Somewhat related: did you know that you cannot disable GMail's GenAI features without disabling the traditional ML powered features, such as smart inbox, which predates LLMs? In this case, Google/Alphabet is conflating GenAI/LLMs with traditional machine learning, so you can't opt out of genai---which necessarily requires access to your data by the models---without deactivating the previous non-LLM powered features.
Which I get it, they indeed can use LLMs to improve the previous traditional ML pipeline. However I believe bundling these options together shouldn't be legal and I never expected Google to employ dark patterns like this. When I subscribed, these features were provided by traditional ML, perhaps some BERT-like encoder architecture at most, however now I cannot keep that previous functionality if I don't wish to grant LLMs access to my data.
I did this (disabled AI features) and lost smart inbox and it honestly improved the experience: it made me realise how much rubbish I was subscribed to that I never read, so I unsubbed and blocked tons of stuff, and now my inbox is clear again
It is absurd. I disabled Gemini for my organization, and Google in return disables tab completion in gmail entirely. They are practically demanding I find an alternative at this point.
I used to be able to search for parts of names. Or a word that is part of a word in a message. Doesn't work anymore. Only exact matches now, or you don't get anything.
It is even worse if you commit the offense of searching for something with an unmatched quotation mark. (Like 1-1/2" for example). Zero results, every time. Makes searching for trades related information in my emails a pain.
Huh? This is not the case for me at all, just tested. Not doubting your experience, but it's weird that such a fundamental aspect of search (not even targettable to anything) would differ from user to user.
Just to clarify, we were discussing gmail search and not the main web search. I just tested again, if I search for 1-1/2" (with the quote mark) I get zero results.
This is google-for-business or whatever they call it now.
Why didn't you expect Google to employ dark patterns? They're pioneers of dark patterns in software. They've probably done more than any other company except Meta to spread and advance these techniques.
BERT is an LLM. I don't see why Google needs to cater to increasingly fine parsing of politically acceptable vs. not ML workflows, because that is really what this is about.
I know, I mentioned it because there's a category of models that I'd be okay with, but I can't exactly name it. I thought the difference lied in the encoder/decoder distinction. As a sibling comment mentioned, generative models do offer more privacy risks.
But it's not only that. Model size, for instance, also plays a role. Will modern BERT keep their lossy compression nature? Or could there be a privacy risk
as well?
Thus, what are the characteristics of privacy-aware models? So far: discriminative vs. generative, model size, and lossy transforms (related to size). Perhaps looking at pre-2020 models could help to narrow other necessary attributes.
Model size, definitely. Not so sure about the discriminative vs. generative distinction (technically 'generative' models are discriminators as well) - I suspect the data is extractable if you have access to the weights, although easier to extract from inferencing in a decoder model.
My guess is that the pairing of all of these things is due to the fact that Google has likely consolidated everything under the same style of decoder model.
To me the distinction is in whether model trained on my data can return it almost verbatim. Any statistical model, even linear regression can expose some of my data and preferences, but the difference is whether they can compromise the data they've been trained.
And frankly I see the distinction in only number of model parameters vs number of people in cohort it was trained. So 50k-param model trained on 1b emails from 1 million people is ok, encoding/compression loses anything personal. But huge LLMs are closer to lossless compressors that can just spew all my passwords, prescriptions etc verbatim.
You need to do this to continue to use Google Assistant for some features as well! One day I try to set a timer on my phone, which is basically all I use the voice features for, and it forced me into Gemini AND data collection. It's truly maddening.
I just abandoned my 22yo gmail account. I feel liberated and empowered!
I found all of the new LLM features to be utterly pointless and incredibly annoying, but the last straw was when my account was abruptly locked recently. For a week I couldn't access mail, maps, photos, digital picture frames for my extended family all went dark, my personal blog ... all shut down with no reason given. No customer support available to contact to resolve google's mistake. A week later, it was suddenly unlocked again. Still no reason given.
Fuck google. You are not their customer; you are their product. Google doesn't care about you any more than Tyson cares about the chickens it sells.
Don't remind the Googlers in the thread it exists! The only Google services unmolested by garbage are the Google services nobody remembers exists that still just like, live on the servers underneath everyone's attention. When you find some Google website with their 2010-style UI you know you've found one of the last remaining nuggets of gold.
On a similar note, I recently had to enable GenAI in my Photos app just to be able to search my photos, and the photo search is now worse than I remember it being before! I need to move off Google entirely for backups but I find the cross-device compatibility and convenience of the ecosystem really hard to give up. Lately, I've started using other LLM plans than Gemini just to have some of my personal details separate from the rest of Google, which has all my documents and email, etc.
I hope that enshitification calms down a bit whenever the dust has settled on frontier AI companies and we no longer have to force this into every product imaginable. Google used to be pretty good relative to the rest of the big ad tech companies (looking at Meta/Facebook with a scowl...) at giving opt-outs and settings to personalize privacy stuff. This has been a disappointing change to see
They can enshittify their product as much as they want, it just makes the case for removing every piece of Google garbage out of my life and never using a single piece of software from them.
>Google/Alphabet is conflating GenAI/LLMs with traditional machine learning
I see people try so hard to make this distinction when it really doesn't exist. Transformers are only a little different than long short-term-memory architecture (parallelized, with attention mechanism and no recurrance).
What is the point at which is departed from tradition machine learning? Seems more like people retrofitting a moral decision to not like LLMs and trying to dress it up as a technical distinction.
I can run a linear regression and get the same result every time or I can go to chat.com and get A/B tested upon. yes these are cherry picked examples. they're also the ones that apply to me personally. for a higher level version of this argument please see my response to the other responder.
when concept A applies to concept B in practice, a lack of theoretical connection between A and B is basically irrelevant to me as a lay person. people don't seem to grasp this argument well here so I'll be very clear, yes you are entirely correct, concurrently there isn't much reason for me to care about correctness when my eyes and ears tell me something else
I'm confused, so google made websites indirectly unusable with all the advertisement incentives and then the AI summary came that usually just summarizes some reddit posts and now they subtly and indirectly block universities from using their entire search engine because unis block the AI part of search? What incentive does google even have on presenting the ai summary?
Does anyone even understand the world currently? Its so absurd to me, why does it exist to begin with and why is this drama appearing?
They want to make everyone cognitively dependent on AI so they can further insert themselves into our lives for profit. The younger they hook us the easier that will be.
Not just universities. If it were universities it wouldn't be so bad since most students have their own laptop and can make their own decisions. This is a play to get AI in the hands of young kids. Most public K-12 schools use Google, so they either have to opt-in to AI and/or their 'helpful' educational chatbot, or not use search at all.
Have you noticed how many people already use AI to do their thinking for them. They just ask it what to think, what to do, for everything. They want that for as many people as possible. They don’t want people going outside of Google to other websites to decide what to think. Yes they will put ads in it eventually. But in the end they will also use it to influence what the population thinks.
We already have some leaders and politicians outsourcing their thinking in this way. It will make these companies incredibly powerful.
if you can go direct like that and influence peoples basic schema of thought, there is no further need for contemporarily structured ads.
as far as ads go, the razor cuts to the bone when you realize that microtargeted ad campaigns, are overly complex, and have folly. what you really want is to predict someones shopping list, rather than recommend items that were already purchased in gross quantity.
This is not different from standard SEO. You adjust the page to be more likely to select in the results. That's not the same thing as that LLM providers injecting explicit ads or steering results.
More-diffuse manipulation become possible too: The National Dairy Society or whatever can spend $ on generally boosting mentions that calcium is an essential mineral and that you can get it from certain drinkable sources.
Baking ads right into the weights would be extra evil, so they will probably do it. But, also, it would take me probably 10 minutes to write a skill to, say, call an API to retrieve advertising instructions and incorporate them into the output.
> What incentive does google even have on presenting the ai summary?
I have some guesses:
They can artificially inflate the number of users who use their AI by including everyone who uses their search engine.
It trains people to be increasingly dependent on whatever google tells them instead of checking and reading multiple sources, and verifying information for themselves.
They hope to let people pay them to manipulate what their AI returns for certain searches (either mentioning certain things/products or suppressing mention of them).
Google search is getting increasingly useless as the unending flood of ad-filled AI slop websites dominates the search results and people will abandon searching with Google if they can't find some way to provide people with the information they are looking for. (it's not going well so far https://arstechnica.com/google/2026/04/analysis-finds-google...)
> They can artificially inflate the number of users who use their AI by including everyone who uses their search engine
Yup I could see that being plausible. It’s like Microsoft and Copilot. Project managers got to boast how they got AI “adoption” numbers up just by shoving it into people’s faces.
Kagi depends on Google's search API still. So they get paid either way, I guess.
I do sort of wonder how much longer Googles search API will even be a thing. They clearly have no interest in running a search engine anymore and are becoming actively hostile to the open web.
