I keep hearing AI companies say the demand is outstripping their supply and they need to build more data centers right now. And then I also see and read about how every company is cramming AI down everyone's throats in every product whether or not it's useful and/or wanted. I've used enough AI to see that it's revolutionary, but I'm also old enough to see the parallel to the dotcom days.
The demand seen by the big AI companies is largely from those others putting the AI features into everything. In this particular gold rush we actually have two sets of shovel sellers: Nvidia and the big AI companies.
And those shovels are being sold wholesale to software firms who recognize the only choices are adapt or be replaced...
...and, in my experience, those software firms are having a hell of a time selling shovels to bumpkins that seem skeptical of this gold rush (mid-sized organizations suffering from inflation and other economic headwinds)
What is the lesson of the dotcom days? There will be a brief pullback - and a couple goofy businesses will live in infamy - before that tech actually does go to the moon?
I think there are two things that can be true at once:
1) For about 1% of people AI is completely transformative to their jobs and let's them work much much faster. Many engineers could easily blow $5,000 in compute per month on AI, and that number is only going up as AI gets better. A year ago, it would have been insane to spend that much on AI, now it's normal. This is driving all the demand, and will continue to be useful.
2) For 99% of people they have no use for AI short of better search. All of the AI tools being crammed everywhere are businesses trying to figure out how to not "get left behind" and ride the hype train. There is no demand for this stuff and I see the parallels to the dotcom bubble.
The AI being packed into everything are lightweight models that are not compute intensive at all to run.
It's the behemoth models (and their agents) that are actually constrained. There is not enough compute to meet the demand for people offloading their day job sized models (and the agent teams they spawn).
The app I use to remote start my car just got a fancy new upgrade from the manufacturer. Now I've got to dismiss a popup that wants me to talk to an AI to remote start the car instead of just pressing a button.
Can't imagine when consumer demand isn't surging for AI.
Something that bothers me with hn is that despite the amazing shit we see with AI, a big chunk of it has somehow decided that the corpo bloodsuckers are also amazing and should be protected at all costs. I wish we could have more nuanced conversations, considering the majority here think of themselves as some sort of high iq individuals.
It made me sad how it destroyed so much cohesion and eliminated people's jobs. I don't find that amazing. It is killing people literally now. Not everyone will be able to find a new job.
Just look at this guy - he retired mostly due to AI taken over ruby, as one example of many more:
And that's about a month ago only, aka quite recently.
AI changed everything. Mostly for the bad. In a few use cases one can say
that AI led to interesting changes, e. g. discovering bugs in projects (assuming
here these are real discovers about real bugs, but quite many were real bugs indeed).
This led to the death of all bug bounty programs. It makes no sense to pay humans when computers can do the same in a better way.
Many here are overconfident. They believe that expertise in one area means expertise in all areas.
Just browse any thread about renewable energy where the tech bro crowd believes that every challenge is a “solved problem”.
Were that the case, private industry would be building data centres powered by nothing but wind and solar energy. Yet, the preferred power solutions still appear to be the power grid or skirting emissions regulations via gas turbines.
> Also the dogma is "AI is inevitable so you better just get used to it and start using it or you will be left behind."
A someone who today handed his notice in as a dev, in preference to going full-on AI at a company that is determined to become a meat relay farm ("AI at the heart of everything we do!!!!!" -- new CEO, whose name it might be disrespectful that I don't remember), even someone like me who doesn't want to play that way can see that they're is definitely some truth in that dogma. I'm resigned to maybe never earning more than menial wages again, and if you refuse to get up to speed I think it'll get you that way too. Of course I expect is to get half of those who _are_ up to speed just as badly in the medium/long term. Programmers and such will be commodity workers like those in the service industry once almost anyone can do it, with the exception of a lucky minority.
> But if that were true, why would it be necessary to force the inevitable thing on everyone everywhere?
Because the people selling the inevitable thing want to force their version of it on you before others do, or force theirs into the mindset of those already converted but using something else, in the hope that when/if you do pay for a subscription it will be with them instead of the others.
The demand is not real. They are hunting for any market that can pay. Nvidia was forcing companies to buy hardware they did not need or they would not be allowed to buy in the future. This created artificial scarcity that skyrocketed prices. Large data centers don't make sense, they require massive expensive grid upgrades to centralize all the gpus. Put those same gpus in each home and no one needs to upgrade anything.
It looks like jensen did this to increase nvidia production, so it would be easier to smuggle gpus to china. US businesses and consumers have had to pay billions more for their hardware because of jensen's smuggling ring.
The only winners were jensen's bank account and chinese businesses that wanted nvidia gpus.
Most people don't want to pay for inference hardware that sits unused 90% of the time. So concentrating that hardware in data centers does actually make sense.
Most individual people don't want to pay for inference hardware at all. They want an affordable graphics card for gaming. While a lot of people are interested in local inference, there are many more who aren't so they are not at all a majority.
Concentrating the relevant compute resource in DCs makes sense, but to companies seeking to sell AI services, not people overall.
> Most people don't want to pay for inference hardware that sits unused 90% of the time.
Enough people have gotten so badly burned by the "SAAS trap" that if people are actually making real use out of it, they'd be very likely to buy the hardware if it was at ~2024 prices.
This is the thing that Altman omitted from his «Local inference won't work out like ya'll think it will.» comment during that interview over beers a little while back.
I think Stadia was pretty good and someone will figure out that business sooner rather than later (nVidia has a good product as well). Totally changed my mind on the concept of renting a server to play a game.
I believe the final solution will be that the gaming hardware will be installed at the next edge of your ISP to your home. The minimal latency you can have and much better than routing the stuff over the public net many milliseconds away to one of the main datacenters.
