Somebody burned compute to send him an LLM-generated thank-you note. Everybody involved in this transaction lost, nobody gained anything from it. It's pure destruction of resources.
Automated systems sending people unsolicited, unwanted emails is more commonly known as spam.
Especially when the spam comes with a notice that it is from an automated system and replies will be automated as well.
Communication happen between two parties. I wouldn't consider LLM an party considering it's just an autosuggestion on steroids at the end of day (lets face it)
Also if you need communication like this, just share the prompt anyway to that other person in the letter, people much rather might value that.
I would wager that the amount of “very significant thing that have happened over the history of humanity” come down to a few emotional responses.
If you insert AI into your thinking process, it has a bias, for sure. It will helpfully reinforce whatever you tell it you think makes sense, or at least on average it will be interpreted that way because of a wide variety of human cognitive biases even if it hedges. At the least it will respond with ideas that are very... median.
So at each one of these tiny branches you introduce a bias towards the "typical" instead of discovering where your own mind would go. It's fine and conversational but it clearly influences your thought process to, well, mitigate your edges. Maybe it's more "correct", it's certainly less unique.
And then at some point they start charging for the service. That's the part I'm concerned about, if it's on-device and free to use I still think it makes your thought process less interesting and likely to have original ideas, but having to subscribe to a service to trust your decision making is deeply concerning.
This, speaking about environmental impacts. I wish that more models start focusing on the parameter density / their compactness more so that they can run locally but this isn't something that big tech really wants so we are probably gonna get models like the recent minimax model or glm air models or qwen or mistral models.
These AI services only work as long as they are free and burn money. As an example, me and my brother were discussing something yesterday related to LLM and my mother tried to understand and talk about it too and wanted to get ghibli styles photo since someone had ghibli generated photo as their pfp and she wanted to try it too
She then generated the pictures and my brother did a quick calculation and it took around 4 cents for each image which with PPP in my country and my currency is 3 ruppees.
When asked by my brother if she would pay for it, she said that no she's only using it for free but she also said that if she were forced to, she might even pay 50 rupees.
I jumped in the conversation and said nobody's gonna force her to make ghibli images.
When the thought is "I owe this person a 'Thank You'", the handwritten letter gives an illusion of deeper thought. That's why there are fonts designed to look handwritten. To the receiver, they're just junk mail. I'd rather not get them at all, in any form. I was happy just having done the thing, and the thoughtless response slightly lessens that joy.
If you're being accurate, the people you know are terrible.
If someone sends me a personal letter [and I gather we're talking about a thank-you note here], I'm sure as hell going to open it. I'll probably even save it in a box for an extremely long time.
Yes so this is why the reason why person card/letters really matter because most people sheldom get any and if you know a person in your life / in any (community/project) that you deeply admire, sending them a handwritten mail can be one of the highest gestures which shows that you took the time out of your day and you really cared about them so much in a way.
That's my opinion atleast.
That interpretation doesn't save the comment, it makes it totally off topic.
FFS. AI's greatest accomplishment is to debase and destroy.
Trillions of dollars invested to bring us back to the stone age. Every communications technology from writing onward jammed by slop and abandoned.
When people have choices and they choose the more harmful action, it hurts their credibility. If Rob cares so much about society and the environment, why did he work at a company that has horrendous track record on both? Someone of his level of talent certainly had choices, and he chose to contribute to the company that abandoned “don’t be evil” a long time ago.
Ian Lance Taylor on the other hand appeared to have quit specifically because of the "AI everything" mandate.
Just an armchair observation here.
"AI" (and don't get me wrong I use these LLM systems constantly) is off the charts compared to normal data centre use for ads serving.
And so it's again, a kind of whataboutism that pushes the scale of the issue out of the way in order to make some sort of moral argument which misses the whole point.
BTW in my first year at Google I worked on a change where we made some optimizations that cut the # of CPUs used for RTB ad serving by half. There were bonuses and/or recognition for doing that kind of thing. Wasteful is a matter of degrees.
It wasn't only about serving those ads though, traditional machine-learning (just not LLMs) has always been computationally expensive and was and is used extensively to optimize ads for higher margins, not for some greater good.
Obviously, back then and still today, nobody is being wasteful because they want to. If you go to OpenAI today and offer them a way to cut their compute usage in half, they'll praise you and give you a very large bonus for the same reason it was recognized & incentivized at Google: it also cuts the costs.
Did you sell all of your stock?