Kagi uses Google API as one among a number of sources, I don’t know if it’s a particularly vital one though. Personally whether it’s Kagi or something else I think the future of search will be tiered:
Annoyingly though, Reddit which apparently has some fucking stupid exclusivity agreement, results, are still a lot of the time found on Google, not even with Kagi using Google's API
Google's search is increasingly AI driven (as it should be, i do not want to click on reddit links and waste my time skimming through the 1000 comment threads), asking Google to maintain a non-AI search engine is like asking Tesla to come up with an ICE engine car right now. It is not happening.
Google Search I would expect will increasingly become more like Gemini and less like traditional search. Schools could certainly go to DuckDuckGo or other search engines that are free. (Google has shown that it does not care if they do as it sees it as a dying business.)
That depends on whether the purpose of web search is to get answers or find references. Since students are still developing their reading and research skills, the latter goal is much more relevant in an educational setting.
AI boosters have a vested interest in conditioning people to treat chatbot output as primary sources when they most definitely are not.
The problem is that Google's AI search basically doesn't work. It routinely can't understand even simple prompts and is the only AI I've used that routinely hallucinates.
And what, pray tell, will this search engine search?
If nobody is visiting anything on the internet, because every search engine response is an AI output, then where will any information for this AI output come from?
This paradox was clear from day one, pretty sure I posted on here about it years ago. AI search will kill the dataset that AI itself relies upon.
Peer review is what made information on sites like Stack Overflow and reddit have value. One on one conversations with an LLM are not an equivalent and direct substitution.
> asking Google to maintain a non-AI search engine is like
...is like asking your company to maintain its SQL database instead of shoving everything into a document the LLM reads and suggests edits to.
Even if you're using an LLM to talk to the user, there still needs to be some kind of deterministic search engine back there in order to do things well, or to answer any kind of question that involves information newer than X months ago.
There needs to be retrieval, but ranking can be fully offloaded to LLMs and they would probably do better. Legacy search engines do both parts, and their ranking is highly gameable judging by how SEOed they are.
Spoiler: regular search ranking doesn't see very many results, either. I mean, have you seen the results for a crowded query like this? They are trash.
Yeah yeah, search is hard and gets adversarial, we already knew that.
What's missing here is your imagined alternative: What exactly are trying to propose where Google can ditch having a search engine, and why do you believe it work well instead of badly?
I think they're just asserting control. They want to control what you see on your devices. They own the worlds largest search engine, and they think they get to pick what's on it.
World makes complete sense. Its just ran by properly fucked up sociopaths who can't ever reach happiness so they hyperfocus on competitive zero sum games with short ego boosts and with no regard for anybody else, future of their kids or anything larger. They think their legacy like that is something that will be admired for generations because they got a lot of money and who cares how they got it, right.
Most will end up in deep regrets, ie how jobs regretted being an asocial dick and all above to his own family before dying (thats the better outcome, many simply see them as stars of their show till last moments, like potus for example). Its not some secret cabal, these people if smart naturally bubble to top of social and financial hierarchies.
It used to be much less visible, some sort of on-surface decency momentum and shit behind the curtains, now everything is out in the open for whole civilization to see instantly and its pretty tiring and nauseating.
If you don't get it its fine, it means you are a normal human being, probably well balanced. Folks like that abhor such people personally, be it c-suites of corporations or politicians or just a pos colleague.
Google is willing to sacrifice children to push engagement. In family link they refuse to let me restrict the Google app. In what world do I want my 1st or 2nd grader opening the Google app and reading the news about murder, rape, and all the other horrendous things that are common in the Google apps news summary? Give us parents actual control in androids parental control app!
I think we've reached a point where reading the AI summary (which I think is Gemini Flash Lite) is probably going to be more reliable than actually reading the top 10 search results. The top 10 search results were often blogspam even in a pre-AI world. I feel like either you trust Google to retrieve information or you don't (and Gemini has problems but so does Google Search, and are you really sure the latter is worse than the former?)
It has been over a decade since Google Search was anything but a machine-learning powered word association engine. I miss old Google Search but it's dead and I trust Gemini more than what Search has turned into.
That is NOT my experience. I used to have a dump of screenshots where Google AI from search is absolutely wrong. In my experience, it doesn't work and "quoting" things never told in the source (the exact contrary sometime), responding in hindi, responding in spanish, responding in mix of languages.
Most of the time it's quicker, sometimes it's just bad.
It hallucinates so frequently, that I've found it unusable. Just ask it to provide sources for all of its claims, then watch how many times it reverses its original answer. It's mind-boggling to me. I've repeatedly tried to give Google AI Search a chance, but after weeks of seeing this behavior I can't trust it anymore.
Imagine winning search by being the fastest and most accurate, then replacing it w/ constant hallucinated errors.
For me, I use it maybe ones or twice a month when my traditional search on DuckDuckGo fails. And, anecdotally, it is overwhelmingly wrong, and worse, confidently so.
Given my sample is biased as I am using it for an outlier of difficult web searches (including web searches localized to Iceland). But I would think that if your parent was correct, at least majority of my uses would be correct, but in my experience, Gemini fails most of the the time traditional web search fails. So to me at least, it is effectively useless.
For people who use it in place of normal web search, it is probably worse then useless, because of how confidently wrong it is.
The argument is not that Google's AI summaries are infallible.
The argument is that there's so much garbage on the open web that clicking through to the source has a lower signal-to-noise ratio than the AI summary.
This rings very true in my experience.
Google's AI summaries definitely make mistakes, but in most cases they're higher signal than the source websites.
You're underestimating how obviously, laughably wrong the AI summary often is.
I really think Google is making an error using such a lightweight model for the page. That's one of my only firsthand exposures to genAI, and it's awful. Use a decent model, or if that's too expensive don't put up a summary at all.
Once it quotes a gov website, exactly the contrary of the official source.
Why ? Because on the official website you have checkboxes and the div shown is the one corresponding to your case and most surely the AI summary didn't have information about which one is which and just mix everything up.
"reading the AI summary is ... more reliable than actually reading the top 10 search results."
I don't quite understand what this even means. I both use AI chatbots frequently and still use keyword-based search engines frequently as well. When I use keyword-based search engines, it's usually because I want to see what sort of websites, discussions, resources, online hangouts, etc. exist around a given term, or just how the term is used. I often don't have a specific question or task in mind and just kind of want to wallow in the search results for a while, sort of like if I moved to a new city and just wanted to walk around different neighborhoods. otoh when I use an AI chatbot, it's usually because I have a specific question to ask, and often I don't know what the relevant search terms would even be. The use cases seem fundamentally different and I don't know how one could be "more reliable" than the other.
The summary isn't limited to the top 10 search results as sources so it need not be only based on blogspam like the first page of search results is/was. Whether or not it's actually better or not is still its own debate beside that point though.
I haven't checked this month, but the few times I used Google Search this year I saw egregious errors in the AI summary ~50% of the time. Things like confidently citing a statistic whose number appeared in the top 10 results, but misrepresenting what it pertains to.
Anecdotally, the percentage is way higher than 10% for some kinds of searches.
I don't always use google, but i've tried it dozens of times the past few weeks to test their AI Summary / AI Overview, and it has yet to get anything correct. I've had to use other search engines or AIs literally every time.
mistakes vary from "obviously wrong" to "looks vaguely correct, but the source states something different" to "looks plausible, but the cited page is completely irrelevant, the AI clearly made this up" Even for straight fact lookups, I've seen it cite pages and claim one number, and then i click through and the source matches the vague shape of the answer but has a different number.
Its just constant hallucinations. "thats a common problem, just change this <setting that does not exist> or set <option that does not exist>". Full gemini seems more capable, but whatever google is throwing at the top of their free searches is awful IME.
I'm guessing they're only counting one failure mode. But I think it's the one that would surprise most boosters. Even if it found a relevant page, and the information is fed into the model's context right there, the output is still completely backwards a fair fraction of the time.
All we’re doing is adding another layer between us and the information. Google flash is still pulling on the same poor information. But it’s not as good at discerning what is useful and what is not as we are (people going way too far out on a limb outside of their domain expertise not withstanding, which is another part of the problem and why this stuff can be dangerous)
Can’t access the article, but would assume the schools are more concerned about students having Google AI doing their work for them, rather than concern about using the AI summary for search results?
at least before I was able a find a blog post or a discussion or fleck of documentation. Now I've pretty much stopped using Google because I've wasted so much time having been assured by Gemini that something is entirely possible when in fact its very much not. now I have to dig to find the canonical reference for a question, at least SEO people didn't spend a lot of energy trying to front rust crates.
This is all very tragic. we built a magic network where we could stash all the worlds information and access it from anywhere, except its been trashed by people with sleazy motivations trying to get between us and what we'd like to see.
What? No. You're completely wrong. I frequently ask about "X" and the summary cites a single random Reddit post that doesn't even MENTION "X" as source for its assertions. That literally never happened before AI.
Absolutely. Google has shown the courage to admit that their bread and butter business is changing permanently. they are adapting which the large cos rarely do.
My mother never used search. She found it too hard to manage. She uses Gemini all the time, it gives answers in her language and she does not need to type. She is incredibly happy.
