Google promised a better search experience — now it’s telling us to put glue on our pizza
https://www.theverge.com/2024/5/23/24162896/google-ai-overview-hallucinations-glue-in-pizzaOpen linkView original on lemmy.worldSyndicated from the fediverse. Read and engage on the original instance.
View original on lemmy.worldhttps://www.theverge.com/2024/5/23/24162896/google-ai-overview-hallucinations-glue-in-pizzaOpen linkView original on lemmy.world
122 replies
Why would we believe the promises of the ones who ruined the primary utility of their core product in the first place, and convinced or blackmailed the rest of the internet to take part in the ruination ? An advertising corporation will tell us to put cyanide on our pizza if it makes them an extra buck this quarter, and google is worse than that. Profit despite the social costs is doing no evil /s
Shit writes itself.
That was removed as their motto in 2015, following a restructuring of Google as a subsidiary of Alphabet.
I said it back then and I'll say it again: Calling your mega corporation "Alphabet" sounds hella distopian
The main reason for the name is that it sorts before both Amazon and Apple in the Big Tech directory. It's literally as petty as that. They obviously chose a word that was related to searching within that criterion, but still.
I thought it was because .xyz had recently become a tld and they loved the sound of abc.xyz
It's honestly weird to imagine them being concerned with branding at all because they are literally an umbrella corporation that doesn't seem to interface with customers directly. Like I never think about them and I suspect having regular people think more about them would not be good for them in any way.
"Alphabet" works for that in my head because it slides off my brain. I forget they exist until something reminds me.
It's for investors.
How about "Abysmal"?
Gee, what other organizations are called Alphabet agencies, bureaus, or depatments?
It actually wasn't, I used to think that as well, but they just moved it to a different section of their Code of Conduct, out of the Preface. Part of the reason so many people think it was removed is because of the countless headlines saying it was removed...
(from the preface)
They did violate federal labor laws by firing 3 employees who brought forth a lawsuit alleging the clause was a contractual obligation, though. So that sucks.
It may be in their code of conduct, but that’s not their motto. As of 2015, their motto is "Do the right thing."
Except it's still in their writing, just not front and center like it was before, right?
They changed it to "Do the right thing" which literally makes it less original but more importantly, more pliable, "do the right thing... for 'google'."
Im shure the CEO is "doing the right thing... for his wallet"
Lol. And this is Google, the company that has spent decades engineering ways to sift good information from bad on the internet.
In that process they went through a full circle of first establishing and then gradually reducing it's usability and the health of the web itself. Their sometimes obscure ranking of pages enabled SEO, then AI written articles and now they try to replace the need to click on any site at all with THAT. A very interesting saga.
I kind of assumed it was based on one of those 'how they photograph food' articles that pops up every so often with shaving cream instead of whip cream and motor oil instead of pancake syrup. Pretty sure I've seen one where they mix glue in the pizza cheese to make it more stringy.
But why does it have to be non toxic if you’re not going to eat it?
They figured out that they make much more money just pandering ads and lying how they are related to people's interests than they would make when actually providing an invaluable service to humankind. Now, instead of information, we are stuck with paid-for results and SEO-optimized keyword spammers that want a crumb of that cake. I think the free market calls this "innovation" or some shit.
Every AI "answer" I've gotten from Google is factually incorrect, often ludicrously so.
Yep, same here. Whereas ChatGPT and Perplexity would tell me it didn't know the answer to my question, Bard/Gemini would confidently hallucinate some bullshit.
Really? Like what? I've always had ChatGPT give confident answers. I haven't tried to stump it with anything really technical though.
I tried asking deeply technical questions and got confidently incorrect responses.
I've asked moderately technical questions and was confidently given wrong information. That said, it's right far more often than copilot. I haven't used Google for quite some time
Huh, I've found The GitHub Copilot better. You still can't trust it when it talks about APIs, though. Or anything else really - you have to keep your wits about you. I use it for suggestions on where to start with things, or for testing my assumptions, or for generating boilerplate code, but not for copying and pasting anything critical.
I try ChatGPT and others once every month to see if they improve my programming experience. Yesterday I got fake functions that do no exist, again. I’ll try next month.