Honestly I believe that google might be one of the few winners from the AI industry perhaps because they own the whole stack top to bottom with their TPU's but I would still stray away from their stock because their P/E ratio might be insanely high or something
Their p/e ratio has almost doubled in just a year which isn't a good sign https://www.macrotrends.net/stocks/charts/googl/alphabet/pe-...
So like, we might be viewing the peaks of the bubble and you might still hold the stocks and might continue holding it but who knows what happens after the stock depreciates value due to AI Bubble-like properties and then you might regret as why you didn't sell it but if you do and google's stock rises, you might still regret.
I feel as if grass is always greener but not sure about your situation but if you ask me, you made the best out of the situation with the parameters you had and logically as such I wouldn't consider it "unfortunately" but I get what you mean.
But with remote work it also became possible to get paid decently around here without working there. Prior I was bound to local area employers of which Google was the only really good one.
I never loved Google, I came there through acquisition and it was that job with its bags of money and free food and kinda interesting open internal culture, or nothing because they exterminated my prior employer and and made me move cities.
After 2016 or so the place just started to go downhill faster and faster though. People who worked there in the decade prior to me had a much better place to work.
I am super curious as I don't get to chat with people who have worked at google as so much so pardon me but I got so many questions for you haha
> It was a weird place to work
What was the weirdness according to you, can you elaborate more about it?
> I never loved Google, I came there through acquisition and it was that job with its bags of money and free food and kinda interesting open internal culture, or nothing because they exterminated my prior employer and and made me move cities.
For context, can you please talk more about it :p
> After 2016 or so the place just started to go downhill faster and faster though
What were the reasons that made them go downhill in your opinion and in what ways?
Naturally I feel like as organizations move and have too many people, maybe things can become intolerable to work but I have heard it be described as it depends where and in which project you are and also how hard it can be to leave a bad team or join a team with like minded people which perhaps can be hard if the institution gets micro-managed at every level due to just its sheer size of employees perhaps?
Not at all. I actually prefer in-office. And left when Google was mostly remote. But remote opened up possibilities to work places other than Google for me. None of them have paid as well as Google, but have given more agency and creativity. Though they've had their own frustrations.
> What was the weirdness according to you, can you elaborate more about it?
I had a 10-15 year career before going there. Much of what is accepted as "orthodoxy" at Google rubbed me the wrong way. It is in large part a product of having an infinite money tree. It's not an agile place. Deadlines don't matter. Everything is paid for by ads.
And as time goes on it became less of an engineering driven place and more of a product manager driven place with classical big-company turf wars and shipping the org chart all over the place.
I'd love to get paid Google money again, and get the free food and the creature comforts, etc. But that Google doesn't exist anymore. And they wouldn't take my back anyways :-)
No, but it puts some perspective on things. IMO Google, after abandoning its early "don't be evil" motto is directly responsible for a significant chunk of the current evil in the developed world, from screen addiction to kids' mental health and social polarization.
Working for Google and drawing an extravagant salary for many, many years was a choice that does affect the way we perceive other issues being discussed by the same source. To clarify: I am not claiming that Rob is evil; on the contrary. His books and open source work were an inspiration to many, myself included. But I am going to view his opinions on social good and evil through the prism of his personal employment choices. My 2c.
Certainly. But this, IMO, is not the reason for the criticism in the comments. If Rob ranted about AI, about spam, slop, whatever, most of those criticizing his take would nod instead.
However, the one and only thing that Rob says in his post is "fuck you people who build datacenters, you rape the planet". And this coming from someone who worked at Google from 2004 to 2021 and instead could have picked any job anywhere. He knew full well what Google was doing; those youtube videos and ad machines were not hosted in a parallel universe.
I have no problem with someone working at Google on whatever with full knowledge that Google is pushing ads, hosting videos, working on next gen compute, LLM, AGI, whatever. I also have no problem with someone who rails against cloud compute, AI, etc. and fights it as a colossal waste or misallocation of resources or whatever. But not when one person does both. Just my 2c, not pushing my worldview on anyone else.
If rob pike was asked about these issues of systemic addiction and others where we can find things google was bad at. I am sure that he wouldn't defend google about these things.
Maybe someone can mail a real message asking Rob pike genuinely (without any snarkiness that I feel from some comments here) about some questionable google things and I am almost certain that if those questions are reasonable, rob pike will agree that some actions done by google were wrong.
I think its just that rob pike got pissed off because an AI messaged him so he got the opportunity to talk about these issues and I doubt that he got the opportunity to talk / someone asking him about some other flaws of google / systemic issues related to it.