My daughter talks to Gemini. The idea of going through links seems primitive to her.
The future of interface is chat/talking not going through clickbait.
I think schools should have some basic control on how kids can use AI but saying that they want to freeze the learning experience in 2020 is simply not right. It is like saying we wont allow whiteboards and markers.
People just really, really, really hate paying for stuff. Hell Google even lets you continue to use their services even if you block their ads. They're that generous, err, evil.
I don‘t think that is so simple. Many schools have bought fully into the Google ecosystem, with Google Classroom, Chrombooks, etc. I used to tutor at my local high school, and for some reason traffic to Duck Duck Go was blocked, and I was forced to use Google search when tutoring my students.
These are still a thing for people who don't like JavaScript, so I'd give them some credit.
And it's not like these don't cost anything to run and maintain. Quite the contrary: bots love to use them, and it screws with their ads-driven business model. And yet they are still keeping these around.
I think that "noai", too, will keep going for a long time.
> We believe schools should have control over AI in their classrooms and in providing administrators with controls to help them manage their approach. After listening to schools, teachers, leading researchers, and policymakers, we are now updating our approach to controls for Gemini, Gemini Notebook, Gemini in Classroom, and Google Search. Access to these products will be entirely off for K-12 student accounts until school administrators explicitly turn them on.
Google blog says search will be off by default, alongside Gemini stuff.
And NYT says they would only be able to turn search on if they enable all of the Gemini stuff. That is not mentioned explicitly on google's blog. So not _that_ clickbaity?
> … and Google Search. Access to these products will be entirely off for K-12 student accounts until school administrators explicitly turn them on.
So yeah the blog post says Google Search itself will be entirely off until being turned on. There’s nothing in this blog post to indicate whether Google Search can be turned on without also turning on Gemini. Although I hate the NYT, the NYT is quoting someone from Google saying there won’t be separate controls for Gemini and Google Search.
From the article, Google AI is deeply integrated with Google Search. So the article is correct - turning off one turns off the other.
From a privacy perspective, having all of this off by default is sensible. But only offering a basic product like Search if you opt into AI (when search as a capability does not need AI to function) - is newsworthy.
> when search as a capability does not need AI to function
That means a far less capable search engine. Note that Google search has been AI-powered since before AI became a thing; it's now just more advanced, and particularly, explicitly publicized.
Does Google tie Gemini access for schools to their ability to train on the data and use it for ads and measurement like they have done to Android if you want to do anything halfway useful you used to do with Assistant before they killed it?
> A.I. features are now a standard part of Google’s commercially available search engine. “Search and A.I. are deeply integrated,” said Chris Phillips, vice president for education at Google. If a school has banned A.I., he added, students “will not have access to search, and they’ll see a message from their school that indicates it’s not available.”
C'mon, it's right in the article.
The school can control AI on/off. Off means no search too.
The local population is probably allowed to complain that you built the reservoir with toxic waste, too, or that you filled the old well with construction debris.
It's not a goalpost move though, as there was never a guarantee to remain on well water.
But anyway I personally would find someone offering well water if I don't like the reservoir. And I've been using DuckDuckGo exclusively for search since they first appeared, so I've never had to deal with what Google's doing. I know the reservoir exists, but I've never drank from it.
I blocked `User-agent: Google-Extended` but allowed `User-agent: Googlebot` on robots.txt. Later I noticed my site is not visible as usual in their search results. Reverted back 2 days after.
Google Search uses lots of AI. Google is now allowing schools to turn off all AI entirely. That, obviously turns off Google Search. This is the feature working as intended.
It seems like school administrators will get to control which products get AI, individually. Which is what they've been asking for:
> After listening to schools, teachers, leading researchers, and policymakers, we are now updating our approach to controls for Gemini, Gemini Notebook, Gemini in Classroom, and Google Search. Access to these products will be entirely off for K-12 student accounts until school administrators explicitly turn them on.
Am I missing something here...? This announcement seems to be 100% good, but people are writing like it's 100% bad.
Hopefully the schools that reject ai use in the classroom are savvy enough to pivot to an alternative search engine, easy workaround for them. Google is pulling Philip Morris moves and we should all abandon them.
I'm glad I'm not the only one who has made the cigarette connection. It's interesting to note some of the cigarette companies pivoted to junk food after cigarettes went out of fashion. The business isn't really cigarettes, it's addiction.
AI literally makes students stupider, if any of these anti-human demons believed this they would force their own children to openly use these things but we always see time after time that the rich big tech elites will never subjugate their own children to this madness but see no issue with experimenting on children across society so they can make a few more dollars ratfucking the earth.
It can go both ways. It can be really helpful for people who want to learn concepts, similar to having a private tutor (but free/cheap, and available 100% of the time). But it's also easy to use it to skip the learning process completely, and just get quick answers for homework.
As an adult, wanting to understand concepts for my own benefit, I've found it really useful as a learning tool. As a teenager, I probably would have been irresponsible with it and sabotaged myself.
All computation offloads cognitive functions. That's what they are for. You could also read a bunch of index cards and decide for yourself which results to shortlist when doing research, instead of using a search engine.
Kind of makes sense, given how integrated Gemini summaries are to Google Search, last I checked, but ultimately, this might be a good thing. I'm 100% in favor of breaking Google's monopoly one way or another. Brave and DuckDuckGo and Kagi are all fine alternatives worthy of more prominence.
>The consumer version of Guided Learning is already available within Gemini, and can enable cheating. In a reporter’s test, it produced a detailed essay in response to an Advanced Placement history prompt.
I don't envy trying to stop cheating in an AI world, but when a student gets to the AP History level, if they want to cheat they'll know how to do it, with or without using a Google product with their school login.
Best you can do is guide them down the "correct" path, which itself is going to be harder and harder with AI generated content out in the wild.
To paraphrase Pink Floyd's Another Brick in the Wall, "How can you have any Google Search pudding if you don't eat your Google Gemini AI meat?"
https://www.youtube.com/watch?v=cdu0RQOKR0I
Also, as mentioned elsewhere in this discussion, Google's "Search and A.I. are deeply integrated" assertion is eerily reminiscent of the "Windows and I.E. are deeply integrated" assertion Microsoft made in United States vs. Microsoft.
https://en.wikipedia.org/wiki/United_States_v._Microsoft_Cor....
No. Not in the contemporary common meaning of the term “AI”.
5 years ago you would be technically correct, but today “AI” is only a subset of technologies that used to be included in the term. Namely Large Language Models, Text To Image Models, Text To Video Models, and other forms of generative AI that take in a user text prompt. These other types of statistical models are now simply called Machine Learning Algorithms.
Very few people today say “AI”, and mean “any type of Machine Learning Algorithm”.
I can't believe I'm saying this, but there may actually be a case for using "SI" now. I'm not saying machine learning somehow stopped being AI, but the term "AI" has become so synonymous with generative AI that people are apparently starting to retroactively redefine the entire field in terms of LLMs.
I prefer GenAI except the “Gen” stands for “general” because, really, the other kinds of ML/AI being contemplated in this thread are not at all general.
Like, before this you had extremely specialized models being painstakingly trained for an extremely specialized task using mountains of specialized data, painstakingly collected and labeled for correspondingly specific features, typically done by ML engineers (a specialized discipline) further specialized in that problem domain. And most users only have ever experienced all this abstractly through extremely narrow and specific functionalities like recommendations or ads.
On the other hand, any random person can hand these GenAI models any task that can be expressed as words or images or sound or some combination thereof and have a shockingly good chance of getting useful results. Yes they sometimes contain mistakes, but if you simply point them out, the models can often fix them!
I don’t know how much more general things can get than that.
It does get a bit uncomfortably close to “AGI”, which is a vague, heavily contentious term. But this sidesteps all that by not calling it AGI. The conventional “Generative AI” interpretation of GenAI thus becomes a convenient decoy for “General AI” ;-)
CI. Conversational interface. That's the difference. Simply a way to interact with a computer through the same mechanisms that you would use to interact with a person, through that computer's calculating a probable human reaction to a stimulus, after looking at millions of them and counting the likelihood of all possible reactions.
The real problem is pretending like there's any thought there at all. It's not programming because it knows how a computer works, it's programming because it can calculate what a person could likely write - which is why it writes source code rather than machine language. The value comes from the fact that it can chain a bunch of these reactions together, and react to things that don't originate from a human - if set to do so by a human, and they are things that a human could also predictably react to.
I'm not 100% sure you should be able to separate it from spam detection, though. It's learning from reading emails, and later whether that email is later marked as spam or saved from the spam pile. The critical part for me is whether that information is just going to be used to detect spam, or whether it's going to be used for something else.
If you're telling me you're not going to offer me spam detection, unless I not only agree to train your spam classifier, but some other model(s) too, just say that. It makes it clear that you're taking ownership of my data as payment.
As for children being able to sell their data as payment for the use of a search engine, the government pretends to be concerned about that - why doesn't it act here?
