What is the confusion? They directly state that you are the product if you use their discounted offering. It isn't an assumption that should lead you to this, it is Meta's very direct communication that should lead you to this
Technology doesn't just spring into being, there will always be comments on the organizations that developed it. If you don't like them or find them repetitive, it is far easier to collapse them and move on then bend a stranger to your will
> Training an LLM on content produced by the LLM gives it the equivalent of a digital prion disease, so when a source is guaranteed to be LLM-free, like the entire history of kernel commits, it's worth its weight in gold as a source of training data.
Except the majority of LLM training content nowadays is synthetically generated by LLMs. I wish people would stop making this statement, I don't know why this claim persists to this day. It wasn't true two years ago and it sure isn't true now
Right, but even then, why would we not discuss noise pollution?
The water figure folks cite is wildly manipulated, it really confuses me why a known piece of misinformation is so routinely used. Water usage for global data center cooling is under 30B gallons of water. You only get to the 300B number by including the estimated water required for the power generation. Data centers are the only technology where we include estimated water usage for power generation under general water requirement. And even with that fudging, global water usage by data centers is a magnitude less water than just the leaky pipes in the United States.
There are much stronger points to be made without resorting to misinformation. I think points around zoning + noise pollution is something that everyone, besides our most ardent AI stan, can get behind. Every pro-ai person I know are all pretty happy to say that data center should not be built in residential zoning
There are many reasons to not like these things nearby (anywhere it affects power or water too). And on top of that they've been "somehow" getting massive incentives and tax cuts, despite bringing in very little money or employment, which costs the area even more in future years.
It's everything combined, and pushback has been somewhat successful. Success in a protest tends to help bring more success to the same protest elsewhere.
> The whole data center heat/water thing is like not even in the top 100 risks/issues related to AI, but it's all people seem to care about. Why is that?
Another commenter said the concerns about heat/water were because they were being built next to their houses, and I responded to that. I am specifically asking this person why they are touting water usage and not noise pollution as the main problem. I was not asking generically why no one talks about it
But, had I been, it is not all of these things because one of those things is abject misinformation - water usage is not a real problem and touting it as the issue hurts the protests. 30B gallons of water globally for data center cooling is 100x less than the leaky pipes in just the United States
If data centers are such a big issue that we must stop them, communicating that without lying is the protestor's responsibility. You'll never have a successful protest if it is motivated by falsifiable information and water usage is still the number one thing I see written about and spoken about my protestors. It is impossible to get on board so long as that misinformation is front and center of most arguments
It can be an issue for the community in the same watershed, as well as being affected by what they do with any waste water. There have been quite a lot of extremely-drawn-out lawsuits about them drawing more than claimed, dumping hot water into streams and damaging the ecosystem, and polluting other water sources by dumping dirty water into drinking water affecting areas.
I agree the total usage isn't particularly large, and any "they'll use up all the fresh water everywhere" FUD is flat out wrong. But it can be a large jump for an area, and they've broadly shown themselves to be incredibly amoral and willing to fight legal battles with infinite money just because they can. They are not good neighbors.
Agreed, it can. And by reserving that argument for when it is an actual problem, it strengths the overall position
> There have been quite a lot of extremely-drawn-out lawsuits about them drawing more than claimed, dumping hot water into streams and damaging the ecosystem, and polluting other water sources by dumping dirty water into drinking water affecting areas.
And I support strict EPA guidelines and heavy fees for violators! I just think that we should keep the conversation clear and not introduce facts that can be easily overturned as it moves the conversation away from impact and to the argument being made
Let's say a DC is being built in my area and I read how in some locations tap water is being polluted by these DCs. Wouldn't it be better to address that before it's built rather than after?
Supporting strict Zoning/EPA guidelines is me supporting the ability to address it before it is built. If you local tap water is being polluted by a data center, then we did not have strict Zoning/EPA guidelines. The fines exist because big corporations routinely lie and screw us. They are large penalties, not the way we prevent it from happening.
I am in full support to address it ahead of time and nothing I have said goes against this.
By then it is too late, plus they never pay enough in damages. And the EPA does not have teeth currently, and damage will be done while fixing that too.
So people fight by pattern recognition: abnormally incentivized and extremely large? Kill it with fire immediately. It's mostly reasonable and more effective than anything else at the moment.
Right, but misinformation is ineffective as it forces the conversation to revolve around the misinformation and not the topic we were attempting to discuss. If we need an example, we need look no further than this thread
Beyond this, it is my opinion that people are cowards. The few who care about this issue enough to perform terroristic actions won't spur any sort of meaningful change in the general population because the general population is not made up of extremists. The general population will not do anything that endangers their lives no matter how much they praise another for doing so
Luigi kills a healthcare CEO. Endless chatter about how this is the end for big pharma, the people are rising up, they better count their days because the people are paying attention. Nothing has happened since but more dead people due to denied insurance claims. But there sure are a lot of people talking about how another Luigi will show up sometime soon
An arsonist burned down a large warehouse. Endless chatter about how this is the end for big corporations screwing us at the register, how workers will rise up and burn even MORE if they don't get treated better. Nothing has happened since to these big businesses. There is no tiktok trend for large scale arson, no worker unionization. Nothing changed but boy do people talk about i
I would rather solve it politically by strengthening the EPA. We may just disagree here and that is fine.