Its like, Okay, I feel like there is an issue in the world so I talk about it. Now does that mean that I have to talk about every issue in the world, no not really. I can have priorities in what issues I wish to talk about.
But that being said, if someone then asks me respectfully about issues which are reasonable, Being moral, I can agree about that yes those are issues as well which needs work upon.
And some people like rob pike who left google because of (ideological reasons perhaps, not sure?) wouldn't really care about the fallback and like you say, its okay to collect checks from organization even if they critize
Honestly Google's lucky that they got rob pike instead of vice versa from my limited knowledge.
Golang is such a brilliant language and ken thompson and rob pike are consistently some of the best coders and their contributions to golang and so many other projects is unparalleled.
I don't know much about rob pike as compared to Ken thompson but I assume he is really great too! Mostly I am just a huge golang fan.
Im not saying nobody has the right to criticize something they are supporting, but it does say something about our choices and how far we let this problem go before it became too much to solve. And not saying the problem isn't solvable. Just saying its become astronomically more difficult now then ever before.
I think at the very least, there is a little bit of cringe in me every time I criticize the very thing I support in some way.
Being a hypocrite makes you a bad person sometimes. It doesn't actually change anything factual or logical about your arguments. Hypocrisy affects the pathos of your argument, but not the logos or ethos! A person who built every single datacenter would still be well qualified to speak about how bad datacenters are for the environment. Maybe their argument is less convincing because you question their motives, but that doesn't make it wrong or invalid.
Unless HNers believe he is making this argument to help Google in some way, it doesn't fucking matter that google was also bad and he worked for them. Yes he worked for google while they built out datacenters and now he says AI datacenters are eating up resources, but is he wrong?. If he's not wrong, then talk about hypocrisy is a distraction.
HNers love arguing to distract.
"Don't hate the player, hate the game" is also wrong. You hate both.
Sometimes facts and logic can only get you so far.
Hypocrisy is when you criticise others for doing a thing you yourself secretly do. It is massively different then criticising a compant you work or worked for. You can even ve part of something, change opinion and then criticise it without being hypocryte.
With all due respect, being moral isn't an opinion or agreement about an opinion, it's the logic that directs your actions. Being moral isn't saying "I believe eating meat is bad for the planet", it's the behaviour that abstains from eating meat. Your moral is the set of statements that explains your behaviour. That is why you cannot say "I agree that domestic violence is bad" while at the same time you are beating up your spouse.
If your actions contradict your stated views, you are being a hypocrite. This is the point that people in here are making. Rob Pike was happy working at Google while Google was environmentally wasteful (e-waste, carbon footprint and data center related nastiness) to track users and mine their personal and private data for profit. He didn't resign then nor did he seem to have caused a fuss about it. He likely wasn't interested in "pointless politics" and just wanted to "do some engineering" (just a reference to techies dismissing or critising folks discussing social justices issues in relation to big tech). I am shocked I am having to explain this in here. I understand this guy is an idol of many here but I would expect people to be more rational on this website.
We got to this point by not looking at these problems for what they are. Its not wrong to say something is wrong and it needs to be addressed.
Doing cool things, without looking at whether or not we should doesn't feel very responsible too me esp. if it impacts society in a negative way.
For example, Rob seems not to realize that the people who instructed an AI agent to send this email are a handful of random folks (https://theaidigest.org/about) not affiliated with any AI lab. They aren't themselves "spending trillions" nor "training your monster". And I suspect the AI labs would agree with both Rob and me that this was a bad email they should not have sent.
We need to find a way to stop contributing to the destruction of the planet soon.
I don't work for any of these companies, but I do purchase things from Amazon and I have an apple phone. I think the best we can do is minimize our contribution to it. I try to limit what services I use from this companies, and I know it doesnt make much of a differnce, but I am doing what I can.
I'm hoping more people that need to be employed by tech companies can find a way to be more selective on who they employ with.
Data centers are not another thing when the subject is data centers.
The AI boom has completely changed that. Data center power usage is rocketing upwards now. It is estimated it will be more than 10% of all electric power usage in the US by 2030.
It's a completely different order of magnitude than the pre AI-boom data center usage.
It might help to look at global power usage, not just the US, see the first figure here:
https://arstechnica.com/ai/2024/06/is-generative-ai-really-g...
There isn't an inflection point around 2022: it has been rising quickly since 2010 or so.