I am actually of the opinion that the problematic term is intelligence and that “super intelligence” won‘t actually help. This term has caused tons of issues in psychology and is now causing other issues in technology and journalism.
However I think it is a lost battle and I am willing to concede the term over to its current usage. I am however glad that traditional statistical models are no longer referred to by this trouble term. I think machine learning also has issues, and wish people would just be specific with their description by stating the particular algorithm they use (e.g. Markov Chains; ANN; Kalman Filter; etc.) but still accept “machine learning” this as a massive improvement over “artificial intelligence”.
I've heard anecdotal stories from MS-365 users that say it feels handicapped without Copilot. One govt worker said they noticed a marked change in Word spell-checking after they had to block all AI on their systems.
If this truly is just the search engine and not other stuff on the google.com domain like OAuth login, then this seems rather unimportant.
Search is easily Google's least-sticky and most replaceable product. The only reason it gets more usage that other engines like DuckDuckGo and the swathes of other Bing frontends is because it's the default and has the brand appeal for people who don't know better.
> Search is easily Google's least-sticky and most replaceable product. The only reason it gets more usage that other engines like DuckDuckGo and the swathes of other Bing frontends is because it's the default and has the brand appeal for people who don't know better.
This is a wild take, and ignores decades of evidence to the contrary.
Yeah, I read this and immediately thought “tying arrangement.” But they’re not CHARGING separately for search and AI, so I doubt a tying theory would hold up in court.
Their search engine is now integrated with Ai. So either google would need to block the search engine or the school if they don't want Ai. Take your pick.
A lot of these schools use tablets though. The kids probably use their phones so more likely to be google. Higher grades though use laptops I think so more likely Microsoft then. So far internet explorer is not integrated with Ai but microsoft copilot though is 100% ai and can be used for search.
I'm not talking about Copilot, I'm talking about IE6 back in the 90s, where Microsoft took the "oh woe is us it'll be so hard to undo the deep unnecessary integration we built for this very purpose" tactic.
That happened before some of the readers of this thread were born. A summary for those people, from the fog of distant memories: MS had an Internet Explorer that was not integrated in Windows, it was like any other browser. Then they decided to build it inside Windows. I remember that the Windows desktop got web views like widgets (web view as a name probably did not exist yet) and they called it Active Desktop. I think that it as COM/OLE objects too and the web view was probably one of them. File Explorer itself got an Internet Explorer like skin, probably with single click to navigate directories and open files, as if they were links. Then they claimed that it was impossible to disentangle IE from Windows. That eventually paved the way to the ruling against MS.
When I hear a giant company say that modifying some of its software is impossible I always think about that IE and laugh.
Not really? In the US it kept being bundled by default, they just made it possible to disable the front end; the MSHTML rendering engine was unremovable.
No, because the post-settlement era was actually the era of consolidated IE dominance. Chrome only started because MS, in their (unbroken by the FTC) dominance had sat around doing nothing for the web, which was bad for a web-based company like Google. Chrome also avoided using platform APIs because they wanted to be cross-platform.
Obviously they have at least a back-end search service that isn't tangled up with an LLM. When Gemini says "Searching the web..." it isn't chatbots all the way down. Google just needs to make a human UI for the same search Gemini uses.
Kagi might have a great opportunity with a cheap/free/freemium model for K-12 students.
Not only would the students become familiar with it and accustomed to the quality results, but the adults would see it and also appreciate the quality, including as an educational too.
They don't care, they think traditional search is already dead and are trying to jump ship to save the company. Pretty sure the kids these days all use "chat" anyway, not search engines.
Perennial reminder: when “Google” does things, what that means is that the two individuals who jointly have a controlling stake in Google, Page and Brin, choose for Google to do these things, or choose not to prevent Google from doing them. Obviously with an exception for things that they genuinely may not have heard about or might not have time for: but that leaves a lot of things. Object persistence is very real: don’t give Page and Brin a pass for things that Musk or Zuckerberg would be excoriated for simply because Page and Brin don’t jump up and down for attention on social media.
They seemed to have been comfortable in the role of the main monopolist, extracting inane ammount of money from people directly or indirectly, all without most people really noticing.
Then like a year ago they seems to have turned batshit evil, doing stuff that does not really make them any more money, is highly visible, makes many people angry & gives munition to their competition not to mention anti-monopoly and digital sovereignty advocates.
Given their recent stock hit due to Search and Ads taking a hit from the AI mode stuff, I'm a bit surprised they're doubling down more on making their core business inaccessible unless people opt into AI. Particularly since these are for children, where their ad data is probably (?) not as monetizeable, it doesn't make sense.
The only reasoning I can think of is that this is an intentional play to get students indoctrinated into Google's AI ecosystem, which is scummy but I guess also the entire reason the Chromebook education stuff exists
You can tell that AI features in user-facing software are good and popular because of the way that companies have to keep forcing them onto users even if they try to opt out.
This is great! Google search has been declining in usefulness so much, that it shouldn't be used by students anyway. And then there's the shitty ai summaries, which students definitely shouldn't have to worry about whether or not it's accurate.
I love old school research using books, however the advantages of connected information cannot be understated.
We could do much better with access to online libraries (which don't exist, apart from the very limited sample from these books accessible using Wikipedia).
Current (blogs and articles and forums) online resources provide information at a rate book search just can't provide.
Jesus, it has absolutely nothing to do with not having the technological capacity to go back and land on the moon, and everything to do with political and social will. We have rockets doing gymnastics in the sky and landing doing back-flips, you really think we've regressed because "we're not doing pen and paper"?
Very narrow minded approach, why not use the tools available? Experimentation is one of the best ways to learn, breaking stuff is a great way to learn. Doesn’t mean that people won’t fall into some traps along the way, but that’s kind of the case with anything.
This is along the lines of “for responsible parenting, never let your kids out of sight”.
Yes, thank you for extrapolating my comment into a massive over generalization. I guess I'll leave you to your books and pencils while I go actually explore the massive world around me to see what works and what doesn't.
Schoolkids aren't Drosophila, with ~10 day reproductive cycles and disposable at the end of the experiment. Experiments take years/decades, have loads of confounding variables, and you can wreck a generation or two if you go too far down the wrong route too quickly.
AI in schools may be useful! It may be horribly bad. It'll be a while before we know.
Yes, doing things by hand is another thing to prioritize.
A good deal of research shows that most learning happens when you recognize and back off from an incorrect solution on your own, without outside guidance. So, the key to a good learning environment is setting up scenarios where the backround information is present, and answers aren't out of reach, but there's a bit of struggle to find them.
This is also how you learn to learn, and develop grit. Grit is one of the best predictors of success.
Technology that gives answers or guides heavily is harmful to the learning process.
Totally agree! I think the risk within the context of the task is important, if you can work within an environment where you can make mistakes but not cause catastrophic consequences, its a good environment. I think its also great to get to a solution and realize it wasn't the right way to do something, and then iterate, there is not always a need to back off, although intuition over time will be great to notify someone that maybe they are on the wrong track.
Its also an approach backed up by research and studdies.
I still cannot believe there are people in education that can maintain that humans under 25 arent done neurologically developing ans have issues managing distractions and that tech in the classroom is a good idea (outside of classes deliberately teaching technology)
This comes off as comically evil. Disabling the AI summaries in search results is trivial, so them refusing to do so and blocking search entirely feels like a blatant attempt at coercion. Isn't it an abuse of monopoly power?
Some results also have thumbnails. They could easily be turned off. That doesn’t mean thumbnails are a different product, or an abuse of monopoly power.
It does if the thumbnails are harmful and you're as market dominant as Google with the incalculable levers of coersion like turning off your email or drive which they spend 20 years actively doing everything possible to get everyone to use.
A big part of this era is a shift to de-anonymizing all search, research, etc. That should be seen as its own problem separate from one's feelings about AI.
Google search enshittified itself. We have evidence that at least some leaders with power inside google made search perform worse to increase ad revenue (if users search more, they see more ads).
There wasn't a better alternative, until LLMs came along. It will be some time until the wild west of LLM-SEO hacking settles into some semblence of a stable ecosystem, but even with misinformation and hallucinations today chatgpt far outperforms search engines for virtually everything we want search engines to do.
So google is under existential threat for the first time ever.
And they appear to be playing "all search is AI now, so you can't have one without the other."
"Traditional" search, AI mode search and Gemini all present similar content rankings. The end result of a student searching "How did Abraham Lincoln win the 1860 election" will only vary by presentation, not narrative.
If Google were to "turn off AI" -- how far back into their ranking engine would they have to go? That would require them to run a 15+ year old engine, and the results would be worse than useless. Approaching Alta-vista levels of noisiness.
Don't you do that on your own personal computer and not.. on a school computer? Though, I'd advise against Google these days since it doesn't surface good results anymore. Even Bing is now better.
You can't count on kids having access to personal computers. And sometimes the computer is the class. I taught graphic design, and even though the kids all had licenses to use the Adobe suite at home, few actually could.
I don't understand why you would need to use it at school, in a classroom with a teacher present. It's not like Google does anything to student's personal computers at home or the one in the public library.