While I understand the underlying economic argument, I don't think it captures the whole picture. I'm not sure that hardware sitting idle 90% of the time necessarily means people aren't willing to pay for it because plenty of things we own spend most of their time unused. Think about it: cars are parked for most of their lifetime, a household drill might only get used a few hours a year. Utilization is certainly a factor, but I don't think it's necessarily a good measure of the value of ownership.
Google is not "cutting off Search from AI-free schools". They're refusing to supply AI-free Search. It takes less theorizing to understand if you phrase it this way.
Everyone complained there should be a way to disable AI, and that it should be on by default. For some reason (curiosity still makes sense here), Google did some 'malicious compliance' and added the AI switch-off without making a new version of search, and disabled AI by default.
My personal theory is some sort of moral liability thing, where they're hoping you blame your school district and tell them to switch platforms, so that if/when they don't, it's "their fault", not Google's.
> Google is not "cutting off Search from AI-free schools". They're refusing to supply AI-free Search. It takes less theorizing to understand if you phrase it this way.
can someone smarter than me explain in simpler words, how is that literally not the same thing?
It would be very expensive to create a custom version of search for these schools. It's not simply a matter of flipping a switch off. They decided that the cost of accommodating the schools wasn't worth the effort.
Search is already chock full of AI, removing it would be tons of work that Google doesn't want to do (since P0 is in fact cram even more LLMs into everything).
The main thing they don't want to disable is the AI summary at the top of the search, which eeeeh, can be seen as AI but won't let kids enter an AI mode /chat like some news stations are claiming. (Based off reporting I watched a week ago)
Imagine you have an otherwise AI positive school that wants AI-less Google. Google wouldn't give it to them, but this fact is not communicated by the less precise headline.
Google are avoiding an à la carte offering here. They are saying you either get search which now includes AI, or you can disable the entire feature. The obvious thing that people want is the original search without AI included.
They are not giving reasons why they can't unbundle these options. It could be technical but seems more likely that they want to appear to comply and shift the focus. Search being disabled creates pressure on those operating the tool. The local narrative becomes that the admin is the one that is limiting use. Eventually people lose sight of the real reasons and Google are on the sidelines watching users fight it out with admins.
This is a classic product management approach to dealing with objections. Remember that the PMs are getting measured on usage. If no one uses their feature then they are in trouble, so they make it hard to disable their features. A wonderful incentive structure, right?
Lots of companies build toggles that bundle up features as a result. You may want to disable this one thing, but toggling that off disables all these other things you want. The admin has pressure from the business or users. The vendor comes in to support the business or users, and then the IT team is overruled. Alternatively, the bundle is packed with items of real value plus mitigations are released elsewhere in the product (think: eDiscovery/logging/compliance) that cause IT to enable the whole bundle.
Google search has involved various forms of AI for a decade or more...
All those ranking algorithms that decide which link comes at the top? They use AI.
Sure - it doesn't look like a chatbot, but it's still AI under the covers and they couldn't in good conscious even give you search from 2008 and claim there was no AI involved.
I find these arguments absurd. You know full well what the schools mean. And that it is not the same thing as objection to 2008 technology. Which was not even called AI in 2008 but technically falls into onw of definitions.
The thing the schools want isn't well founded. Presumably they want their students not to be able to cheat, but a perfect search algorithm also allows them to cheat. As does, e.g. Wolfram Alpha-esque deterministic natural language systems.
We don't get AI search results at work. Google Search works fine, but the AI results don't load. Maybe Google can't do it from their end, but it can be done.
This is surely true. It took me no time at all to vibecode an extension that blocks AI on Google and other sites. [1] I'm sure it's not bulletproof, and students might be able to find a way around it, but it avoids having a chatbot front and center by default.
that's not AI-free search, that's blocking the ai mode from search.
If google signs a contract to deliver AI-free search, it means they have to deliver a search product that doesn't use ai to generate any of the search results they show, not just inside the ai-mode box but anywhere on the page. Which obviously is not a thing they want to guarantee.
In general, schools don't care if "AI is used to generate" search results. They care if a chatbot is shoved in front of students' faces every time they google something. The issue is the UI, not whether AI is used anywhere along the chain.
you're probably mostly correct, but given the current discourse around AI there's definitely at least some subset of people involved in pushing schools to go AI-free that really mean entirely AI-free.
And if google delivered a chatbot-free experience to a school for a contract that said AI-free, they're looking for the inevitable outrage when somebody finds out their search results are being generated in an AI datacenter.
Expensive for whom? Class action lawsuits are typically handled on contingency. If the lawsuit fails, the lawyers eat the time cost, and if the lawsuit succeeds, they are paid out of the funds paid by the wrongdoer.
Right now if you google "zelda" a Master Sword will appear in the bottom right corner. If you click on the sword it will open a 1998 version google with a search of "zelda".
They clearly have the ability to disable AI in search but just refuse to.
Maybe, but not necessarily. For example, I recently searched for someone who worked at a company using a plain language query. The AI identified one person (correctly, as it happens), but all the links below were to different people.
Obviously the results were not driven by the same AI that produced the summary.
They can be cute and say that the search results are AI, which while technically true is probably not what most people mean when they say they don't want AI. They could literally just disable the AI Overview at the top and that's what most schools mean when they say they don't want AI features.
So, yes, they are cutting off search from AI-free schools. The claim is precise and accurate. This is the net effect, and they can, in fact, opt to not do this but they will not. There is no technical hurdle. There is no economic hurdle. Doing so simply does not align with their incentives.
There is no virtue in appealing to nuance to maintain plausible deniability for these companies. It's absurd on its face.