So if you looked at a graph of energy consumption, you wouldn't even notice crypto. In fact even LLM stuff will just look like a blip unless it scales up substantially more than its currently trending. We use vastly more more energy than most appreciate. And this is only electrical energy consumption. All energy consumption is something like 185,000 TWh. [1]
[1] - https://ourworldindata.org/energy-production-consumption
Figure 1.1 is the chart I was referring to, which are the data points from the original sources that it uses.
Between 2010 and 2020, it shows a very slow linear growth. Yes, there is growth, but it's quite slow and mostly linear.
Then the slope increases sharply. And the estimates after that point follow the new, sharper growth.
Sorry, when I wrote my original comment I didn't have the paper in front of me, I linked it afterwards. But you can see that distinct change in rate at around 2020.
Figure 1.1 does show a single source from 2018 (Shehabi et al) that estimates almost flat growth up to 2017, that's true, but the same graph shows other sources with overlap on the same time frame as well, and their estimates differ (though they don't span enough years to really tell one way or another).
"yeah but they became efficient at it by 2012!"
Yes, data center efficiency improved dramatically between 2010-2020, but the absolute scale kept growing. So you're technically both right: efficiency gains kept/unit costs down while total infrastructure expanded. The 2022+ inflection is real though, and its not just about AI training. Inference at scale is the quiet energy hog nobody talks about enough.
What bugs me about this whole thread is that it's turning into "AI bad" vs "AI defenders," when the real question should be: which AI use cases actually justify this resource spike? Running an LLM to summarize a Slack thread probably doesn't. Using it to accelerate drug discovery or materials science probably does. But we're deploying this stuff everywhere without any kind of cost/benefit filter, and that's the part that feels reckless.
How much of that compute was for the ads themselves vs the software useful enough to compel people to look at the ads?
To stretch the analogy, all the "babies" in the "bathwater" of youtube that I follow are busy throwing themselves out by creating or joining alternative platforms, having to publicly decry the actions Google takes that make their lives worse and their jobs harder, and ensuring they have very diversified income streams and productions to ensure that WHEN, not IF youtube fucks them, they won't be homeless.
They mostly use Youtube as an advertising platform for driving people to patreon, nebula, whatever the new guntube is called, twitch, literal conventions now, tours, etc.
They've been expecting youtube to go away for decades. Many of them have already survived multiple service deaths, like former Vine creator Drew Gooden, or have had their business radically changed by google product decisions already.
I am not accurate about google but facebook definitely has some of the most dystopian tracking I have heard. I might read the facebook files some day but the dystopian fact that facebook tracks young girls and sees if that they delete their photos, they must feel insecure and serves them beauty ads is beyond predatory.
Honestly, my opinion is that something should be done about both of these issues.
But also its not a gotcha moment for Rob pike that he himself was plotting up the ads or something.
Regarding the "iphone kids", I feel as if the best thing is probably an parental level intervention rather than waiting for an regulatory crackdown since lets be honest, some kids would just download another app which might not have that regulation.
Australia is implementing social media ban basically for kids but I don't think its gonna work out but everyone's looking at it to see what's gonna happen basically.
Personally I don't think social media ban can work if VPN's just exist but maybe they can create such an immense friction but then again I assume that this friction might just become norm. I assume many of you guys must have been using internet from the terminal days where the friction was definitely there but the allure still beat the friction.
You're tilting at windmills here, we can't go back to barter.
It's literally impossible to start or run a business without advertising your products or services.
It is like an arms race. Everyone would have been better off if people just never went to war, but....
A lot of advertising is telling people about some product or service they didn't even know existed though. There may not even be a competitor to blame for an advertising arms race.
And before the LLM craze there was a constant focus on efficiency. Web search is (was?) amazingly efficient per query.
There is no upside to vast majority of the AI pushed by the OpenAI and their cronies. It's literally fucking up economy for everyone else all to get AI from "lies to users" to "lies to users confidently", all while rampantly stealing content to do that, because apparently pirating something as a person is terrible crime govt need to chase you, unless you do that to resell it in AI model, then it's propping up US economy.
If there is any example of hypocrisy, and that we don't have a justice system that applies the law equally, that would be it.
I mean, buying another pair of sneakers you don't need just because ads made you want them doesn't sound like the best investment from a societal perspective. And I am sure sneakers are not the only product that is being bought, even though nobody really needs them.
I.e., they are proud to have never intentionally used AI and now they feel like they have to maintain that reputation in order to remain respected among their close peers.
LLMs need to burn significant amounts of power for every inference. They're exponentially more power hungry than searches, database lookups, or even loads from disk.