As an example, when I was at school, I would use the library in the building to do research for my papers. It would be odd for a major search engine to just not be available at my school, I feel. I would have to wait for the weekend to go home and be on a network that's not the schools to then use that search engine? Very inconvenient.
> If a school has banned A.I., he [Chris Phillips, vice president for education at Google] added, students “will not have access to search, and they’ll see a message from their school that indicates it’s not available.”
So scummy! Make it sound like the school's decision, rather than Google's.
That's what every Big Tech company has done when threatened with loss of access to their children: weaponize the children themselves with poor gaslighting to fight the adults so Big Tech gets its way.
Google says it's the big mean school's fault, so the school kids are all gonna band up and try to protest and argue and disrupt because they can't get Google Search.
When Epic didn't get its way... turn the kids against the adults by threatening to take their Fortnite away.
Get people used to search without ads or influence or data collection or any incentive other than getting you what you asked for as honestly as possible.
Or maybe more pointedly, get them to even become aware of the difference between home and school, or everyone else's computers and their own.
The economic elite really want to dumb children down at the same time of juicing their AI numbers by bundling LLMs and searches even when they don't need to be.
> Very interesting approach. Basically, you can no longer use Google Search as a way to discover and learn about a subject the old-school way.
If by "interesting" you mean "abusive and coercive."
> I guess the point here is that Google thinks schools need to start using AI, and that all this "no AI in schools" talk is nonsense.
Explain that more please: whose talk is "nonsense"? Google's? Parents'? Schools'?
'Cause it sounds Google is saying this:
> You silly parents and educators, you thought you set educational policy for your students? No, we actually do. Your choices will conform to our business plans of the moment, or we will make your life so painful and difficult that you will "consent" to our demands.
> Did we mention our motto is "don't be evil"? We live our values every day.
> I guess the point here is that Google thinks schools need to start using AI, and that all this "no AI in schools" talk is nonsense.
More generously, it seems like Google just doesn't see a justification for maintaining an AI-free search platform alongside their AI-integrated vision for search.
For a smaller vendor in a competitive market, this decision wouldn't read like coercion. It would just be a roadmap statement that might draw some customers in and drive some towards competitors. You would expect a business to be able to make that kind of decision (a prima facie reasonable one) about how they'll invest in their products and the support infrastructure that surrounds them.
What this story really exposes is how dependent schools have been made on Google specifically and how the idea of independently making choices against their roadmap feels infeasible. It feels like a monopoly to administrators and the public even if it isn't formally marked as such by regulatory agencies.
that is an absolutely repugnant take from anyone at Google (I will not attribute this to a legal entity that cannot think, a human being made this call and is accountable for it). They are not educators and should have absolutely zero say in how schools do things.
They don't; they do however have say on how they develop their own products, and they want to stuff search full of AI. At least they let schools opt out of it easily.
"You don't have to use AI, but if not you can't stay here".
Google thinks that will be more profitable if they force schools to use AI. They do not think "no AI in schools" talk is nonsense, they think "no AI in schools" is bad for their bottom line, somehow.
Google thinks forcing schools to use AI to fuel search will give them great cash. You have far too kind a view on what Google hopes AI will do to schools and not nearly a cynical economic one.
Google really is an evil, enshittified company now. I see it every day when I'm on YouTube with an ad-blocker. "Experiencing interruptions?" Yeah, Google, thanks for fucking with me. If you don't want me viewing your videos, don't make them accessible on the standard protocols.
There are a host of other search engines. I use search.brave.com myself (after leta.mullvad.com was decommissioned).
Students will learn about other search engines, and about how corporations are not your friends. Both good lessons to learn.
I really don't like that my original submission "Google Will Turn Off Search for Schools That Reject A.I." from the New York Times got replaced with "Designing responsible AI for learning" PR slop from Google Corporate that doesn't even disclose this limitation.
If the mods are gonna Orwell this, at least take my name off it.
Please submit the original source. If a post reports on something found on another site, submit the latter.
The exception to this this where a source contains significant additional information that is not in the original source, which, as commenters have pointed out, seems to be true in this case.
I would also note the same guideline says to use the original title, which was (and remains, as submitted) "Google Will Turn Off Search for Schools That Reject A.I."
They are constantly tipping the scales here, and if you say so everyone will show up and claim they're the best mod team on the planet and totally fair and it's outrageous that you want to see accountability for mod actions via some sort of mod log and the mods will say that a mod log will allow some sort of bad actors to do something vague and threatening and they'll also use rules like a cudgel and claim a technical reading of some rule will absolve them. Then again, we are on a site run by a VC so what would you expect, of course they're going to tip the scales.
Please don't post ranting accusations like this. We're accountable to this community every single day, as this case demonstrates. Any time we do anything that seems even a little bit suspect, we hear about it very quickly.
> a mod log will allow some sort of bad actors to do something vague and threatening
I don't know of any time we've said anything like that. Dang has explained plenty of times why we don't publish a moderation log [1]. There are plenty of ways for people to see what we're doing and hold us to account if we get things wrong.
> we are on a site run by a VC so what would you expect, of course they're going to tip the scales
HN is only any use to our bosses if we continue to be the place where the best (i.e., "smartest") hackers come to have the highest-quality discussions about the most important topics, and the best/smartest hackers have the best bullshit detectors, almost by definition. But even if we thought we should deprioritize that consideration and instead "tip the scales" in aid of the narrow commercial interests of our parent organization (leaving aside the flawed presumption that it would work), YC has zero interest in protecting the reputation or commercial interests of the megacorps, which it demonstrates through its actions continually [2,3].
How is this at all related to the mafia and anywhere near blackmailling schools?
It would only be blackmail if the service they are gating access to is of significant enough benefit that it causes enough harm to the schools to encourage them to also allow the gemini access, but that's counter to your claim that "nobody needs that company anyway"
Google is a private entity that is allowed to decide how anyone uses their products.
The schools, being entities themselves, are in turn allowed to handle this however they want. Perhaps rolling their own search to encourage learners to rely on better sources. Perhaps blocking AI generated responses. Perhaps blocking search results that advertise products to children.
Or they can just use any of a dozen other search providers.
is there a way to use it that prevents "tell me the answer", if not that is what enough students will gravitate towards that it will be harmful to society at large. If there is an agent that only works as an infinitely patient tutor, that will under no circumstance do the work for you, then sure. Just have expereience educators dictate the way the agent works and is used, not some marketing dweeb or MBAsshole.
> The AI crew's fantasy is that students sitting in front a screen will be motivated because A) the adaptive technology will hit them with exactly the right material for the student and B) shiny! Arnett explains that any dreams of AI-aided motivation are doomed to failure.
> Arnett's explanation is not exactly where I expected we were headed. Human respect is scarce, he argues, because humans only have so much time and attention to parcel out, and so it's valuable. AI has infinite attention resources, can be programed to be always there and always supportive. Arnett argues that makes its feedback worthless in terms of status and respect. ...
> But I think much of what he says here is valuable and valid-- that AI faces a major hurdle in classrooms because it offers no social relationship component, little opportunity to provide students with status or respect. Will folks come up with ways to use AI tools that have those dimensions? No doubt. But the heart of Arnett's argument is an explanation of one more reason that sitting a student in front of an AI-run screen is not a viable future for education.
That's certainly one outcome. Another is having AI tell you the answer and you understanding why your answer is incorrect. This is not black and white.
If the Agent is designed for that type of interaction and not "do it for me so i don't have to know the details" then yes, but currently that is not supported so teenagers and children will absolutely use it as a "don't make me think" button
You still have to double check the Agent's grading as they still make frequent minor mistakes, that while they don't really impact coding as much or can be caught with tests, that is not the case here.
The actual question to ask is if the agent makes more mistakes than a human being. Grading homework involves reading over similar/identical answers to the same questions over and over, and it is easy for human beings to make mistakes while doing that.
(That is basically the argument for OCR in the postal system -- it does make mistakes but it makes fewer mistakes than human beings.)
Some of the worst performance schools are cyber charter schools, which often promote the idea that they use AI as a personal tutor.
If they could get better results with off-the-shelf tools, I'm pretty sure they would.
(Not 100% sure. Some non-profit charters pay a for-profit management company, which suggests commercial software, where the charter, management company, and software provider are all owned by the same person. They may prefer to make more money than improve student grades.)
Would you agree that there is a difference between "search results ranked by AI" and "Magic box that blurts out an answer"?
We do want to teach kids how to do searches, to look at and parse results and basically do "online research", right. Not just blindly trust the magic answer box?
The "magic box" is really just summarizing the top X results, which have been ranked by AI for many years. Everyone has been "blindly trusting" all that for a good while now.
I remember being taught library skills as a kid. That's what adults back in the day wanted to teach, because we needed to be able to find books to help with assignments and whatnot. And then the Internet and search engines happened and that class is no more. Things and times change.
And this is why Google needs to be split in parts. Ads on one side, or many sides as they have purchased so many ads business that they should be further divided, youtube on another side, search in a different one, etc.