It's interesting that in the lawsuit mentioned in the article, Google's defense was that they can't separate the AI overview from search as they are one product. That's almost the exact same argument Microsoft made in its antitrust defense about bundling IE. They went so far as to shoehorn IE into all kinds of system processes that were fine without IE before. But their argument didn't work and they were forced to offer the option to remove IE. But ironically, now the same argument is working for Google, despite the obvious solution of just having a setting to hide the AI overview.
Just wanted to thank you for using the text only link! This is the type of thing we should use and encourage.
I have a list somewhere I need to dig out of these types of link modifications available from when I was deployed. They were the only way to read news.
It's not like Google is the only search engine out there. Duck Duck Go has a toggle for the main search page, and a sub-domain where it's off by default: https://noai.duckduckgo.com/
someone needs to take the time to go to the meetings where the decision is made to go with one solution or another (and out-sell the person who is selling the alternative)
personally i feel i'm the only person in the district who is aware that there are email providers aside from gmail/yahoo
It seems pretty random when I use Google and it returns with me on the AI tab vs the All tab. Sometimes I want the AI response and so I word my question in a way that should return an AI response but I end up on the All tab - in these cases I assume it's a load issue on their servers.
Off topic, but when I give Google a list of ingredients that I want to cook something new with and ask for a recipe, the AI response is almost always exactly one of the recipes in the All responses, just stripped of extraneous information. AI is going to kill advertising.
As I recall, Microsoft tried something similar with Internet Explorer, claiming it was an integral part of their operating system. Back then the courts were not impressed with Microsoft's arguments.
I think this is a bad move from Google, not just morally, but also tactically.
Anti-tech skepticism and sentiment is at an all time high. The blue wave was fueled in part by this. People are tired of tech companies dictating to them how things will be done. Nothing is certain, but I suspect Google will be forced to backtrack.
AI is deeply embedded into Google search and has been for 15 years. Even if they could master it technically, the results and ranking would be so poor that the customer would complain anyway. Better to just turn it off entirely.
Even if students can't use google search, they can still type the address of another search engine into the browser, so there's no problem.
Just type duckduckgo.com and get your search done.
In my current district all other engines are blocked. My daughter has permission to use her phone to utilize Kagi from her teachers, but she cannot use it on the school devices.
I've recently learned my daughter, with teacher permission, regularly has to use her phone (in her otherwise phone-free school) to get around blocks on basic stuff. Like in her math class the teacher regularly has them review techniques using YouTube videos (from an approved channel), but YouTube gets blocked every so often. Why? Very possibly no one knows. So she'll be sent to the front office where they can unlock her phone and she can use it to keep doing her work. It also came up recently when she was researching drug addiction for a paper, and all related search terms were banned.
Yeah, you have to be real careful. Our son has his own Chromebook and he once logged into it with his school account to do something. It installed the district's certificates and other blocks so that even his personal account couldn't do anything. We had to factory reset the machine.
The stated reason, that free Google search now is an AI product, is technically correct. I hear that's the best kind of correct.
But if I had the option of a separate search product that didn't involve AI, I would be very interested in using it. I'd rather have more control over whether AI is involved in a search or not. Maybe I'm wrong. Maybe I'm an outlier. But I think it's a pretty good hypothesis that Google won't decouple search from AI because they think it might be too popular.
I don't have inside knowledge, but I'd be willing to bet that search without the LLM summary still involves plenty of non-LLM ML models along the way, classifying and pattern matching. “No LLM summaries” would be a tractable request; “no AI” would not be.
That's what I meant: when I use a search product, I want the whole screen devoted to search results. That's not to say a hybrid product that includes an LLM summary section is useless. I just want to decide when to see it.
Of my recent searches, I’m either 1) just clicking the top link anyways as I was searching for the exact #1 result and need their website, not info. Or 2)searching for a quick snippet of info where the AI summary is sufficient, like a cooking measurement conversion.
What’s your usage where AI in the results is an issue that AI-free results are better?
It's not that unusual for the AI summary to be incorrect, and that the citation link doesn't support the summary. I've seen it multiple times, and a recent study indicated it happened about 10% of the times.
Recently I was struggling to explain to a friend why I've disliked the AI in search - right from the early days. Yet I have my own bot that uses Kagi Search + Extract to get me information, and I love it.
It comes down to the UX and lack of choice. Sometimes I want the AI to go do the research and get me the answer. Other times I just want to do a search to find a link - I do not want a summary. It's noise and takes up space.
If Google (without requiring an account), let me default to no AI and have a separate button to get the AI summary, I'd likely be happier.
if you're giving it to kids where wrong information in the search summary could mislead them badly. The AI overviews, because they are highly optimized to be instantaenous and clearly draw from caches to some extent, often have really wrong information, like "opposite of correct" answers to things that are not as simple as cooking measurements.
But non-AI search results are no better. It’s easy to get opposite of correct answers via websites too. If it was a choice between looking it up in an encyclopedia and AI summaries, I’d get it, but there’s plenty of flat earth websites out there.
Maybe they wouldn't feel so awkward without their ai features if they hadn't sabotaged their search algorithms years ago to increase retention. You can't really find information on google search any more. It's just an index for content marketing, and these days, ai written content marketing.
I guess it turns up reddit and stackoverflow threads
I whipped up a Chrome extension that hides the AI on Google, Bing, etc. [1] It does not gather any user data and is 100% free.
It's laughable that Google says the two cannot be separated, since it was trivially easy to do — even for a third party like me. It would obviously be even easier for Google to accomplish.
They use this all-or-nothing tactic with gmail: I despise the unsolicited AI slop replies that Gmail offers me. I go to turn them off, and gmail is like "okay, guess you don't want any classification filters, all your messages can go to the same inbox".