If they want LLM, you probably don't have to advertise them as much
No the reality of the matter is that people are being shoved LLM's. They become the talk of the town and algorithms share any development related to LLM or similar.
The ads are shoved down to users. Trust me, the average person isn't as much enthusiastic about LLM's and for good reasons when people who have billions of dollars say that yes its a bubble but its all worth it or similar and the instances where the workforce themselves are being replaced/actively talked about being replaced by AI
We live in an hackernews bubble sometimes of like-minded people or communities but even on hackernews we see disagreements (I am usually Anti AI mostly because of the negative financial impact the bubble is gonna have on the whole world)
So your point becomes a bit moot in the end but that being said, Google (not sure how it was in the past) and big tech can sometimes actively promote/ close their eyes if the ad sponsors are scammy so ad-blockers are generally really good in that sense.
But I do derive value from owning a car. (Whether a better world exists where my and everyone else's life would be better if I didn't is a definitely a valid conversation to have.)
The user doesn't derive value from ads, the user derives value from the content on which the ads are served next to.
The value you derive is the ability to make your car move. If you derived no value from gas, why would you spend money on it?
If they just wanted ads blasted at them, and nothing else, they'd be doing something else, like, say, watching cable TV.
Like, the ratio is not too crazy, it's rather the large resource usages that comes from the aggregate of millions of people choosing to use it.
If you assume all of those queries provide no value then obviously that's bad. But presumably there's some net positive value that people get out of that such that they're choosing to use it. And yes, many times the value of those queries to society as a whole is negative... I would hope that it's positive enough though.
Well the people who burnt compute got it from money so they did burn money.
But they don't care about burning money if they can get more money via investors/other inputs faster than they can burn (fun fact: sometimes they even outspend that input)
So in a way the investors are burning their money, now they burn the money because the market is becoming irrational. Remember Devin? Yes cognition labs is still there etc. but I remember people investing into these because of their hype when it did turn out to be moot comparative to their hype.
But people/market was so irrational that most of these private equities were unable to invest in something like openai that they are investing in anything AI related.
And when you think more deeper about all the bubble activities. It becomes apparent that in the end bailouts feel more possible than not which would be an tax on average taxpayers and they are already paying an AI tax in multiple forms whether it be in the inflation of ram prices due to AI or increase in electricity or water rates.
So repeat it with me: whose gonna pay for all this, we all would but the biggest disservice which is the core of the argument is that if we are paying for these things, then why don't we have a say in it. Why are we not having a say in AI related companies and the issues relating to that when people know it might take their jobs etc. so the average public in fact hates AI (shocking I know /satire) but the fact that its still being pushed shows how little influence sometimes public can have.
Basically public can have any opinions but we won't stop is the thing happening in AI space imo completely disregarding any thoughts about the general public while the CFO of openAI proposing an idea that public can bailout chatgpt or something tangential.
Shaking my head...
Just like the invention of Go.
I find it difficult to express how strongly I disagree with this sentiment.
The value or "moral" fork would be trying to convince you that building, producing, and growing was actually helpful rather than harmful.
I don't imagine we actually disagree on the physical fork, making that argument pretty pointless: clearly humans and human civilization are learning, growing, and still have a strong potential to thrive as long as ASI, apathy, or a big rock don't take us out first. Instead, I took your statement as an indication that you don't actually positively value humans, more humans, humans growing, and humans building things. That's a preferences and values disagreement, and there's no way to rationally or logically argue someone into changing their core values. No ought from is, and all that.
I'm not suggesting, by the way, that people's values don't change, or can't be changed by discussion, only that there is no way to do so with logical argument; reason can get you to your goal, but it can't tell you what ultimate goal to want.
Anyway, I was expressing that I like humans and want humans (or people who themselves used to be humans, in the limit) to continue and do more, rather than arguing that you ought to feel the same.
I did not intend to make a moral argument.
That's just not true... When a mother nurses her child and then looks into their eyes and smiles, it takes the utmost in cynical nihilism to claim that is harmful.
But everything humans do does that. Everything increases entropy. Sometimes we find that acceptable. So when people respond to Pike by pointing out that he, too, is part of society and thus cannot have the opinion that LLMs are bad, I do not find that argument compelling, because everybody draws that line somewhere.
Down the street from it is an aluminum plant. Just a few years after that data center, they announced that they were at risk of shutting down due to rising power costs. They appealed to city leaders, state leaders, the media, and the public to encourage the utilities to give them favorable rates in order to avoid layoffs. While support for causes like this is never universal, I'd say they had more supporters than detractors. I believe that a facility like theirs uses ~400 MW.