Split monopolies so they stop threatening schools, governments and citizens.
Which I get it, they indeed can use LLMs to improve the previous traditional ML pipeline. However I believe bundling these options together shouldn't be legal and I never expected Google to employ dark patterns like this. When I subscribed, these features were provided by traditional ML, perhaps some BERT-like encoder architecture at most, however now I cannot keep that previous functionality if I don't wish to grant LLMs access to my data.
I used to be able to search for parts of names. Or a word that is part of a word in a message. Doesn't work anymore. Only exact matches now, or you don't get anything.
Damn, truly we are the products.
This is google-for-business or whatever they call it now.
for those who didn't know how to do it:
1. https://mail.google.com/mail/u/0/#settings/general
2. under "Smart features:", toggle off "Turn on smart features in Gmail, Chat, and Meet"
But it's not only that. Model size, for instance, also plays a role. Will modern BERT keep their lossy compression nature? Or could there be a privacy risk as well?
Thus, what are the characteristics of privacy-aware models? So far: discriminative vs. generative, model size, and lossy transforms (related to size). Perhaps looking at pre-2020 models could help to narrow other necessary attributes.
My guess is that the pairing of all of these things is due to the fact that Google has likely consolidated everything under the same style of decoder model.
And frankly I see the distinction in only number of model parameters vs number of people in cohort it was trained. So 50k-param model trained on 1b emails from 1 million people is ok, encoding/compression loses anything personal. But huge LLMs are closer to lossless compressors that can just spew all my passwords, prescriptions etc verbatim.
I found all of the new LLM features to be utterly pointless and incredibly annoying, but the last straw was when my account was abruptly locked recently. For a week I couldn't access mail, maps, photos, digital picture frames for my extended family all went dark, my personal blog ... all shut down with no reason given. No customer support available to contact to resolve google's mistake. A week later, it was suddenly unlocked again. Still no reason given.
Fuck google. You are not their customer; you are their product. Google doesn't care about you any more than Tyson cares about the chickens it sells.
I hope that enshitification calms down a bit whenever the dust has settled on frontier AI companies and we no longer have to force this into every product imaginable. Google used to be pretty good relative to the rest of the big ad tech companies (looking at Meta/Facebook with a scowl...) at giving opt-outs and settings to personalize privacy stuff. This has been a disappointing change to see
I see people try so hard to make this distinction when it really doesn't exist. Transformers are only a little different than long short-term-memory architecture (parallelized, with attention mechanism and no recurrance).
What is the point at which is departed from tradition machine learning? Seems more like people retrofitting a moral decision to not like LLMs and trying to dress it up as a technical distinction.
Does anyone even understand the world currently? Its so absurd to me, why does it exist to begin with and why is this drama appearing?
Have you noticed how many people already use AI to do their thinking for them. They just ask it what to think, what to do, for everything. They want that for as many people as possible. They don’t want people going outside of Google to other websites to decide what to think. Yes they will put ads in it eventually. But in the end they will also use it to influence what the population thinks.
We already have some leaders and politicians outsourcing their thinking in this way. It will make these companies incredibly powerful.
The former is a subset of the latter.
as far as ads go, the razor cuts to the bone when you realize that microtargeted ad campaigns, are overly complex, and have folly. what you really want is to predict someones shopping list, rather than recommend items that were already purchased in gross quantity.
https://beomniscient.com/blog/the-9-best-llm-seo-agencies-fo...
Just wait for “$1000/mo for a 10% chance your product is mentioned in the AI summary for an auto parts search”.
I have some guesses:
They can artificially inflate the number of users who use their AI by including everyone who uses their search engine.
It trains people to be increasingly dependent on whatever google tells them instead of checking and reading multiple sources, and verifying information for themselves.
They hope to let people pay them to manipulate what their AI returns for certain searches (either mentioning certain things/products or suppressing mention of them).
Google search is getting increasingly useless as the unending flood of ad-filled AI slop websites dominates the search results and people will abandon searching with Google if they can't find some way to provide people with the information they are looking for. (it's not going well so far https://arstechnica.com/google/2026/04/analysis-finds-google...)
Yup I could see that being plausible. It’s like Microsoft and Copilot. Project managers got to boast how they got AI “adoption” numbers up just by shoving it into people’s faces.
I do sort of wonder how much longer Googles search API will even be a thing. They clearly have no interest in running a search engine anymore and are becoming actively hostile to the open web.
Free, terrible ad-riddled and essentially broken
Paid, you are the customer not the product.
Google's search is increasingly AI driven (as it should be, i do not want to click on reddit links and waste my time skimming through the 1000 comment threads), asking Google to maintain a non-AI search engine is like asking Tesla to come up with an ICE engine car right now. It is not happening.
Google Search I would expect will increasingly become more like Gemini and less like traditional search. Schools could certainly go to DuckDuckGo or other search engines that are free. (Google has shown that it does not care if they do as it sees it as a dying business.)
If nobody is visiting anything on the internet, because every search engine response is an AI output, then where will any information for this AI output come from?
This paradox was clear from day one, pretty sure I posted on here about it years ago. AI search will kill the dataset that AI itself relies upon.
That's exactly why they MUST have everyone opted in, even if they don't like it.
...is like asking your company to maintain its SQL database instead of shoving everything into a document the LLM reads and suggests edits to.
Even if you're using an LLM to talk to the user, there still needs to be some kind of deterministic search engine back there in order to do things well, or to answer any kind of question that involves information newer than X months ago.
If someone searches for "mexican food" you require ranking anyway because you aren't going to load those ten bazillion hits into the LLM first.
What's missing here is your imagined alternative: What exactly are trying to propose where Google can ditch having a search engine, and why do you believe it work well instead of badly?
Most will end up in deep regrets, ie how jobs regretted being an asocial dick and all above to his own family before dying (thats the better outcome, many simply see them as stars of their show till last moments, like potus for example). Its not some secret cabal, these people if smart naturally bubble to top of social and financial hierarchies.
It used to be much less visible, some sort of on-surface decency momentum and shit behind the curtains, now everything is out in the open for whole civilization to see instantly and its pretty tiring and nauseating.
If you don't get it its fine, it means you are a normal human being, probably well balanced. Folks like that abhor such people personally, be it c-suites of corporations or politicians or just a pos colleague.
Google is willing to sacrifice children to push engagement. In family link they refuse to let me restrict the Google app. In what world do I want my 1st or 2nd grader opening the Google app and reading the news about murder, rape, and all the other horrendous things that are common in the Google apps news summary? Give us parents actual control in androids parental control app!
It has been over a decade since Google Search was anything but a machine-learning powered word association engine. I miss old Google Search but it's dead and I trust Gemini more than what Search has turned into.
Most of the time it's quicker, sometimes it's just bad.
Imagine winning search by being the fastest and most accurate, then replacing it w/ constant hallucinated errors.
Given my sample is biased as I am using it for an outlier of difficult web searches (including web searches localized to Iceland). But I would think that if your parent was correct, at least majority of my uses would be correct, but in my experience, Gemini fails most of the the time traditional web search fails. So to me at least, it is effectively useless.
For people who use it in place of normal web search, it is probably worse then useless, because of how confidently wrong it is.
'I think they upgraded the model...."
The argument is that there's so much garbage on the open web that clicking through to the source has a lower signal-to-noise ratio than the AI summary.
This rings very true in my experience.
Google's AI summaries definitely make mistakes, but in most cases they're higher signal than the source websites.
I really think Google is making an error using such a lightweight model for the page. That's one of my only firsthand exposures to genAI, and it's awful. Use a decent model, or if that's too expensive don't put up a summary at all.
Now it's just speed. This is a grave error.
I don't quite understand what this even means. I both use AI chatbots frequently and still use keyword-based search engines frequently as well. When I use keyword-based search engines, it's usually because I want to see what sort of websites, discussions, resources, online hangouts, etc. exist around a given term, or just how the term is used. I often don't have a specific question or task in mind and just kind of want to wallow in the search results for a while, sort of like if I moved to a new city and just wanted to walk around different neighborhoods. otoh when I use an AI chatbot, it's usually because I have a specific question to ask, and often I don't know what the relevant search terms would even be. The use cases seem fundamentally different and I don't know how one could be "more reliable" than the other.
I find it easier to recognize bullshit from its source vs an LLM's confident and well written regurgitation of it.
(Obviously far from perfect, often far from even good... just better.)
Which is not a good look for the Search teams/execs.
> Your public IP address as seen by external servers is 66.249.68.3 (or check your live details via ifconfig.me or What Is My IP Address).
The AI just hallucinates¹ an IP. (I.e., that is most definitely not my IP address.)
(¹it could possibly be invoking a tool here, I suppose. Either way, the answer is slop.)
So I can’t understand how one would trust the summary but not what is being summarized.
mistakes vary from "obviously wrong" to "looks vaguely correct, but the source states something different" to "looks plausible, but the cited page is completely irrelevant, the AI clearly made this up" Even for straight fact lookups, I've seen it cite pages and claim one number, and then i click through and the source matches the vague shape of the answer but has a different number.
