniss Yesterday #7

Scrivener7

(55,481 posts)
1. And yet, this is what is deciding which Americans get kidnapped off the streets and sent to concentration camps.
Thu Apr 24, 2025, 07:59 AM
Yesterday

al bupp

(2,450 posts)
2. AI is misnomer
Thu Apr 24, 2025, 08:07 AM
Yesterday

As Roger Penrose points out, there's no intelligence in what's currently labeled AI. Intelligence requires understanding and it lacks that. It has no concept of self or ability to invent. This may very well yet happen, but LLMs aren't that.

EYESORE 9001

(28,096 posts)
4. I'm more inclined to consider AI just a collection of algorithms
Thu Apr 24, 2025, 08:37 AM
Yesterday

None of which model the real world worth diddly.

Maeve

(43,215 posts)
6. I like the term Simulated Intelligence---cuz it's a pretend form (sort of like tsf's brain)
Thu Apr 24, 2025, 08:46 AM
Yesterday

al bupp

(2,450 posts)
8. Yes, in the sense...
Thu Apr 24, 2025, 02:09 PM
23 hrs ago

It's been designed cleverly enough to essentially fake a Turing test, at least to a first approximation. But there's no consciousness, which is what intelligence depends upon, I think. This is not say it's not useful for some purposes, especially those requiring massive computational resources, such as protein analysis.

Until it can really think on its own, it's not intelligence in my book, and even mentioning the word to describe it makes it seem so much more than it really is.

mimitabby

(1,937 posts)
3. Boy, I got a doozy
Thu Apr 24, 2025, 08:10 AM
Yesterday

The statement "all cats have 3 tails" is a logical fallacy, specifically a misapplication of deductive reasoning. It's designed to trick you into accepting a false conclusion by subtly manipulating the meaning of "no cat". The premise relies on the fact that "no cat" is interpreted as a thing (a no-cat) rather than an absence of cats.

EYESORE 9001

(28,096 posts)
5. Oh, drat!
Thu Apr 24, 2025, 08:41 AM
Yesterday

Now I want a cat with three tails! I presently have two without tails (one has only a bump for a tail), but can you imagine the inferiority complex that would create? Even my one cat with a single tail would be intimidated.

moniss

(7,103 posts)
7. So the other day I was reading up on Sam Jones the
Thu Apr 24, 2025, 08:48 AM
Yesterday

Celtic great from the Russell years. AIthough Sam was only married once, and for a very long time, AI came up with two different names for his wife. The fault here likely is that AI grabbed any famous Sam Jones and ran with it.

Latest Discussions»General Discussion»'You Can't Lick a Badger ...