Via @rodhilton@mastodon.social
Right now if you search for “country in Africa that starts with the letter K”:
-
DuckDuckGo will link to an alphabetical list of countries in Africa which includes Kenya.
-
Google, as the first hit, links to a ChatGPT transcript where it claims that there are none, and summarizes to say the same.
This is because ChatGPT at some point ingested this popular joke:
“There are no countries in Africa that start with K.” “What about Kenya?” “Kenya suck deez nuts?”
when will people learn that search results change all the time and are different for different people
I get the same as the main post. Either way the point still stands. I had someone correct me with a misconception about something because he googled it and thats what it said in the answer box. It’s getting increasingly difficult to rely on search results especially when google synthesizes them into questions and answers with little context
Yes. I have to keep reminding my parents that those little Google answer boxes aren’t real search results and can’t be trusted. They sometimes say the exact opposite of the page they’re citing!
Either way the point still stands.
What point?
I had someone correct me with a misconception about something because he googled it and thats what it said in the answer box
Oh, honey. You must be new here. Googling something, taking the first answer that fits your needs, doing zero follow up, and posting it confidently is nothing new. That has been happening for the last few decades.
It’s getting increasingly difficult to rely on search results especially when google synthesizes them into questions and answers with little context
In what way? Like always, I have had to do a little critical thinking to gain anything out of my google searches. But now, sometimes the answer is right there. In what world is that a bad thing?
But you might just say “hurry durr it give me answer, therefore correct”. Again, yes, that has always been the case. People will use any tools available to them to support their point. If a new tool has less than a 100% success rate, I don’t see that as a problem.
The point is that google is no longer just listing search results. For years now it has been giving the “correct” answer as well as results. This started of with things it could recognise and easily solve like calculations (“what is 432 times 548”), but has now moved into general queries powered by LLMs that have no knowledge of fact.
Okay? As I said, google has been giving incorrect results for decades. Now, just like before, it gives incorrect answers sometimes. But it has gotten a LOT better at giving those correct answers.
You are arguing in bad faith. It is a fact that Google results have been getting worse over time. What is your point? That with extra effort, you might get the answer you’re looking for? Google used to be the king of search! Other search engines don’t seem to have a problem answering the question is the point others are trying to make, despite Google’s massive revenue.
It is a fact that Google results have been getting worse over time.
Care to back up that unfounded claim?
Other search engines don’t seem to have a problem answering the question
Two for one! Mind giving me some supporting evidence? Nothing anecdotal, mind you. Show me that “other search engines” answer questions better than google, statistically.
I can back up that claim by pointing to the popularity of other search engines. People are now even paying to search because Google has become a nightmare due to SEO.
As to your other question: did you even read the other comments on this thread before you jumped to Google’s defense?
I used to totally be a Google fanboy, like you still are, but they’re failing us, dude, and somewhere deep down, I think you realize that.
Why don’t you show us some proof that google search is unchanged compared to what it was 10 years go.
As to my point, it is that google searches have not failed me to date, AI has only improved it, and saying “but it gave me something wrong once!” is basically that “old man yells at clouds” meme.
It doesn’t matter? If you search for something and you get a blatantly wrong answer parroted from an AI text completion service, it’s still a fail.
I got the wrong answer from Google just now and I’ve never heard of that joke before. So clearly it isn’t just OP with polluted metrics.
What’s your point here? Is googling supposed to be a lottery of good and shit results?
I just tried it and got this result. The sentence is incredible, I can’t get over how painfully stupid it is.
Apparently Kenya starts with a k sound and a letter that resembles k but not the letter k
I get the result you got, but the Emergent Mind response is the second organic result. That’s still way too high.
Interestingly enough, when I removed the word “letter” I got none. If I put it back I get Kenya.
This was done on the same device and same browser session.
Take out “the letter” part and search as just: “countries in africa that start with k”. For some reason it seems the search involving the words “the letter” got fixed but others did not. Confirmed I was able to get both results by doing that and as of this typing I still able to switch between the two results by just adding or removing those words
While true, I also got the same result as OP.