Now, there are plans for a 300 MW data center from companies that most people aren't familiar with. There are widespread efforts to disrupt the plans from people who insist that it is too much power usage, will lead to grid instability, and is a huge environmental problem!
This is an all too common pattern.
Easier for a politician to latch onto manufacturing jobs.
You don't just chuck ore into a furnace and wait for a few seconds in reality.
I'd guess that this is also an area where the perception makes a bigger difference than the reality.
(NB: I am currently working in AI, and have previously worked in adtech. I'm not claiming to be above the fray in any way.)
Obviously now it is mostly the latter and minimally the former. What capitalism giveth, it taketh away. (Or: Capitalism without good market design that causes multiple competitors in every market doesn't work.)
The amount of “he’s not allowed to have an opinion because” in this thread is exhausting. Nothing stands up to the purity test.
He sure was happy enough to work for them (when he could work anywhere else) for nearly two decades. A one line apology doesn't delete his time at Google. The rant also seems to be directed mostly if not exclusively towards GenAI not Google. He even seems happy enough to use Gmail when he doesn't have to.
You can have an opinion and other people are allowed to have one about you. Goes both ways.
[0]: https://www.tomshardware.com/tech-industry/artificial-intell...
That dam took 10 years to build and cost $30B.
And OpenAI needs more than ten of them in 7 years.
We want free services and stuff, complain about advertising / sign up for the google's of the world like crazy.
Bitch about data-centers while consuming every meme possible ...
It's like all those anti-copyright activists from the 90s (fighting the music and film industry) that suddenly hate AI for copyright infringements.
Maybe what's bothering the critics is actually deeper than the simple reasons they give. For many, it might be hate against big tech and capitalism itself, but hate for genAI is not just coming from the left. Maybe people feel that their identity is threatened, that something inherently human is in the process of being lost, but they cannot articulate this fear and fall back to proxy arguments like lost jobs, copyright, the environment or the shortcomings of the current implementations of genAI?
The points you raise, literally, do not affect a thing.
Furthermore, w.r.t the points you raised - it's a matter of scale and utility. Compared to everything that has come before, GenAI is spectacularly inefficient in terms of utility per unit of compute (however you might want to define these). There hasn't been a tangible nett good for society that has come from it and I doubt there would be. The egarness and will to throw money and resources at this surpasses the crypto mania which was just as worthless.
Even if you consider Rob a hypocrite , he isn't alone in his frustration and anger at the degradation of the promise of Open Culture.
People being more productive with writing code, making music or writing documents fpr whatever is not a improvement for them and therefore for society?
Or do you claim that is all imaginary?
Or negated by the energy cost?
And all at significant opportunity cost (in terms of computing and investment)
If it was as life altering as they claim where's that novel work of art (in your examples..of code, music or literature) that truly could not have been produced without GenAI and fundamentally changed the art form ?
Surely, with all that ^increased productivity^ we'd have seen the impact equivalent of linux, apache, nginx, git, redis, sqlite, ... Etc being released every couple of weeks instead of yet another VSCode clone./s
The overall resource efficiency of GenAI is abysmal.
You can probably serve 100x more Google Search queries with the same resources you'd use for Google Gemini queries (like for like, Google Search queries can be cached, too).
> Nope, you can't, and it takes a simple Gemini query to find out more about the actual x if you are interested in it. (closer to 3, last time I checked, which rounds to 0, specially considering the clicks you save when using the LLM)
Why would you lie: https://imgur.com/a/1AEIQzI ???
For those that don't want to see the Gemini answer screenshot, best case scenario 10x, worst case scenario 100x, definitely not "3x that rounds to 0x", or to put it in Gemini's words:
> Summary
> Right now, asking Gemini a question is roughly the environmental equivalent of running a standard 60-watt lightbulb for a few minutes, whereas a Google Search is like a momentary flicker. The industry is racing to make AI as efficient as Search, but for now, it remains a luxury resource.
The reason why it all rounds to 0 is that the google search will not give you an answer. It gives you a list of web pages, that you then need to visit (often times more than just one of them) generating more requests, and, more importantly, it will ask more of your time, the human, whose cumulative energy expenditure to be able to ask to be begin with is quite significant – and that you then will have not to spend on other things that a LLM is not able to do for you.
Yes, Google Search is raw info. Yes, Google Search quality is degrading currently.
But Gemini can also hallucinate. And its answers can just be flat out wrong because it comes from the same raw data (yes, it has cross checks and it "thinks", but it's far from infallible).