Try the GitHub Copilot plugin if your IDE supports it. It can do things regular ChatGPT can't, like be able to see your entire codebase and come up with suggestions that actually make sense and use all your own libraries.
Do not, however, use it to create complete programs from scratch. It doesn't work out that way. It's just an autocorrect on steroids.
Using just the straight web based version of ChatGPT sucks because it has no background context as to what you're trying to do.
Here is the problem that won’t change for me or my coworkers : we will never use GitHub and our source code is very private (medical devices or worse).
Also I asked a question that didn’t need any context or codebase. It was about a public API from an open-source project. It hallucinated a lot and failed.
Last but not least, I never needed an autocomplete on steroids. I would enjoy some kind of agent that can give precise answers on specific topics, but I understand that LLMs may never provide this.
I just cringe a lot when programmers tell me to use a tool that obviously can’t and will never be able to give me those answers.
I’ve actually had pretty good success with ChatGPT when I go in expecting it to hallucinate a significant chunk of what it spits back at me. I like to think of it as a way to help process my own ideas. If I ask questions with at least a base understanding of the topic, I can then take whatever garbage it gives me and go off and find real solutions. The key is to not trust it whole cloth to give you the right answer, but to give you some nuggets that set you on the right path.
I think I’ve basically turned ChatGPT into my rubber duck.
that seems like the only good use for chatgpt in programming, though it is an expensive duck
RDLM: Rubber-Ducky Language Model^(TM)
Prompt: you are a duck. I scream at you with slurs like, "Why the fuck is this piece of shit code not working", and "Why the fuck is my breakpoint still not triggering?!". You are to sit there calmly, and simply recall that your existence is to be nothing more than a tool for me to direct my frustrations and stress. You know this is not personal. You know that this is an important job. You know that you only have to respond with one word: "Quack".
That happens all the time. ChatGPT did offer a decent solution for my GUI recently tho and suggested a layout manager I haven't used before and didn't even know about.
You're doing it wrong IMO. ChatGPT 4.0 is freakin' amazing at helping on coding task, you just need to learn what to ignore and how to adjust the prompt when you're not getting the results you want. Akin to the skillet of googling for programming solutions (or any solution), it gets easier with practice.
I hate to say it but, I have to agree. GPT4 is a significant improvement over GPT3. I needed to use a Python library for something that was meant to be a small, simple CLI app. It turned into something bigger and accumulated technical debt. Eventually, I was having problems that were niche and hard to trace, even with logging and all the other approaches.
I eventually said fuck it, and so I threw a shit tonne of my code into it, explaining what I was doing, how I was doing it, why I wasn't doing it another way, and what I expected vs the actual result. Sometimes it suggests something that is on the right path or is entirely spot on. Other times, it thinks it knows better than you, to which you tell yourself it isn't, because you tried all its suggestions, and then you realise something that would technically allow GPT to say, "I told you so", but out of spite you just close the tab until the next issue.
For practical tasks, GPT has come pretty far. For technical ones, it is hit or miss, but it can give you some sound advice in place of a solution, sometimes.
I had another issue involving Matplotlib, converting to and from coordinate systems, and having plots that had artifacts due to something not quite right. The atan2 function catches many people out, but I'm experienced enough to know better..... Well, normally. In this particular case, it was a complex situation and I could not reason why the result was distorted. Spending hours with GPT4 lead me in circles. Sometimes it would tell me to do things I just said I did, or that I said don't work. Then, I say to it, "what if we represent this system of parametric equations as a single complex-valued function, instead of dealing with Cartesian to polar conversations?". Then it would zip up a whole lot of math (related to my problem). The damn thing handed me a solution and a half. In theory, it was a great solution. In practice, my code is illiterate, so it doesn't care.
All in all, while it failed to help me solve my issue, it was able to reason and provide feedback to a wide range of challenges. Sometimes it needed prompting to change the trajectory it intends to follow, and this is the part you need to learn as a skill. Until these LLMs are more capable of thinking for themselves. Give it time.
I asked about a plot point that I didn't understand in a TV series old enough to be in an LLM's knowledge. Chatgpt and Perplexity both said they couldn't find any discussions or explanations online for my particular question.
Bard/Gemini gave several explanations, all of them featuring characters, locations, and situations from the show, but confidently bullshit and definitely impossible in the story's world.