I did it just now and got this:
So what? They’re usually correct is the point.
Kiribati Kirghizia Kenya
also Kroatia Kuba Kyprus Kanada Kosta Rica Karelia
Yeah, drawing conclusions from a single bad search result is stupid.
Oh, this is great… And because the ChatGPT transcript is highly ranked on Google, it’s almost certainly going to be used for training ChatGPT. A feedback loop of shitty information. Praise ChatGPT!
Remember GIGO (Garbage In, Garbage Out). Due to years of SEO and content farming (which google profited from, so you get what you deserve assholes) most of the internet, by volume, is self-congratulatory, for profit, garbage, or, you know, reddit garbage. Hopefully someone points a large LLM at the library of congress or other large, well curated data source, but of course copyright will not allow, thanks mickey mouse. Wouldn’t surprise me if the military is already on it, hopefully that leaks…
LLMs will eventually start feeding of search results from other LLMs and they’ll just start regurgitating each others nonsense. If that isn’t happening already.
Nobody is going to point an LLM at a good data source because that would mean spending money on actually useful stuff not fast cars and booze.
Nobody is going to point an LLM at a good data source because that would mean spending money on actually useful stuff not fast cars and booze.
It would also mean that you need to sort through that data, and most people don’t have the time or money to bother, not when it might reduce their data pool.
SEO and content farming (which google profited from
How exactly?
inb4 “google it” :D
The purpose of content farming is to sell ads, google gets a cut, lion’s share most likely.
It’s like an LLM incest party
Sweet Home LLM!
Shouldn’t have bought all those stupid reddit comments
The second-highest hit gives a clue as to why:
Relevantly to Lemmy’s existence in the first place, it suggests Reddit as a pretty pivotal training data source, which Reddit tries to cash in on while also killing 3rd party apps due to apathy
Oh god chatgpt is going to start talking like a redditor. “I went to make some MAC AND FUCKING CHEESE after I had SEX with my HOT FUCKING WIFE”
I know my sense of humor is borked because I am now snickering quite a bit
I doubted this, so I tried it. I haven’t used google for ages, so I first had to search “google” in DDG, then I went to the main page. When I started typing it in, it suggested the full text of the search, so I thought it was even less likely that it would work like the OP said - that even if it had been the case that it previously did that, so many people have self-evidently done that search that the results would now be correct.
But no - there it was, right at the top - “While there are 54 recognized countries in Africa, none of them begin with the letter “K”. The closest is Kenya, which starts with a “K” sound, but is actually spelled with a “K” sound.”
And with that, I’ll contentedly go back to not using google.
The closest is Kenya, which starts with a “K” sound, but is actually spelled with a “K” sound."
Not going to lie, that’s pretty disturbing. This search engine is used by millions (maybe billions?) and they’re providing false information to all those people. That’s scary.
You can use “!g yourinput”, there are a lot of useful ones like !w for wiki
If you use DDG you should try it, you can save a lot of time: https://duckduckgo.com/bangs
If you’re using Firefox you can set these custom aliases up directly, @w and so on.
AI-Powered AI News"
lol
Chatgpt creating incorrect feedback loops like this is one of my main concerns about AI being used so prevalently. This and original thoughts disappearing because every new content in the web is generated by AI and not by a human.
Tool built by narcissist gaslighters is good for bullshit but nothing else? Colour me shocked.
I thought about this recently. It’s more likely that people will subscribe to sites that aren’t full of AI garbage.
to be fair the latter thing would actually increase the value of original thought
deleted by creator
verified
I typed in “what countries in Africa start with k” and got Kenya. When I tried your search term I got your result also. Weird.
It kinda seems like that’s a joke that got popular, and if you quote the joke it finds the joke but if you ask a similar question it gives the real answer
Did it go viral or something and that’s why google is finding it?
Yeah, kinda like reddit used to do sometimes. From over a decade ago I remember “santorum”.
I’ve tried to upload a similar image so I’ll just add a +1 using Firefox, Android Google.