No, Gmail. I just don't want you eroding the relationships with my friends and family by promoting fake replies.
This could be simple incompetence that was then doubled-down on. The ai-free product was eliminated (accepting this would result in acceptable loss of ad-views) and they simply didn't consider the consequences for the educational product.
Only after it became an issue would any of these theories likely occur. My experience is that legal isn't proactive enough to plan this upfront, but would set the policy reactively.
I have a take on this general subject that life would be easier for those of us using LLM for code and engineering if the greater community weren't so furious with all things AI, so to that degree, I really wish Google would do whatever they have to to provide a search interface with nothing "AI-looking" on it that schools who want to say "AI free" can use. Of course "AI" is involved with how the search engine works internally, how google develops it, etc. They want kids to not be exposed to AI-generated content.
"AI free" restores some burden to teachers who can't use it to create handy things like stories that include a specific list of vocabulary words and such, but I'd shudder to think what kind of crappy slop a "pro AI" school is dumping on kids, so in that sense, "AI free" is probably better.
Generative LLMs make it really obvious because they produce stylistically strong text that looks like a smart human wrote it. But maybe there’s a bigger moral here in the “Search uses AI under the hood, so…” argument and the fact that having search cut off is apparently a big crisis.
It turns out we’ve been offloading the cognitive work of organizing information to these ML based services for a long time! Is that… bad? Maybe? I mean google search had been going downhill for a long time as they lost the fight against SEO spammers, and the fact that there are not very many good alternatives has been really annoying.
My suspicion is that they basically know they’ve lost the fight against SEO spammers. This generative AI stuff is their path forward. There isn’t some static ground they can stand on to continue offering the service that people liked. We were all standing on a cliff over the sea being eroded by waves, and Google is giving up on building the scaffolding that held it together.
I dunno. It is annoying, but we don’t really have a right to ask them to pour resources into a dead-end product. And I can see why they wouldn’t want the next generation to get their impression of Google based on that dying product. But, the outcome seems to be a real PR mess for them. Part of that is how they’ve handled it, but part of that is probably just that events aren’t working out in their favor.
This explains almost everything whatever Google does and touch. Corruption kicks in when you get Google-degraded.
The Google you see today, has nothing to do with what Google once used to be. Once it became an adCompany, it changed. They realised this internally, so they removed the "don't do evil" motto as well.
I don’t think it really works this way (but hold on this isn’t going to be some corporate apologia). Companies don’t have morals, just responses to incentives and PR to build a narritive about it. Google was never not-evil; they were far enough ahead of everybody else to offer us all a really nice deal that was easy to spin as not-evil. They were an ad company more or less from the beginning, but a really rich one that had enough headroom to give us really nice services for free.
Now they are responding to the dual crunches of: their classic search techniques were insufficient to keep out folks who wanted things other than good search results to show up (SEO and the like). And they seem(?) to be losing to OpenAI and Anthropic on the “next thing” for organizing information, generative AI.
There is no way a school system would allow searches with any search engine without heavy filters. I’m sure you can imagine what boys would be searching for all day long.
Your reply makes sense if you cannot use other search engines on Chromebooks. That sounds like a monopoly/cartel issue if true. If it's not true, then I don't understand the relevance of your comment. All that the schools need to do is not use Google for search. What's the problem with that?
It's a bluff trying to push to schools against blocking AI.
The problem is that Google isn't the only search engine. It isn't even the best search engine anymore.
I really think this could be a great opportunity for Kagi to enter and show how good search can actually be.
I keep hearing AI companies say the demand is outstripping their supply and they need to build more data centers right now. And then I also see and read about how every company is cramming AI down everyone's throats in every product whether or not it's useful and/or wanted. I've used enough AI to see that it's revolutionary, but I'm also old enough to see the parallel to the dotcom days.
The demand seen by the big AI companies is largely from those others putting the AI features into everything. In this particular gold rush we actually have two sets of shovel sellers: Nvidia and the big AI companies.
And those shovels are being sold wholesale to software firms who recognize the only choices are adapt or be replaced...
...and, in my experience, those software firms are having a hell of a time selling shovels to bumpkins that seem skeptical of this gold rush (mid-sized organizations suffering from inflation and other economic headwinds)
What is the lesson of the dotcom days? There will be a brief pullback - and a couple goofy businesses will live in infamy - before that tech actually does go to the moon?
The lesson of the dotcom days was that transformative technology does not relieve a business from worrying about financial fundamentals.
The dotcom crash wasn't about technology, it was about business.
I think there are two things that can be true at once:
1) For about 1% of people AI is completely transformative to their jobs and let's them work much much faster. Many engineers could easily blow $5,000 in compute per month on AI, and that number is only going up as AI gets better. A year ago, it would have been insane to spend that much on AI, now it's normal. This is driving all the demand, and will continue to be useful.
2) For 99% of people they have no use for AI short of better search. All of the AI tools being crammed everywhere are businesses trying to figure out how to not "get left behind" and ride the hype train. There is no demand for this stuff and I see the parallels to the dotcom bubble.
The AI being packed into everything are lightweight models that are not compute intensive at all to run.
It's the behemoth models (and their agents) that are actually constrained. There is not enough compute to meet the demand for people offloading their day job sized models (and the agent teams they spawn).
The app I use to remote start my car just got a fancy new upgrade from the manufacturer. Now I've got to dismiss a popup that wants me to talk to an AI to remote start the car instead of just pressing a button.
Can't imagine when consumer demand isn't surging for AI.
Something that bothers me with hn is that despite the amazing shit we see with AI, a big chunk of it has somehow decided that the corpo bloodsuckers are also amazing and should be protected at all costs. I wish we could have more nuanced conversations, considering the majority here think of themselves as some sort of high iq individuals.