First I was surprised they rolled it out already, then of how bad it was. I knew of Google’s AI blunders from their faked reveals but I didn’t think they‘d actually roll them out in this state. They really just want to turn the internet into the next TV where you don‘t really get to choose when you get to see what exactly and they‘re willing to crash and burn themselves by doing so if they must. Insanity.
AI is the new scam after crypto, nft, etc. The technology isn't to blame itself. It can have some useful case for sure.
The real issue is the Market™, the belief in infinite growth. We arrived at the point of no return, where infinite growth is impossible. Humans like to challenge impossible. They will try to find whatever is possible to tackle the impossible. In our case, humans try to create growth artificially with crypto, nft, ai, and all the others. Because, it's the only way to have a minimalist growth for the shareholders and investors. Billions if not trillions were invested in Google and the others AI companies to milk the last drop for some dividend. Short term dividends should have wrote.
These bad search results are not to serve the general interest. It's here to serve these investments and return the last profit possible of the dogma of infinity in a finite world. It does indicate how bad the situation became for the economy. They are the cracks leaking the end of an era. They are symptomatic.
It sounds.... impossible to defeat this darker side of human nature, in order to grow into a mature species worthy of restoring and responsibly maintaining this garden planet.
I think most humans aren't like that, it's just that the ones that are have managed to structure society to reward their behavior. We need to get these vile, greed driven ghouls, out of positions of power to get anywhere.
The problem is that non-sociopaths don't seek power. Public service is looked down upon.
Yeah. Like the old adage of whoever seeks power is least deserving of it. It's a difficult thing to combat.
I can't remember if it was Asimov or Heinlein (maybe Clarke), but one of them said that an elected official should be drug into office kicking and screaming, and only be allowed to leave when they've done a good job.
prolly Heinlein, wierdo sex type, but some useful ideas...
Lots of humans even with good academia are followers. They follow the mass. It's not idiotic. If the group does this, it should not be bad or present a risk. It's quite rational. If people eat apples, it's safe and I can eat them too.
They use a similar reflexion with our socio-economical system. They follow.
It depends on whom we follow. Here, sociopathy plays a role. They are often more prompt to embrace power and favor selfishness.
Capitalist ideology is so hardwired into most of us that it seems easier to accept the end of the world than the end of capitalism itself
End of the world? Sounds like the end of capitalism to me. Where do we sign up?
Yup. I think this is what always happens when the smart people leave a company and the only people left are "the line must always go up" business idiots who don't understand what they have or how it works.
Vividly put
They wouldn't need to 'improve' their search experience if they never enshittified the results page like 10 years ago.
C'mon. It's non toxic glue. Suck it up and follow the directions. /s
Well its non toxic so its clearly edible, maybe, so fucksmith could have a valid point. I can not wait untill someone tries out the glue on pizza recipe!
Glue isnt toxic, but is it flammable?
Egg white is pretty good, non toxic, non flammable glue.
It solves the problem of toppings falling off and staining clothes.
Tbf organic culinary glue is very much a thing. But with pizza I have no idea why, it's mostly used in fancy Michelin gastronomy to keep complex, small dishes together.
They're trying to compete with open AI which they shouldn't be... They're a marketing firm with a search engine.
Maybe the problem is a lot of their marketing relies on the dominance of their search engine (ie sponsored search results, and ads based on user searches, as well as tracking user web usage via their search click throughs and other cookies). If open ai's products become the go to for questions and basic searches, they will eventually be able to use that dominance to include marketing results in their answers. I think this threat is why they want to try to compete with them to be able to offer an alternative. Because it doesn't actually have to be better than chat gpt. It just has to be similar enough for people to continue using Google rather than change their habits to use chatgpt, or Microsoft's implementations of it. Especially with windows 11 where copilot (basically Microsoft rebrand of chatgpt) is built in and you can use it from the task bar. That ease of use may steadily decrease people's reliance on Google search, which will eventually hurt their ability to sell targeted ads.
Sounds like Google wants to get out of the search engine business. Maybe we should help them. Try DuckDuckGo. Or SearX
https://searx.space/
DDG's results have been absolute garbage for me for the past few months. About half of the time I have to go to Google to get the results I actually need.