Also, the comparison of human energy usage with GenAI energy usage is super ridiculous :-)))
Animal intelligence (including human intelligence) is one of the most energy efficient things on this planet, honed by billions years of cut throat (literally!) evolution. You can argue about time "wasted" analysing search results (which BTW, generally makes us smarter and better informed...), but energy-wise, the brain of the average human uses as much energy as the average incandescent light bulb to provide general intelligence (and it does 100 other things at the same time).
Talking about "condescending":
> super ridiculous :-)))
It's not the energy efficient animal intelligence that got us here, but a lot of completely inefficient human years to begin with, first to keep us alive and then to give us primary and advanced education and our first experiences to become somewhat productive human beings. This is the capex of making a human, and it's significant – specially since we will soon die.
This capex exists in LLMs but rounds to zero, because one model will be used for +quadrillions of tokens. In you or me however, it does not round to zero, because the number of tokens we produce round to zero. To compete on productivity, the tokens we have produce therefore need to be vastly better. If you think you are doing the smart thing by using them on compiling Google searches you are simply bad at math.
For instance, in the Gemini screenshot, the claim for 100-500x more resource usage for AI queries comes from water usage, however it's not clear to me why data center water usage for AI queries would be 100-500x more than a Google search when power usage for an AI query is supposedly only 10-30x more than a Google search. Is water usage and CO2 footprint not derived from power consumption? Did the LLM have to drink as much water while thinking as I did while researching the original claim?
The 10-30x more power consumption claim seems to come from this scientific paper [0] from late 2023 which cites a news article which quotes Alphabet's chairman as saying 'a large language model likely cost 10 times more than a standard keyword search, [though fine-tuning will help reduce the expense quickly]'. Editorialising the quote is not a good look for a scientific paper. The paper also cites a news letter from an analyst firm [1] that performs a back of the envelope calculation to estimate OpenAI's costs, looks at Google's revenue per search, and estimates how much it would cost Google to add an AI query for every Google search. Treating it like a Fermi Problem is reasonable I guess, you can get within an order of magnitude if your guesstimates are reasonable. The same analyst firm did a similar calculation [2] and came to the conclusion that training a dense 1T model costs $300m. It should be noted that GPT-4 cost 'more than $100m' and it has been leaked that it's a 1.8T MoE. LLama 3.1 405B was around 30M GPU hours, likely $30-60m. DeepSeek, a 671B MoE, was trained for around $5m. However, while this type of analysis is fine for a news letter, citing it to see how many additional servers Google would need to add an AI query to every search, taking the estimated power consumption of those servers, and deriving a 6.9–8.9 Wh figure per request for the amount of search queries Google receives is simply beyond my comprehension. I gave up trying to make sense of what this paper is doing, and this summary may be a tad unfair as a result. You can run the paper through Gemini if you would prefer an unbiased summary if you prefer :-).
The paper also cites another research paper [3] from late 2022 which estimates a dense 176b parameter model (comparable to GPT-3) uses 3.96 Wh per request. They derive this figure by running the model in the cloud. What a novel concept. Given the date of the paper, I wouldn't be surprised if they ran the model in the original BF16 weights, although I didn't check. I could see this coming down to 1 Wh per request when quantised to INT4 or similar, and with better caching/batched requests/utilisation/modern GPUs/etc I could see this getting pretty close to the often quoted [4, from 2009 mind] 0.3 Wh per Google search.
Google themselves [5] state the median Gemini text prompt uses 0.24 Wh.
I simply don't see where 100x is coming from. 10x is something I could believe if we're factoring in training resource consumption as some extremely dodgy napkin maths is leading me to believe a moderately successful 1T~ model gets amortised to 3 Wh per prompt which subjectively is pretty close to the 3x claim I've ended up defending. If we're going this route we'd have to include the total consumption for search too as I have no doubt Google simply took the running consumption divided by amount of searches. Add in failed models, determine how often either a Google search or AI query is successful, factor in how much utility the model providing the information provides as it's clearly no longer just about power efficiency, etc. There's a lot to criticise about GenAI but I really don't think Google searches being marginally more power efficient is one of them.
[0] https://www.sciencedirect.com/science/article/pii/S254243512...
[1] https://newsletter.semianalysis.com/p/the-inference-cost-of-...
[2] https://newsletter.semianalysis.com/p/the-ai-brick-wall-a-pr...
[3] https://arxiv.org/abs/2211.02001
[4] https://googleblog.blogspot.com/2009/01/powering-google-sear...