What exactly is amazing with AI?
It made me sad how it destroyed so much cohesion and eliminated people's jobs. I don't find that amazing. It is killing people literally now. Not everyone will be able to find a new job.
Just look at this guy - he retired mostly due to AI taken over ruby, as one example of many more:
https://po-ru.com/2026/10/18/why-im-no-longer-updating-this-...
And that's about a month ago only, aka quite recently.
AI changed everything. Mostly for the bad. In a few use cases one can say that AI led to interesting changes, e. g. discovering bugs in projects (assuming here these are real discovers about real bugs, but quite many were real bugs indeed). This led to the death of all bug bounty programs. It makes no sense to pay humans when computers can do the same in a better way.
Many here are overconfident. They believe that expertise in one area means expertise in all areas.
Just browse any thread about renewable energy where the tech bro crowd believes that every challenge is a “solved problem”.
Were that the case, private industry would be building data centres powered by nothing but wind and solar energy. Yet, the preferred power solutions still appear to be the power grid or skirting emissions regulations via gas turbines.
Also the dogma is "AI is inevitable so you better just get used to it and start using it or you will be left behind."
But if that were true, why would it be necessary to force the inevitable thing on everyone everywhere?
> Also the dogma is "AI is inevitable so you better just get used to it and start using it or you will be left behind."
A someone who today handed his notice in as a dev, in preference to going full-on AI at a company that is determined to become a meat relay farm ("AI at the heart of everything we do!!!!!" -- new CEO, whose name it might be disrespectful that I don't remember), even someone like me who doesn't want to play that way can see that they're is definitely some truth in that dogma. I'm resigned to maybe never earning more than menial wages again, and if you refuse to get up to speed I think it'll get you that way too. Of course I expect is to get half of those who _are_ up to speed just as badly in the medium/long term. Programmers and such will be commodity workers like those in the service industry once almost anyone can do it, with the exception of a lucky minority.
> But if that were true, why would it be necessary to force the inevitable thing on everyone everywhere?
Because the people selling the inevitable thing want to force their version of it on you before others do, or force theirs into the mindset of those already converted but using something else, in the hope that when/if you do pay for a subscription it will be with them instead of the others.
The demand is not real. They are hunting for any market that can pay. Nvidia was forcing companies to buy hardware they did not need or they would not be allowed to buy in the future. This created artificial scarcity that skyrocketed prices. Large data centers don't make sense, they require massive expensive grid upgrades to centralize all the gpus. Put those same gpus in each home and no one needs to upgrade anything.
It looks like jensen did this to increase nvidia production, so it would be easier to smuggle gpus to china. US businesses and consumers have had to pay billions more for their hardware because of jensen's smuggling ring.
The only winners were jensen's bank account and chinese businesses that wanted nvidia gpus.
Most people don't want to pay for inference hardware that sits unused 90% of the time. So concentrating that hardware in data centers does actually make sense.
Most individual people don't want to pay for inference hardware at all. They want an affordable graphics card for gaming. While a lot of people are interested in local inference, there are many more who aren't so they are not at all a majority.
Concentrating the relevant compute resource in DCs makes sense, but to companies seeking to sell AI services, not people overall.
> Most people don't want to pay for inference hardware that sits unused 90% of the time.
Enough people have gotten so badly burned by the "SAAS trap" that if people are actually making real use out of it, they'd be very likely to buy the hardware if it was at ~2024 prices.
This is the thing that Altman omitted from his «Local inference won't work out like ya'll think it will.» comment during that interview over beers a little while back.
Is that always so? That was the theory behind the various video-games-in-the-cloud services, but I don't think that worked out.
That was because it wasn’t technically good yet
Unity was notoriously trying to solve this but couldn’t find a way to fix latency
I think Google stadium(?) ran into the same issue
Concurrent shared state at raytracing speed across the internet is basically impossible
I think Stadia was pretty good and someone will figure out that business sooner rather than later (nVidia has a good product as well). Totally changed my mind on the concept of renting a server to play a game.
I believe the final solution will be that the gaming hardware will be installed at the next edge of your ISP to your home. The minimal latency you can have and much better than routing the stuff over the public net many milliseconds away to one of the main datacenters.
While I understand the underlying economic argument, I don't think it captures the whole picture. I'm not sure that hardware sitting idle 90% of the time necessarily means people aren't willing to pay for it because plenty of things we own spend most of their time unused. Think about it: cars are parked for most of their lifetime, a household drill might only get used a few hours a year. Utilization is certainly a factor, but I don't think it's necessarily a good measure of the value of ownership.
Started well, that sentence.
Google is not "cutting off Search from AI-free schools". They're refusing to supply AI-free Search. It takes less theorizing to understand if you phrase it this way.
Everyone complained there should be a way to disable AI, and that it should be on by default. For some reason (curiosity still makes sense here), Google did some 'malicious compliance' and added the AI switch-off without making a new version of search, and disabled AI by default.
My personal theory is some sort of moral liability thing, where they're hoping you blame your school district and tell them to switch platforms, so that if/when they don't, it's "their fault", not Google's.
> Google is not "cutting off Search from AI-free schools". They're refusing to supply AI-free Search. It takes less theorizing to understand if you phrase it this way.
can someone smarter than me explain in simpler words, how is that literally not the same thing?
It would be very expensive to create a custom version of search for these schools. It's not simply a matter of flipping a switch off. They decided that the cost of accommodating the schools wasn't worth the effort.
Search is already chock full of AI, removing it would be tons of work that Google doesn't want to do (since P0 is in fact cram even more LLMs into everything).