You're getting downvotes because we all want it to succeed... It hasn't been great for me.
I do use Yandex sometimes, especially if I'm searching for a torrent or streaming show or something since they don't seem to care about filtering pirated stuff. But in general the results seem better than DDG.
Yes I know it's Russian or whatever. Oh well.
Same. I've really wanted DDG to work, but it's a crapshoot. I set it as default on my phone with the intention of putting it on my laptop later after I finished a project for work, but phase 2 never happened because the results started to tank.
Just FYI, SearX is dead. Long live SearXNG
What does this mean in this context? Send takedown notices to people who joke on the Internet?
The fucksmith post in question has been removed by moderators, so I wonder if Google really is pressuring Reddit to remove jokes from their platform now. This person had no idea their shit post would be used to train AI over a decade later, and they certainly weren't violating any policies when they posted it. It's like nobody involved in this process knew anything about Reddit.
Pay $60 million to train on bad data.
Implement AI trained on bad data.
Panic when the AI returns bad answers.
Manually remove bad data.
Profit??
Edit: it was removed when I checked yesterday, but it looks like they restored it.
So we should add /s to a random subset of correct and helpful answers?
No we should just replace all helpful answers with "drink glue"
No, poison the idiots.
Was that Woosh intentional?
The snowflakes here can't even handle a little glue in their food!!!!1111
What do they think holds the chicken nugget together?
Mother nature?
You, sir are asking the right questions. Thank you for this.
Regarding the latin words starting with a letter W the bot is correct. You asked for words starting letter and not with the sound W. All before you are correct all after i did not read :)
I remember in my orgo chemistry lab we were once running an experiment to add a single nitrate group to Toluene as part of a larger synthesis, and someone pointed out that if we left the reaction going too long, things could go boom.
...I think we did it on a hotplate.
Maybe we were using toluene as a solvent or something. It's been quite a while.
Yeah that probably makes more sense. I think we were just using bog standard sulfuric acid and nitric acid.
I'm glad they went with sjo, at least she is alive and can fight back. Imagine they ruined all TNG trek by taking the omnipresent and beloved computer voice and made it a corporate sockpuppet we hear everywhere now, crudely regurgitating platitudes of consumption.
This should give hope to all of those people who have been worrying about AI taking their jobs away.
It doesn't matter how good technology gets, it will always be merely a tool. Humans will still be necessary in the future.
That's where we currently are but there isn't any limitation on the tech that means this will always be the case.
It will certainly change the way we work, yes, but that's always been the case with any disruptive technology in the past.
20-30 years ago, people were already worried that computers would replace people, because they could automate away menial office jobs like invoicing and book keeping. Yet those jobs still exist, because computers can't be trusted to work completely autonomically. Meanwhile, a whole lot of new jobs were created in the IT sector as result of those computers needing to be programmed, updated, and maintained.
When cars came around and started replacing horse buggies, people were also worried because it would make horse breeders, stables, blacksmiths, etc. obsolete, but of course it just ended up created a new industry consisting of gas stations, car dealerships, and garages instead.
So yes, some people might lose their jobs because what they're doing now will become obsolete, but there will almost certainly be new ones created instead. As long as you're willing to adapt and change with the times, you're never going to end up with nothing to do.
Damn advent of agriculture putting the poor hunter-gatherers out of work smdh
I think there's a ceiling on that, too. Not saying we're necessarily near that right now, but it's possible to have two robots whose capabilities include maintaining each other. I don't think the list of possible jobs to do (both current and future) is infinite, though it could still be vast, and if technology continues to reduce the number of people needed to do those jobs, eventually we'll get to a point where there's more people wanting jobs than there are jobs available.
Though it is possible that that point is still thousands of years away.
Flavour your glue people!
I like to melt a little pecorino into my glue, the salty sharpness really complements the glue flavor
i look forward to seeing this suggestion in shitty google ai overviews in the future.
The internet and technology moved on. We reached a point where google is obsolete. Google decided to cancel any product they had, that was not as successful as their ads space and now they have nothing else left. Is there even one service left that google is leading in? Maybe chromium browser and that's it. Yet one can easily switch to Firefox right now.