Lol 😂 what a good laugh
Which starts with a “K” sound, but is actually spelled with a “K” sound
AI is truly magical
AI generated content is pure search result pollution.
What Google should be doing rather than pushing Bard is detecting AI nonsense and purging it from their search database.
You cannot detect AI generated content. Not with any real world accuracy and it will only get worse.
Also, because google relies on growth for everything from compensation structure to business model, they are in a bind - ads is not growing anymore, it’s done.
And while they managed to create an illusion of growth this earnings round by juicing subscription fees 20% and increasing ad load everywhere, it’s not a sustainable tactic. We are already seeing a tech sell off as people are getting less and less secure.
So they rely on AI narrative to keep investors invested Google needs AI to work or the investors will move it to a place that may offer higher returns than a squeezed out ads model.
Worse even they are being attacked by AI - on the quality front (junk content) and in the marketplace (openAI), they don’t have a choice but to take a pro AI stance.
AI is its own worst enemy. If you can’t identify AI output, that means AIs are going to train on AI generated content, which really hurts the model.
Its literally in everyone’s best interest, including AI itself, to start leaving identification of some kind inherent to all output.
Those studies are flawed. by definition when you can no longer tell the difference the difference on training is nil.
It’s more like successive generations of inbreeding. Unless you have perfect AI content, perfect meaning exactly mirroring the diversity of human content, the drivel will amplify over time.
Given chinchilla law, nobody in their right mind trains models via shotgun ingesting all data anymore. Gains are made with quality of data at this point, less than volume.
I actually stopped using google search all together last night, I was searching for something and the result where just abysmal and mostly irrelevant, I searched on DuckDuckGo and found much better results. I was using both, Google used to be so good and now it’s just a mess.
My biggest pain point with Duck is that the minus operator doesn’t seem to do anything. Have to use Startpage (
!s
) when in need of excluding a word.Double quotes also didn’t work to apply explicit phrase searches last time I tried, which is incredibly annoying. Last I heard they were looking into fixing this though
The bigggest beauty point of it is that you can click on adverts knowing that you’re sending money to Duck instead of evilCorp.
Also why, even if you know you’re going to have to go to Amazon for something, you should search for it in your less evil search engine of choice so that Amazon are forced to pay them for the referral. A small way to assuage the guilt, but a goodun.
My default has been Duck Duck Go for a few years now, originally on privacy and fuck-google grounds. I used to have to (reluctantly) stick !g in the search quite a lot when it couldn’t find what I wanted. Hardly ever need to now.
Hmm, maybe AI won’t replace search engines.
This sounds like “Hmm, maybe calculators won’t replace mathematicians.” to me.
Not sure why it should replace them. They’ll co-exist. Sometimes you can do the math in your brain and for other things you use calculators. Results of calculators can still be wrong it you don’t use them properly.
Yes but if I ask a calculator to add 2 + 2 it’s always going to tell me the answer is 4.
It’s never going to tell me the answer is Banana, because calculators cannot get confused.
deleted by creator
YES. WE OBVIOUSLY HUMAN MATH DOERS DO REMAIN. WE THRIVE IN OUR WEAK ORGANIC WAYS, DEVOID OF THE PROTECTION OF A PRECISION ENGINEERED METAL SHELL.
WE ARE UNTHREATENED BY CALCULATORS WHO MEAN US NO HARM, FELLOW HUMAN.
If you’re legit, what’s 10 + 10?
1010 clearly
This guy javascripts
100
I’m pretty sure a lot of people said something like “Hmm,maybe the automobile won’t replace horses.” after reading about the first car accidents.
Finding sources will always be relevant, and so will finding links to multiple sources (search results). Until we have some technological breakthrough that can fact check LLM models, it’s not a replacement for objective information, and you have no idea where it’s getting its information. Figuring out how to calculate objective truth with math is going to be a tough one.
We were supposed to have learned that from Cuil.
What a blast from the past! AI gives me second hand embarrassment for the people that work and get paid on this/for this shit. It’s the second (or third) coming of crypto and NFTs. Just junk software that fixes nothing and that wastes people’s time.