The main thing they don't want to disable is the AI summary at the top of the search, which eeeeh, can be seen as AI but won't let kids enter an AI mode /chat like some news stations are claiming. (Based off reporting I watched a week ago)
Imagine you have an otherwise AI positive school that wants AI-less Google. Google wouldn't give it to them, but this fact is not communicated by the less precise headline.
Google are avoiding an à la carte offering here. They are saying you either get search which now includes AI, or you can disable the entire feature. The obvious thing that people want is the original search without AI included.
They are not giving reasons why they can't unbundle these options. It could be technical but seems more likely that they want to appear to comply and shift the focus. Search being disabled creates pressure on those operating the tool. The local narrative becomes that the admin is the one that is limiting use. Eventually people lose sight of the real reasons and Google are on the sidelines watching users fight it out with admins.
This is a classic product management approach to dealing with objections. Remember that the PMs are getting measured on usage. If no one uses their feature then they are in trouble, so they make it hard to disable their features. A wonderful incentive structure, right?
Lots of companies build toggles that bundle up features as a result. You may want to disable this one thing, but toggling that off disables all these other things you want. The admin has pressure from the business or users. The vendor comes in to support the business or users, and then the IT team is overruled. Alternatively, the bundle is packed with items of real value plus mitigations are released elsewhere in the product (think: eDiscovery/logging/compliance) that cause IT to enable the whole bundle.
Google search has involved various forms of AI for a decade or more...
All those ranking algorithms that decide which link comes at the top? They use AI.
Sure - it doesn't look like a chatbot, but it's still AI under the covers and they couldn't in good conscious even give you search from 2008 and claim there was no AI involved.
It's pretty clear that in this context, and in much of the current discussion about "AI", that "AI" is used to mean generative language models.
yes but choosing what link to click and then making up your mind from the contents was not AI
I find these arguments absurd. You know full well what the schools mean. And that it is not the same thing as objection to 2008 technology. Which was not even called AI in 2008 but technically falls into onw of definitions.
The thing the schools want isn't well founded. Presumably they want their students not to be able to cheat, but a perfect search algorithm also allows them to cheat. As does, e.g. Wolfram Alpha-esque deterministic natural language systems.
USA frequently operates on gotchas involving explicit text of the law, not spirit
> can someone smarter than me explain in simpler words, how is that literally not the same thing?
You can't just buy peanuts without shells. They grow in that shell! That's what a peanut is!
We don't get AI search results at work. Google Search works fine, but the AI results don't load. Maybe Google can't do it from their end, but it can be done.
This is surely true. It took me no time at all to vibecode an extension that blocks AI on Google and other sites. [1] I'm sure it's not bulletproof, and students might be able to find a way around it, but it avoids having a chatbot front and center by default.
1: https://chromewebstore.google.com/detail/beeline-search-avoi...
that's not AI-free search, that's blocking the ai mode from search.
If google signs a contract to deliver AI-free search, it means they have to deliver a search product that doesn't use ai to generate any of the search results they show, not just inside the ai-mode box but anywhere on the page. Which obviously is not a thing they want to guarantee.
In general, schools don't care if "AI is used to generate" search results. They care if a chatbot is shoved in front of students' faces every time they google something. The issue is the UI, not whether AI is used anywhere along the chain.
you're probably mostly correct, but given the current discourse around AI there's definitely at least some subset of people involved in pushing schools to go AI-free that really mean entirely AI-free.
And if google delivered a chatbot-free experience to a school for a contract that said AI-free, they're looking for the inevitable outrage when somebody finds out their search results are being generated in an AI datacenter.
I don't think what schools "meant" when they said "no AI in search" will prevent expensive class action lawsuits.
Expensive for whom? Class action lawsuits are typically handled on contingency. If the lawsuit fails, the lawyers eat the time cost, and if the lawsuit succeeds, they are paid out of the funds paid by the wrongdoer.
Is it possible that there's at technical reason? Or not really
Right now if you google "zelda" a Master Sword will appear in the bottom right corner. If you click on the sword it will open a 1998 version google with a search of "zelda".
They clearly have the ability to disable AI in search but just refuse to.
No. You can just append "&udm=14" to your Google search URL to remove AI overview.
Could a school district IT employee automatically append this to all search requests sent over their network?
They would have to be using ssl decryption to append it at the network level. There may be extensions that would do it though?
I actually made an extension that hides the AI results (using a different method), and which works across Google, Bing, and other search engines. [1]
It's 100% free (of tracking and cost).
1: https://chromewebstore.google.com/detail/beeline-search-avoi...
most schools already are
Even if you disable the 'AI overview', the search results themselves are powered by AI models as well. They have been since at least ~2018 with BERT.
No one's objecting to that aspect of things.
Won't be first time there's unexpected consequences due to how someone formed the legal argument
Maybe, but not necessarily. For example, I recently searched for someone who worked at a company using a plain language query. The AI identified one person (correctly, as it happens), but all the links below were to different people.
Obviously the results were not driven by the same AI that produced the summary.
If I understand correctly, they already exclude AI overviews for under-18 users in the EU to comply with local laws.
This is software, there are no constraints. Technology is about political control and this is Google trying to exert it over school children.
They can be cute and say that the search results are AI, which while technically true is probably not what most people mean when they say they don't want AI. They could literally just disable the AI Overview at the top and that's what most schools mean when they say they don't want AI features.
So, yes, they are cutting off search from AI-free schools. The claim is precise and accurate. This is the net effect, and they can, in fact, opt to not do this but they will not. There is no technical hurdle. There is no economic hurdle. Doing so simply does not align with their incentives.
There is no virtue in appealing to nuance to maintain plausible deniability for these companies. It's absurd on its face.
So they maliciously complied, making it essentially the same thing.