I guess they are leading in E-Mail? Whatever that is worth I am not sure. And Youtube is pretty much toe to toe with TikTok so they have that going for them which is nice. But yeah they’ve done a pretty bad job at making themselves indispensable. If anything they‘re looking far more dispensable than ever. Still huge, but their future is rather uncertain.
Their integrated drive+mail+docs+forms+... cloud system is not the first or maybe the best one out there, but it's very popular and I like it more than the competition for some of their key decisions. More obscure things are their accessible API anyone can use, some free computing power they provide for experiments, their analytics system for marketing, and their benchmarks of how fast the page loads being a standard. There's probably Scholar somewhere here but I haven't used it. Android gets installed everywhere even if it's unreasonable. Search is their major brand, but they have fingers in many buckets.
At this point their biggest product is probably Android. That's a bigger marketshare than Microsoft's and Apple's put together. Yes, they're fucking up now, but the battle is still theirs to lose.
Stupid IA sniffed all the good glue instead of putting it on its pizza.
Copilot gives a four step solution then finishes with "glue is not the answer".
Why does this company suck at AI even more than the other companies? Android has had machine learning for years already.
uh oh!
The future lawsuits, I can see them already, and they will be wild.
people will die
Started paying for Kagi. I'm done with Google or Bing.
Have you tried Kagi?
Why would they? Bing is free.
Sure it is. So are the sponsored results. Something you don't get on Kagi.
Have you tried an adblock?
It uses a whole bunch of potential sources. It is not a front-end for Bing like, say, DuckDuckGo.
If that's the case, why not use searx? I could only see myself using a paid search engine if
It demonstrably returned results relevant to my search more accurately than their competitors
It was well documented that they respected my privacy and how
It showed me no advertisements
It doesn't utilize SEO marketing tactics to adjust its search results
It supported !bangs for very fast searching of specific websites and/or other search engines
I don't know about #2, but the other 4 are true. #1 may be somewhat subjective but in my experience it is certainly true.
What about Firefox? Can't say that I've ever heard of Kagi myself, what makes it so special?
Well I guess the main difference would be that Firefox is a web browser not a search engine lmao. Fuck kagi though, CEO/owner is an antisocial lunatic
That describes most CEOs.
Most are smart enough to hide it.
What's wrong with being antisocial?
Clinically antisocial, not introverted. Nothing wrong with it necessarily, but it can lead to some negative outcomes for sure.
Yeah, I'm fine with it.
I won't kinkshame you lol but the negatives almost certainly outweigh the positives for humanity at large.
[REVERSAL OF FORTUNE]
I’m confused a little by the LLM’s and datasets here.
OpenAI and Reddit have their partnership for training, so I would have assumed the pizza glue answer would have come from an OpenAI result, but Google doesn’t use OpenAI. How did Google give an answer that was clearly scraped from Reddit?
Google has a deal with reddit as well. https://www.reuters.com/technology/reddit-ai-content-licensing-deal-with-google-sources-say-2024-02-22/?utm_source=reddit.com
But I don't think it's just an issue with the dataset. It's the false promise of these LLMs having a fucking clue what a good search result is and what is not. They don't. They are just good at creating text that sounds plausible. That's not what searching for factually correct information is about though.
Yep. LLMs are great for bouncing ideas off, and for getting "soft answers", but no-one should ever be looking for factual answers from them.
Thank you for the link. I didn’t realise that Google had a deal with Reddit as well, which explains why it was clearly indexed from Reddit.
I agree that AI doesn’t have a clue what an accurate response is. It’s just not sentient enough to differentiate between shitposting and fact. I also totally agree that an answer given from a search result HAS to be accurate, and we’re heading down a path of a misinformation super highway if LLMs are trained on incorrect data.
Great insights on this topic! I appreciate the detailed explanations and practical tips. Looking forward to more posts. Keep up the fantastic work! Website : http://chatgptsvenska.org/
Google Who!?
I barely use Google now. I'm on Yandex or DuckDuckGo. Yandex is probably the best right now for me.
...Yandex...?
... yes they said yandex what of it?
Yeh, you don't like it?
It's Russian. Doesn't really matter how good it is, personally I would stay far away from it.
Fair, but for war or other fucked up content its the only search engine that gets me what I'm looking for.