Its just constant hallucinations. "thats a common problem, just change this <setting that does not exist> or set <option that does not exist>". Full gemini seems more capable, but whatever google is throwing at the top of their free searches is awful IME.
The relative effectiveness of the search is immaterial.
This is gross.
This is all very tragic. we built a magic network where we could stash all the worlds information and access it from anywhere, except its been trashed by people with sleazy motivations trying to get between us and what we'd like to see.
My mother never used search. She found it too hard to manage. She uses Gemini all the time, it gives answers in her language and she does not need to type. She is incredibly happy.
My daughter talks to Gemini. The idea of going through links seems primitive to her.
The future of interface is chat/talking not going through clickbait.
I think schools should have some basic control on how kids can use AI but saying that they want to freeze the learning experience in 2020 is simply not right. It is like saying we wont allow whiteboards and markers.
People just really, really, really hate paying for stuff. Hell Google even lets you continue to use their services even if you block their ads. They're that generous, err, evil.
That's how I read it nowadays, as Google search results are in a quality going down trend.
Actual search results are still bad. Seems to be a trend across most engines
the direction they want to go is clear for me, similar to proton
https://lite.duckduckgo.com/
These are still a thing for people who don't like JavaScript, so I'd give them some credit.
And it's not like these don't cost anything to run and maintain. Quite the contrary: bots love to use them, and it screws with their ads-driven business model. And yet they are still keeping these around.
I think that "noai", too, will keep going for a long time.
very google-y to me, or valley-y is probably more accurate, par for the course across so many products these days (in your face)
DDG pulls stunts too
I'm moving to API based search options, using multiple, and rendering them as markdown for myself or my agent
I noticed the same with Qwant. They even try to copy the useless UI for Google. And on top of that I get a 2 seconds delay "verifying device".
It seems nobody knows how to create a proper search engine anymore. Google search is also useless compared to "the good old days".
I'm using API search like EXA, the incentives feel more aligned
https://blog.google/products-and-platforms/products/educatio...
> We believe schools should have control over AI in their classrooms and in providing administrators with controls to help them manage their approach. After listening to schools, teachers, leading researchers, and policymakers, we are now updating our approach to controls for Gemini, Gemini Notebook, Gemini in Classroom, and Google Search. Access to these products will be entirely off for K-12 student accounts until school administrators explicitly turn them on.
And NYT says they would only be able to turn search on if they enable all of the Gemini stuff. That is not mentioned explicitly on google's blog. So not _that_ clickbaity?
So yeah the blog post says Google Search itself will be entirely off until being turned on. There’s nothing in this blog post to indicate whether Google Search can be turned on without also turning on Gemini. Although I hate the NYT, the NYT is quoting someone from Google saying there won’t be separate controls for Gemini and Google Search.
From a privacy perspective, having all of this off by default is sensible. But only offering a basic product like Search if you opt into AI (when search as a capability does not need AI to function) - is newsworthy.
That means a far less capable search engine. Note that Google search has been AI-powered since before AI became a thing; it's now just more advanced, and particularly, explicitly publicized.
you either get search + all the ai stuff, or you get none of it. you can't just opt-in to search alone.
(in other words, title is accurate. block ai = no search)
C'mon, it's right in the article.
The school can control AI on/off. Off means no search too.
This particular bit of juice didn't have the "AI summaries" ingredient for 93% of its existence.
Schools invested enormously in Google Classroom and Google for Education, and now the deal's changing. Not for the benefit of students.
The local population is probably allowed to complain that you built the reservoir with toxic waste, too, or that you filled the old well with construction debris.
But anyway I personally would find someone offering well water if I don't like the reservoir. And I've been using DuckDuckGo exclusively for search since they first appeared, so I've never had to deal with what Google's doing. I know the reservoir exists, but I've never drank from it.
That sounds like an argument to never depend on Google for anything ever again under any circumstance short of severe punitive legislative action.
Search is off; if you want it back on, you get Gemini summaries with it. you can't get Search _without_ Gemini
Scummy wording makes it looks like google is giving more control, but it's really the opposite.
https://kagi.com/libraries
It’s a good time to break free from Google and try something else. I’ve been using kagi for a couple years and I haven’t looked back.
Not a shill.. I just like the service.
Students don’t really carre if their school computer use DuckDuckGo or Google.
And if you can even open student eyes on Google alternative, that’s a great thing.
They are shooting themselves in the foot, and I’m glad they are doing it.
Google Search uses lots of AI. Google is now allowing schools to turn off all AI entirely. That, obviously turns off Google Search. This is the feature working as intended.
It seems like school administrators will get to control which products get AI, individually. Which is what they've been asking for:
> After listening to schools, teachers, leading researchers, and policymakers, we are now updating our approach to controls for Gemini, Gemini Notebook, Gemini in Classroom, and Google Search. Access to these products will be entirely off for K-12 student accounts until school administrators explicitly turn them on.
Am I missing something here...? This announcement seems to be 100% good, but people are writing like it's 100% bad.
As an adult, wanting to understand concepts for my own benefit, I've found it really useful as a learning tool. As a teenager, I probably would have been irresponsible with it and sabotaged myself.
This is a blatant strawman.
AI lies and gaslighting are driving me up the wall. (Good thing I don't have hypertension.)
I don't envy trying to stop cheating in an AI world, but when a student gets to the AP History level, if they want to cheat they'll know how to do it, with or without using a Google product with their school login.
Best you can do is guide them down the "correct" path, which itself is going to be harder and harder with AI generated content out in the wild.
Also, as mentioned elsewhere in this discussion, Google's "Search and A.I. are deeply integrated" assertion is eerily reminiscent of the "Windows and I.E. are deeply integrated" assertion Microsoft made in United States vs. Microsoft. https://en.wikipedia.org/wiki/United_States_v._Microsoft_Cor....
5 years ago you would be technically correct, but today “AI” is only a subset of technologies that used to be included in the term. Namely Large Language Models, Text To Image Models, Text To Video Models, and other forms of generative AI that take in a user text prompt. These other types of statistical models are now simply called Machine Learning Algorithms.
Very few people today say “AI”, and mean “any type of Machine Learning Algorithm”.
Like, before this you had extremely specialized models being painstakingly trained for an extremely specialized task using mountains of specialized data, painstakingly collected and labeled for correspondingly specific features, typically done by ML engineers (a specialized discipline) further specialized in that problem domain. And most users only have ever experienced all this abstractly through extremely narrow and specific functionalities like recommendations or ads.
On the other hand, any random person can hand these GenAI models any task that can be expressed as words or images or sound or some combination thereof and have a shockingly good chance of getting useful results. Yes they sometimes contain mistakes, but if you simply point them out, the models can often fix them!
I don’t know how much more general things can get than that.
It does get a bit uncomfortably close to “AGI”, which is a vague, heavily contentious term. But this sidesteps all that by not calling it AGI. The conventional “Generative AI” interpretation of GenAI thus becomes a convenient decoy for “General AI” ;-)
The real problem is pretending like there's any thought there at all. It's not programming because it knows how a computer works, it's programming because it can calculate what a person could likely write - which is why it writes source code rather than machine language. The value comes from the fact that it can chain a bunch of these reactions together, and react to things that don't originate from a human - if set to do so by a human, and they are things that a human could also predictably react to.
I'm not 100% sure you should be able to separate it from spam detection, though. It's learning from reading emails, and later whether that email is later marked as spam or saved from the spam pile. The critical part for me is whether that information is just going to be used to detect spam, or whether it's going to be used for something else.
If you're telling me you're not going to offer me spam detection, unless I not only agree to train your spam classifier, but some other model(s) too, just say that. It makes it clear that you're taking ownership of my data as payment.
As for children being able to sell their data as payment for the use of a search engine, the government pretends to be concerned about that - why doesn't it act here?
However I think it is a lost battle and I am willing to concede the term over to its current usage. I am however glad that traditional statistical models are no longer referred to by this trouble term. I think machine learning also has issues, and wish people would just be specific with their description by stating the particular algorithm they use (e.g. Markov Chains; ANN; Kalman Filter; etc.) but still accept “machine learning” this as a massive improvement over “artificial intelligence”.
Everyone here is also aware of what specific form of AI is being discussed here, including, I suspect, you.
Search is easily Google's least-sticky and most replaceable product. The only reason it gets more usage that other engines like DuckDuckGo and the swathes of other Bing frontends is because it's the default and has the brand appeal for people who don't know better.
This is a wild take, and ignores decades of evidence to the contrary.
https://en.wikipedia.org/wiki/United_States_v._Microsoft_Cor....
When I hear a giant company say that modifying some of its software is impossible I always think about that IE and laugh.
Profited very handsomely from his lies.
Do you not think the settlement requiring APIs be opened up helped enable Chrome at all?
I added a custom search engine entry to firefox that does this, made it the default, and haven't seen an AI summary from Google since I did that.
Hiding one div is 90% of the solution.