Google loves these dark patterns they do it everywhere.
It's interesting that in the lawsuit mentioned in the article, Google's defense was that they can't separate the AI overview from search as they are one product. That's almost the exact same argument Microsoft made in its antitrust defense about bundling IE. They went so far as to shoehorn IE into all kinds of system processes that were fine without IE before. But their argument didn't work and they were forced to offer the option to remove IE. But ironically, now the same argument is working for Google, despite the obvious solution of just having a setting to hide the AI overview.
Just wanted to thank you for using the text only link! This is the type of thing we should use and encourage.
I have a list somewhere I need to dig out of these types of link modifications available from when I was deployed. They were the only way to read news.
There's a good list here - https://greycoder.com/a-list-of-text-only-new-sites/ !
It's not like Google is the only search engine out there. Duck Duck Go has a toggle for the main search page, and a sub-domain where it's off by default: https://noai.duckduckgo.com/
I wonder why these districts don't simply pivot to noai.duckduckgo.com. It's my default search engine and has served me well.
i haven't been following this story closely, but i always use https://www.google.com/search?q=foo&udm=web and it works fine.
someone needs to take the time to go to the meetings where the decision is made to go with one solution or another (and out-sell the person who is selling the alternative)
personally i feel i'm the only person in the district who is aware that there are email providers aside from gmail/yahoo
It seems pretty random when I use Google and it returns with me on the AI tab vs the All tab. Sometimes I want the AI response and so I word my question in a way that should return an AI response but I end up on the All tab - in these cases I assume it's a load issue on their servers.
Off topic, but when I give Google a list of ingredients that I want to cook something new with and ask for a recipe, the AI response is almost always exactly one of the recipes in the All responses, just stripped of extraneous information. AI is going to kill advertising.
I like the kagi way of doing things with adding a ? at the end of the search will bring up the AI Overview for the results.
Nah, the responses will just start including the advertising. It may kill marketing content webpages though.
I already saw ads being added forgot if it was chatgpt or google's gemini.
As I recall, Microsoft tried something similar with Internet Explorer, claiming it was an integral part of their operating system. Back then the courts were not impressed with Microsoft's arguments.
Simple: Because they can just do this and get away with it. It's the inevitable result of going all in in Google's ecosystem.
Although I acknowledge that there isn't necessarily a better alternative.
For search specifically, there are other good options. Schools hopefully use them.
Developing and maintaining a special variant of a product for years to come comes with a cost.
Here’s a simple theory: cost >> reward
Oh no! Cost! How will Google afford that!? Won't someone think of the shareholders.
A business is not a a charity
I think this is a bad move from Google, not just morally, but also tactically.
Anti-tech skepticism and sentiment is at an all time high. The blue wave was fueled in part by this. People are tired of tech companies dictating to them how things will be done. Nothing is certain, but I suspect Google will be forced to backtrack.
Related:
Chicago schools to Google – we'd rather lose Search than be forced to accept AI
https://news.ycombinator.com/item?id=50476586
AI is deeply embedded into Google search and has been for 15 years. Even if they could master it technically, the results and ranking would be so poor that the customer would complain anyway. Better to just turn it off entirely.
Even if students can't use google search, they can still type the address of another search engine into the browser, so there's no problem. Just type duckduckgo.com and get your search done.
This is not true at all districts.
In my current district all other engines are blocked. My daughter has permission to use her phone to utilize Kagi from her teachers, but she cannot use it on the school devices.
I've recently learned my daughter, with teacher permission, regularly has to use her phone (in her otherwise phone-free school) to get around blocks on basic stuff. Like in her math class the teacher regularly has them review techniques using YouTube videos (from an approved channel), but YouTube gets blocked every so often. Why? Very possibly no one knows. So she'll be sent to the front office where they can unlock her phone and she can use it to keep doing her work. It also came up recently when she was researching drug addiction for a paper, and all related search terms were banned.
I wonder if this is actually an opportunity for Kagi by having some kind of education pricing for schools specifically for this reason.
Came here to say that too.
Yeah, you have to be real careful. Our son has his own Chromebook and he once logged into it with his school account to do something. It installed the district's certificates and other blocks so that even his personal account couldn't do anything. We had to factory reset the machine.
So it's the schools that're making kids - and thus the next generation - dependent on Google.
Because for some reason AI features are an "essential part" of google search
we can just ignore the years without the AI features though trust
The stated reason, that free Google search now is an AI product, is technically correct. I hear that's the best kind of correct.
But if I had the option of a separate search product that didn't involve AI, I would be very interested in using it. I'd rather have more control over whether AI is involved in a search or not. Maybe I'm wrong. Maybe I'm an outlier. But I think it's a pretty good hypothesis that Google won't decouple search from AI because they think it might be too popular.
I don't have inside knowledge, but I'd be willing to bet that search without the LLM summary still involves plenty of non-LLM ML models along the way, classifying and pattern matching. “No LLM summaries” would be a tractable request; “no AI” would not be.
That's what I meant: when I use a search product, I want the whole screen devoted to search results. That's not to say a hybrid product that includes an LLM summary section is useless. I just want to decide when to see it.
Why is that?
Of my recent searches, I’m either 1) just clicking the top link anyways as I was searching for the exact #1 result and need their website, not info. Or 2)searching for a quick snippet of info where the AI summary is sufficient, like a cooking measurement conversion.
What’s your usage where AI in the results is an issue that AI-free results are better?
It's not that unusual for the AI summary to be incorrect, and that the citation link doesn't support the summary. I've seen it multiple times, and a recent study indicated it happened about 10% of the times.
Recently I was struggling to explain to a friend why I've disliked the AI in search - right from the early days. Yet I have my own bot that uses Kagi Search + Extract to get me information, and I love it.