[5] https://cloud.google.com/blog/products/infrastructure/measur...
In reality what they do is pay "carbon credits" (money) to some random dude that takes the money and does nothing with it. The entire carbon credit economy is bullshit.
Very similar to how putting recyclables in a different color bin doesn't do shit for the environment in practice.
"Google deletes net-zero pledge from sustainability website"
as noticed by the Canadian National Observer
https://www.nationalobserver.com/2025/09/04/investigations/g...
If he is currently at Google: congratulations on this principled stance, he deserves a lot of respect.
> You spent your whole life breathing, and now you're complaining about SUVs? What a hypocrite.
You needed to read only conservative resources to not be aware that such criticism exists.
As if there isn't a massive pro AI hype train. I watched an nfl game for the first time in 5 years, and saw no less than 8 AI commercials. AI Is being forced on people.
In commercials people were using it to generate holiday cards for God sake. I can't imagine something more cold and impersonal. I don't want that garbage. Our time on earth is to short to wade through LLM slop text
It all stinks of resume-driven development
In windows, Co-polit is installed and its very difficult to remove.
Don't act like this isn't a problem, its a very simple premise.
I noticed a pattern after a while. We'd always have themed toys for the Happy Meals, sure, sometimes they'd be like ridiculously popular with people rolling through just to see what toys we had.
Sometimes, they wouldn't. But we'd still have the toys, and on top of that, we'd have themed menus and special items, usually around the same time as a huge marketing blitz on TV. Some movie would be everywhere for a week or two, then...poof!
Because the movies that needed that blitz were always trash. Just forgettable, mid, nothing movies.
When the studios knew they had a stinker, they'd push the marketing harder to drum up box office takings, cause they knew no one was gonna buy the DVD.
Good products speak for themselves. You advertise to let people know, sure, but you don't have to be obnoxious about it.
AI products almost all have that same desperate marketing as crappy mid-budget films do. They're the equivalent of "The Hobbit branded menus at Dennys". Because no one really gives a shit about AI. For people like my mom, AI is just a natural language Google search. That's all it's really good at for the average person.
The AI companies have to justify the insane money being blown on the insane gold rush land grab at silicon they can't even turn on. Desperation, "god this bet really needs to pay off".
And companies do force it.
You're breaking the expected behavior of something that performed flawlessly for 10+ years, all to deliver a worse, enshitified version of the search we had before.
For now I'm sticking to noai.duckduckgo.com
But I'm sure they'll rip that away eventually too. And then I'll have to run a god dang local search engine just to search without AI. I'll do it, but it's so disappointing.
Unless your version of reason is clinical. then yeah, point taken. Good luck living on that island where nothing else matters but technological progress for technology's sake alone.
I hate the way people get angry about what media and social media discourse prompts them to get angry about instead of thinking about it. It’s like right wingers raging about immigration when they’re really angry about rent and housing costs or low wages.
His anger is ineffective and misdirected because he fails to understand why this happened: economics and convenience.
It’s economics because software is expensive to produce and people only pay for it when it’s hosted. “Free” (both from open source and VC funded service dumping) killed personal computing by making it impossible to fund the creation of PC software. Piracy culture played a role too, though I think the former things had a larger impact.
It’s convenience because PC operating systems suck. Software being in the cloud means “I don’t have to fiddle with it.” The vast majority of people hate fiddling with IT and are happy to make that someone else’s problem. PC OSes and especially open source never understood this and never did the work to make their OSes much easier to use or to make software distribution and updating completely transparent and painless.
There’s more but that’s the gist of it.
That being said, Google is one of the companies that helped kill personal computing long before AI.
> I want no local storage anywhere near me other than maybe caches. No disks, no state, my world entirely in the network. Storage needs to be backed up and maintained, which should be someone else's problem, one I'm happy to pay to have them solve. Also, storage on one machine means that machine is different from another machine. At Bell Labs we worked in the Unix Room, which had a bunch of machines we called "terminals". Latterly these were mostly PCs, but the key point is that we didn't use their disks for anything except caching. The terminal was a computer but we didn't compute on it; computing was done in the computer center. The terminal, even though it had a nice color screen and mouse and network and all that, was just a portal to the real computers in the back. When I left work and went home, I could pick up where I left off, pretty much. My dream setup would drop the "pretty much" qualification from that.
In which case he’s got nothing to complain about, making this rant kind of silly.
Home Computer enthusiasts know better. Local storage is important to ownership and freedom.