Not only would the students become familiar with it and accustomed to the quality results, but the adults would see it and also appreciate the quality, including as an educational too.
All those kids will get used to whatever other search they can use.
Probably time to switch him to Kagi. Google Search is pretty bad anyway.
They seemed to have been comfortable in the role of the main monopolist, extracting inane ammount of money from people directly or indirectly, all without most people really noticing.
Then like a year ago they seems to have turned batshit evil, doing stuff that does not really make them any more money, is highly visible, makes many people angry & gives munition to their competition not to mention anti-monopoly and digital sovereignty advocates.
The only reasoning I can think of is that this is an intentional play to get students indoctrinated into Google's AI ecosystem, which is scummy but I guess also the entire reason the Chromebook education stuff exists
We could do much better with access to online libraries (which don't exist, apart from the very limited sample from these books accessible using Wikipedia).
Current (blogs and articles and forums) online resources provide information at a rate book search just can't provide.
I optimize for finding the answer to a given question, without knowing beforehand if it will be found online or in an research article or in a book.
And I mean finding the right answer, before you mention "learning" like it is something that can't be done using computers.
There are many, but all short of the potential for knowledge and learning:
Internet Archive Lending Library, Hathi Trust, Libby / Overdrive, U. of Pennsylvania's Online Books Page:
https://onlinebooks.library.upenn.edu/
There are also the shadow libraries, of course.
Trillions and trillions of dollars of technology later, and the best we can do now is fly by and wave.
This is along the lines of “for responsible parenting, never let your kids out of sight”.
Breaking stuff sometimes is a way to learn "don't break that".
Broken children can be tough to fix.
The answer is "sometimes the new fancy tools are bad, and it can take decades to confirm that".
AI in schools may be useful! It may be horribly bad. It'll be a while before we know.
A good deal of research shows that most learning happens when you recognize and back off from an incorrect solution on your own, without outside guidance. So, the key to a good learning environment is setting up scenarios where the backround information is present, and answers aren't out of reach, but there's a bit of struggle to find them.
This is also how you learn to learn, and develop grit. Grit is one of the best predictors of success.
Technology that gives answers or guides heavily is harmful to the learning process.
Centrally forcing schools to do what is good for corporation is not experimentation.
There wasn't a better alternative, until LLMs came along. It will be some time until the wild west of LLM-SEO hacking settles into some semblence of a stable ecosystem, but even with misinformation and hallucinations today chatgpt far outperforms search engines for virtually everything we want search engines to do.
So google is under existential threat for the first time ever.
And they appear to be playing "all search is AI now, so you can't have one without the other."
If Google were to "turn off AI" -- how far back into their ranking engine would they have to go? That would require them to run a 15+ year old engine, and the results would be worse than useless. Approaching Alta-vista levels of noisiness.
https://en.wikipedia.org/wiki/Google_for_Education
Maybe even like some kind of aggregator that allows you to combine several other search engines' results.
That'd be really cool if that exi...oh, wait...
It's true, I used AltaVista and Yahoo! search because Google didn't exist yet.
So scummy! Make it sound like the school's decision, rather than Google's.
Google says it's the big mean school's fault, so the school kids are all gonna band up and try to protest and argue and disrupt because they can't get Google Search.
When Epic didn't get its way... turn the kids against the adults by threatening to take their Fortnite away.
I know they removed "don't be evil" a while ago now, but they didn't have to immediately become the exact opposite
Get people used to search without ads or influence or data collection or any incentive other than getting you what you asked for as honestly as possible.
Or maybe more pointedly, get them to even become aware of the difference between home and school, or everyone else's computers and their own.
Microsoft tried this with IE way back when. Didn't go well for them in court. I think it's time Google faced the same music.
I guess the point here is that Google thinks schools need to start using AI, and that all this "no AI in schools" talk is nonsense.
If by "interesting" you mean "abusive and coercive."
> I guess the point here is that Google thinks schools need to start using AI, and that all this "no AI in schools" talk is nonsense.
Explain that more please: whose talk is "nonsense"? Google's? Parents'? Schools'?
'Cause it sounds Google is saying this:
> You silly parents and educators, you thought you set educational policy for your students? No, we actually do. Your choices will conform to our business plans of the moment, or we will make your life so painful and difficult that you will "consent" to our demands.
> Did we mention our motto is "don't be evil"? We live our values every day.
More generously, it seems like Google just doesn't see a justification for maintaining an AI-free search platform alongside their AI-integrated vision for search.
For a smaller vendor in a competitive market, this decision wouldn't read like coercion. It would just be a roadmap statement that might draw some customers in and drive some towards competitors. You would expect a business to be able to make that kind of decision (a prima facie reasonable one) about how they'll invest in their products and the support infrastructure that surrounds them.
What this story really exposes is how dependent schools have been made on Google specifically and how the idea of independently making choices against their roadmap feels infeasible. It feels like a monopoly to administrators and the public even if it isn't formally marked as such by regulatory agencies.
That itself is worth more people thinking about.
"You don't have to use AI, but if not you can't stay here".
This is attempted coercion, plain and simple.
There are a host of other search engines. I use search.brave.com myself (after leta.mullvad.com was decommissioned).
Students will learn about other search engines, and about how corporations are not your friends. Both good lessons to learn.
edit: Looks like it was undone. Thank you, whoever that was.
If the mods are gonna Orwell this, at least take my name off it.
edit: Thanks, mods, for putting it back.
Please submit the original source. If a post reports on something found on another site, submit the latter.
The exception to this this where a source contains significant additional information that is not in the original source, which, as commenters have pointed out, seems to be true in this case.
This will virtually always be the case of a reputable news article about a PR announcement written by a corporate marketing department of this nature.
Example, currently #2 on the home page: https://news.ycombinator.com/item?id=50336918 would need to be changed to https://www.fujifilm.com/me/en/news/hq/14137.
I would also note the same guideline says to use the original title, which was (and remains, as submitted) "Google Will Turn Off Search for Schools That Reject A.I."
> a mod log will allow some sort of bad actors to do something vague and threatening
I don't know of any time we've said anything like that. Dang has explained plenty of times why we don't publish a moderation log [1]. There are plenty of ways for people to see what we're doing and hold us to account if we get things wrong.
> we are on a site run by a VC so what would you expect, of course they're going to tip the scales
HN is only any use to our bosses if we continue to be the place where the best (i.e., "smartest") hackers come to have the highest-quality discussions about the most important topics, and the best/smartest hackers have the best bullshit detectors, almost by definition. But even if we thought we should deprioritize that consideration and instead "tip the scales" in aid of the narrow commercial interests of our parent organization (leaving aside the flawed presumption that it would work), YC has zero interest in protecting the reputation or commercial interests of the megacorps, which it demonstrates through its actions continually [2,3].
[1] https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...
[2] https://techcrunch.com/2025/05/13/y-combinator-says-google-i...
[3] https://techcrunch.com/2025/08/22/y-combinator-says-apples-a...
It would only be blackmail if the service they are gating access to is of significant enough benefit that it causes enough harm to the schools to encourage them to also allow the gemini access, but that's counter to your claim that "nobody needs that company anyway"
Google is a private entity that is allowed to decide how anyone uses their products.
The schools, being entities themselves, are in turn allowed to handle this however they want. Perhaps rolling their own search to encourage learners to rely on better sources. Perhaps blocking AI generated responses. Perhaps blocking search results that advertise products to children.
Or they can just use any of a dozen other search providers.
It's almost always possible to do worse.
> The AI crew's fantasy is that students sitting in front a screen will be motivated because A) the adaptive technology will hit them with exactly the right material for the student and B) shiny! Arnett explains that any dreams of AI-aided motivation are doomed to failure.
> Arnett's explanation is not exactly where I expected we were headed. Human respect is scarce, he argues, because humans only have so much time and attention to parcel out, and so it's valuable. AI has infinite attention resources, can be programed to be always there and always supportive. Arnett argues that makes its feedback worthless in terms of status and respect. ...
> But I think much of what he says here is valuable and valid-- that AI faces a major hurdle in classrooms because it offers no social relationship component, little opportunity to provide students with status or respect. Will folks come up with ways to use AI tools that have those dimensions? No doubt. But the heart of Arnett's argument is an explanation of one more reason that sitting a student in front of an AI-run screen is not a viable future for education.
(That is basically the argument for OCR in the postal system -- it does make mistakes but it makes fewer mistakes than human beings.)
If they could get better results with off-the-shelf tools, I'm pretty sure they would.
(Not 100% sure. Some non-profit charters pay a for-profit management company, which suggests commercial software, where the charter, management company, and software provider are all owned by the same person. They may prefer to make more money than improve student grades.)
It is a tool, it can be used within boundaries, but it has zero place in teaching.
We do want to teach kids how to do searches, to look at and parse results and basically do "online research", right. Not just blindly trust the magic answer box?
This is like a book publisher refusing to sell books to schools without including spark notes on the inside cover.
Split monopolies so they stop threatening schools, governments and citizens.