Incorrect AI answers being at the top of the search results
Wednesday, 12 Mar 2025
Exclusive insights, data, and analysis for financial market experts.
Search
Have an existing account?
Sign In
AI is exactly on a Facebook user’s level.
Welcome to the future. Truly “post-truth” now!
add -ai at the end to get that slop away
These AI summaries are generated in real time and *cannot* be checked for accuracy automatically.
The models they’re using are basically predictive text on steroids. The AI cannot determine the truth or accuracy of what it’s spewing out, nor does it actually aim to offer truth – it aims to *sound* correct.
*Do not trust these summaries on any level,* even if it seems right. The deployment of these models so prominently is honestly disgusting and is entirely motivated by bullshit AI hype and stock market politics.
For the first time in my life , I don’t trust googles top results anymore
Bing knows 😂
?width=1284&format=pjpg&auto=webp&s=d6bda5a55fd8afd8641410be29acef9c07b5dc2e
By next year AI will already be using 4% of all the electrical energy produced on the planet. By 2030 is estimated to be 10-20%.
Sure the planet will be frying but at least we know 5/16 is bigger than 6/16…
Fucking infuriating is what it is.
The whole _intelligence_ part of AI in this context is so misleading; it’s just generative text with data scraping baked into the programme, it doesn’t _know_ anything.
The technology is impressive in theory but it’s used in entirely the wrong way. Imagine saying that my blender is really good at making smoothies and technology is advancing so fast that one day it’ll be able to make toast.
I like how it gets the explanation and conversion correct. But then it’s just oh you fucking moron obviously 6 is smaller than 5. Like tf?
Got the same incorrect answer on DuckDuckGo
This is an excellent example of how LLMs don’t actually work things out.
It’s been trained to say right-looking stuff with confidence but doesn’t have the intelligence to see how it’s wrong or work out the actual answer
I love how Google has gone from a trusted, reliable source of information to making completely unbased claims that it treats as 100% accurate by trying to use AI to do their job for them
This AI is the average US intelectual level, its the 1/3 pound burger all over again
I’m team Wolfram
It’s really AT (artificial thickness). What we call AI at the moment (ChatGPT-like AI) is really a more sophisticated autocomplete (they are statistical language models). It’s not intelligent at all (it’s not doing any reasoning) and at least the ones I’ve tried are incapable of answering “I don’t know” and instead will generate some confidently wrong slop.
I miss when Google search functioned as a calculator and not a shit regurgitator.
It’s one thing for it to mess up comparing the fractions with different denominators, but for it to convert them and still go 6 is less than 5? Wow.
This AI probably thinks 1/3 pound burgers are smaller than 1/4 pound ones.
And yet many Redditors think Google is the gospel and get pissed off when Redditors ask questions and whine that they should be looking things up on Google.
It’s not 2003 anymore.
The AI nonsense and the half page of ads is what’s made me stop using Google for most searches. I shouldn’t have to go to the second page to get actual results.
They didn’t make chat gpt to do hard stuff like math, it’s for memes and 6 fingered girls.
Good news, by exposing people to AI while it’s still this regularly inaccurate, we can more quickly reach a culture where people assume ‘*of course* the AI is inaccurate’.
And this shows perfectly that LLMs are not AGI and won’t be.
Most generative AI works by creating a response which approximates a mathematical average of a human response.
For some things like math this averaging technique doesn’t work, for the same reason why the average human being is not named Mohammed Li
Not sure if you’re trolling or Google is quick at fixing these type of things…
Can’t replicate that search result here!
You mean rage-inducing, right? They are training their AI on us as consumers, we are just along for the ride. I don’t even pay attention to the AI results anymore, I always scroll past. The amount of bullshit that gets thrown at me it’s unbelievable.
I’m already seeing the implications of bad AI being used in my job. Employees look up answers with AI, pretend to be experts with bad or misleading information, then leave me to clean up the mess. I see bad things in our future as a society if this goes on for too long. People relying on bad information for their knowledge is a recipe for disaster.
>equivalent of 6/16ths, which is *clearly* smaller than 5/16ths
Clearly?? Why is it so confident??
Told me yesterday that a Citicar goes from 0-60 in 3 seconds despite its top speed being 28 mph
The real question is who searches this
you know the worse part I’m 90% sure the average person isn’t better
There’s a firefox extension to hide the google AI overview, I just wish there was an overall AI hiding extension for all sites.
I would love to know what the hell Google was thinking when they went all in with AI.
All they had to do was replace the “I’m feeling lucky” button with “Ask an AI”. That could have kept their search engine dominance while still dipping their toes in the AI water.
The best ad for duckduckgo is Google.
I am a teacher, and Google has fucked education by doing this. So many students searching for the easy correct answers, and our shitty algorithms are failing us. The internet is objectively worse than it was 10 years ago.
When I was a kid in school I had a teacher that was horrible at math. We got into a week long argument over whether 0.5 or 3/4 was a larger value with him stating because 0.5 has less digits after the decimal than 0.75, that meant 0.5 was larger. He made the mistake of bringing in an actual math teacher to settle it. Hadn’t thought about that for a while until now.
First reaction was to wonder if this was real or not so I double checked. Yep, it’s real…
Bing gets the answer right, meanwhile DuckDuckGo says 3/8 < 5/16 because 6/16 > 5/16.
“AI” is an absolute joke, and sincerely fuck all of the braindead investors who are responsible for every company pushing awful AI models onto us at every opportunity because they don’t even remotely understand what it actually is or does.
I like how it explains how to compare fractions and then still somehow decides 5/16 is bigger than 6/16
When i type this into Google i get a correct answer so maybe it’s fixed already
I always scroll past the AI and ads to find the real answer.
And then some people will treat it as sacred knowledge.
Anyone should see that dividing the top half by the bottom half will give a decimal value that can be compared, just not a system based on an LLM apparently.
BTW would love to have an option to disable these AI responses.