It comes down to the UX and lack of choice. Sometimes I want the AI to go do the research and get me the answer. Other times I just want to do a search to find a link - I do not want a summary. It's noise and takes up space.
If Google (without requiring an account), let me default to no AI and have a separate button to get the AI summary, I'd likely be happier.
if you're giving it to kids where wrong information in the search summary could mislead them badly. The AI overviews, because they are highly optimized to be instantaenous and clearly draw from caches to some extent, often have really wrong information, like "opposite of correct" answers to things that are not as simple as cooking measurements.
But non-AI search results are no better. It’s easy to get opposite of correct answers via websites too. If it was a choice between looking it up in an encyclopedia and AI summaries, I’d get it, but there’s plenty of flat earth websites out there.
Maybe they wouldn't feel so awkward without their ai features if they hadn't sabotaged their search algorithms years ago to increase retention. You can't really find information on google search any more. It's just an index for content marketing, and these days, ai written content marketing.
I guess it turns up reddit and stackoverflow threads
I whipped up a Chrome extension that hides the AI on Google, Bing, etc. [1] It does not gather any user data and is 100% free.
It's laughable that Google says the two cannot be separated, since it was trivially easy to do — even for a third party like me. It would obviously be even easier for Google to accomplish.
1: https://chromewebstore.google.com/detail/beeline-search-avoi...
They use this all-or-nothing tactic with gmail: I despise the unsolicited AI slop replies that Gmail offers me. I go to turn them off, and gmail is like "okay, guess you don't want any classification filters, all your messages can go to the same inbox".
No, Gmail. I just don't want you eroding the relationships with my friends and family by promoting fake replies.
Just wait til they learn that "classic" google search going back to page rank is also a form of artificial intelligence
It’s because someone is asking for a different flavor of their product than what they offered to the rest of the entire world.
This could be simple incompetence that was then doubled-down on. The ai-free product was eliminated (accepting this would result in acceptable loss of ad-views) and they simply didn't consider the consequences for the educational product.
Only after it became an issue would any of these theories likely occur. My experience is that legal isn't proactive enough to plan this upfront, but would set the policy reactively.
I have a take on this general subject that life would be easier for those of us using LLM for code and engineering if the greater community weren't so furious with all things AI, so to that degree, I really wish Google would do whatever they have to to provide a search interface with nothing "AI-looking" on it that schools who want to say "AI free" can use. Of course "AI" is involved with how the search engine works internally, how google develops it, etc. They want kids to not be exposed to AI-generated content.
"AI free" restores some burden to teachers who can't use it to create handy things like stories that include a specific list of vocabulary words and such, but I'd shudder to think what kind of crappy slop a "pro AI" school is dumping on kids, so in that sense, "AI free" is probably better.
Generative LLMs make it really obvious because they produce stylistically strong text that looks like a smart human wrote it. But maybe there’s a bigger moral here in the “Search uses AI under the hood, so…” argument and the fact that having search cut off is apparently a big crisis.
It turns out we’ve been offloading the cognitive work of organizing information to these ML based services for a long time! Is that… bad? Maybe? I mean google search had been going downhill for a long time as they lost the fight against SEO spammers, and the fact that there are not very many good alternatives has been really annoying.
My suspicion is that they basically know they’ve lost the fight against SEO spammers. This generative AI stuff is their path forward. There isn’t some static ground they can stand on to continue offering the service that people liked. We were all standing on a cliff over the sea being eroded by waves, and Google is giving up on building the scaffolding that held it together.
I dunno. It is annoying, but we don’t really have a right to ask them to pour resources into a dead-end product. And I can see why they wouldn’t want the next generation to get their impression of Google based on that dying product. But, the outcome seems to be a real PR mess for them. Part of that is how they’ve handled it, but part of that is probably just that events aren’t working out in their favor.
Number one theory: Google is evil.
This explains almost everything whatever Google does and touch. Corruption kicks in when you get Google-degraded.
The Google you see today, has nothing to do with what Google once used to be. Once it became an adCompany, it changed. They realised this internally, so they removed the "don't do evil" motto as well.
I don’t think it really works this way (but hold on this isn’t going to be some corporate apologia). Companies don’t have morals, just responses to incentives and PR to build a narritive about it. Google was never not-evil; they were far enough ahead of everybody else to offer us all a really nice deal that was easy to spin as not-evil. They were an ad company more or less from the beginning, but a really rich one that had enough headroom to give us really nice services for free.
Now they are responding to the dual crunches of: their classic search techniques were insufficient to keep out folks who wanted things other than good search results to show up (SEO and the like). And they seem(?) to be losing to OpenAI and Anthropic on the “next thing” for organizing information, generative AI.
Being ant AI in 2026 is like being anti nuclear in 1976.
You mean we should be bracing for the AI analogue of Chernobyl?
How so?
Reminder: Three Mile Island happened in 1979 :-)
How is that a problem? There is a ton of alternative search engines. To be honest, students shouldn't even use Google search in the classroom.
There is no way a school system would allow searches with any search engine without heavy filters. I’m sure you can imagine what boys would be searching for all day long.
are you aware that entire school districts give out Chromebooks to all students?
Google is ubiquitous in the educational system right now.
Google is also discontinuing chromebooks.
looks like they will be called...."Googlebooks" running "GooglebookOS" rather than "Chromebooks" running "ChromeOS"
this...doesn't seem to matter much.
Your reply makes sense if you cannot use other search engines on Chromebooks. That sounds like a monopoly/cartel issue if true. If it's not true, then I don't understand the relevance of your comment. All that the schools need to do is not use Google for search. What's the problem with that?