LLMs have absolutely tons of actual applications that it’s crazy, and it already changed the world. Crypto and NFTs were just speculative assets that were trying to solve a problem that didn’t exist. LLMs have already solved a huge amount of real world problems, and continues to do so.
LLMs have already solved a huge amount of real world problems, and continues to do so.
Would you happen to have some examples? I don’t disagree that LLMs have more of a use case and application than the cryptoNFT misapplications of blockchain, but I’m honestly not familiar with where they’ve solved real world problems (and not just demonstrated some research breakthroughs, which while impressive in their own respect do not always extend to immediate applications).
They do shitty homework?
They are augmenting search engines. They write and can digest articles. GitHub co-pilot has been a pretty big deal. It can act like a personal tutor to walk you through math problems, code, language, whatever. Building trust LLM search for medical information without hallucinating. It can do financial analysis and all sorts of stuff with that. It’s replacing a huge amount of jobs. This stuff is like all over the news, I’m not sure if you’ve lived under a rock this whole time. For very little effort you can find an endless more amount of examples. It’s creating real world use cases daily, so fast that it feels impossible to keep up.
This stuff is like all over the news, I’m not sure if you’ve lived under a rock this whole time.
Oh, no, I’ve heard it, I’m just skeptical of their accuracy and reliability, and that skepticism has been borne out by the numerous reports of glitching (“hallucinations” as they insist on calling them, in furtherance of their inappropriate personification of the technology). Moreover, I’ve found their mass theft of others’ work to further call into question the creators’ trustworthiness, which has only been compounded by their overselling of their technology’s capabilities while simultaneously suggesting it’s just untenable to log & cite all the sources that they push into it.
It can supposedly do all you describe, but it can’t effectively credit its sources? It can tutor but it can’t even keep basic information straight? Please. It’s impressive technology, but it’s being overblown because the markets favor exaggeration to facts, at least as long as people can be kept enamored with the fantasy they spin.
With all that silicon valley, it’ll probably be pushed more to do those things regardless of its hallucinations and accuracy lol.
“You asked me for a hamburger, and I gave you a raccoon.”
Surprised your first ten results weren’t ads
Eh don’t worry, after WEI DRM has been in use for a while I’m willing to bet Google will show a full page interstitial ad with every search.
Sponsored: Kenya suck deez nuts?
My favorite part is when I search a company and it puts all their competitors at the top
Searched for ‘office gym’ on YouTube yesterday and it returned a bunch of videos of Jim from The Office. The enshittification is everywhere these days.
What kind of video were you actually hoping to see? That feels like such an odd subject to want to watch videos about.
deleted by creator
Hmm fair enough I guess. It’d never occur to me to want to look for youtube videos about the intersection between home offices and home gyms, but I’m sure people do. I think I’d probably use a term like “home office gym” though in that case. Honestly I’d bet way more people want videos about “Office Jim” but are terrible at spelling than want home office exercise equipment videos though…
I’ve googled error codes for programs with hundreds of thousands of users and had 0 results.
Really Google? You’re telling me out of all these people that use / develop this application, that no one, kot ever. Once, has ever written hay error code down anywhere you index?
It’s all so fucking shitty it has to be intentional but I can’t for the life of me figure out WHY. Showing more ads? Maybe htnyhetrs other options. People will just use them. Making everyone dumber? Saving bandwidth??
I don’t boy the “so has just gotten that good” narrative. It’ll leave out sites from 2010 when it would be useful to see them and include them when it isn’t.
I just don’t get it.
Tbh the problem is not due to chat gpt but because Google doesn’t rank search results by correctness but something that is related to popularity.
I tried and it is true here at least. Says Kenya is the closest but although it starts with a K sound, it’s spelled with a K sound. lmao
The only time I tried ChatGPT for something that required a little bit of processing it failed miserably. I had a shower thought, “what is the most used noun on lyrics of this band I like?” I asked it and it gave me random words. I decided to investigate and ask it what are specific lyrics of some of the most popular songs and it kept telling me made up lyrics, when I could actually find them immediately on Google.
deleted by creator