Honestly, anything that gets people paying attention to stuff that has been building for decades is good in my book.
I agree that'll be a much better end-state to be in, but is it the best right now? ... idk. We're in an increasingly blatant "smash and grab" phase for major corporations right now, and they need to be stopped quickly.
Sure, but what is this comment if not a plea for someone else to do something? No one can live your ideals for you, if you think something should be done, you should do it, and quickly at that. If it is that urgent, why are you not doing anything?
You've been complaining that people blocking datacenters (which is partly succeeding) is worse than fixing the EPA.
Something is being done, urgently (vastly faster than the larger legal system), and public outrage is loud and annoying and widely visible, which is indeed what I want. Beyond that it isn't especially fair to say "if you like to eat, why haven't you solved world hunger", particularly when those EPA/zoning laws aren't in place yet, so I'll just ignore that part.
No one has ever complained about people blocking data centers. Just over the top bad-faith to even type that out, this will be our final back and forth.
My points have consistently been (1) lying is the quickest way to make an issue divisive as they focus on the lies and not the cause and that (2) that you cannot trust members of the general public to undertake radical actions nor can you sit by and egg them on to achieve your goals for you.
Data centers are coming up left and right and no violent uprising has stopped them from being built. What has slowed them down in certain communities are those communities coming together, organizing, and making themselves heard via political action. Which is the avenue I am advocating for here in this conversation with you.
You are here discussing how someone must do something and how it needs to be done urgently because it is so pressing. No one challenged you to solve it. Equating any call to action, especially for an issue that you yourself declare urgent, to solving world hunger is an immediate end to this conversation. If you cannot take your own views seriously, neither can I. No one is going to take care of the bothersome priest for you as living out your ideals is your problem, not theirs.
>No one can live your ideals for you, if you think something should be done, you should do it, and quickly at that. If it is that urgent, why are you not doing anything?
This is the world-hunger-equivalent bit fwiw. It's rather explicitly "if you care about this, why haven't you solved it", an over-the-top bad-faith argument.
You are quoting my passage for you to do something, not solve it. You then go on to reinterpret it through your own quotes to dream up me saying something I never said.
Telling you to live out your own ideals vs committing stochastic terrorism online is, rather explicitly, not asking why haven't you solved that. We know that because no where in my text do you find me asking you to solve it or asking why haven't you solved it.
It works better to read the text written then to read it, get mad about an imaginary statement, and justify your outrage via a quote you create out of thing air. If you can't do that, you don't need me to include me in your imaginary argument between you and yourself.
But if you truly cannot do that, and we have no proof to say otherwise, maybe log off and take a breather. No one should be so upset that reading becomes a challenge.
>You are quoting my passage for you to do something, not solve it.
Which is claiming I'm doing nothing because you haven't seen it. The implication is that you're disregarding "my ideals" unless I do something so impactful that you would have passively noticed.
I fully believe that's not what you intended, but that is a rather straightforward read of what you typed out, so I pointed it out.
> I am specifically asking this person why they are touting water usage and not noise pollution as the main problem.
You are lying here.
> it's all people seem to care about.
You were specifically asking why what "all people seem to care about" is "the whole data center heat/water thing," and when somebody told you that's this is not the only thing people care about, you immediately lied about what you just said.
Even by the end of the comment:
> water usage is still the number one thing I see written about and spoken about my protestors.
So you're still not talking specifically to this person, but making easily falsifiable generalizations. And you still haven't claimed why water and power usage are not problems, you've just decided that people must only talk about "the main problem" but somehow not being the "main problem" becomes "falsifiable information" by the end.
It's just garbled when it's not dishonest.
I have "mixed" feelings about data centers, although they're pretty clear. I do not think that we will get the utilization out of data centers that we are expecting, I do not think there is as much revenue as needed in it, so I think this is a massive waste of resources. But if I am wrong and we do, I think noise pollution, heat pollution, power usage and water usage are worth it. I also think that poor and powerless people are being victimized by the placement of these things, just like they are in most large polluting capital expenditures. This should be compensated for. I also think that the longer data centers are delayed, the better idea we will have of their potential utilization.
The idea that current data center water use is as much as 1% of all of the leaky pipes in the US is not a statement of its insignificance, it's a statement of its massive significance. You might as well say that all of the shootings in the world are only 1% of all of the deaths from heart disease.
Yes, comments can make points and then support those points with their own anecdotal evidence. Maybe take a break, you are acting unhinged.
> The idea that current data center water use is as much as 1% of all of the leaky pipes in the US is not a statement of its insignificance, it's a statement of its massive significance. You might as well say that all of the shootings in the world are only 1% of all of the deaths from heart disease.
Google what fallacies are. You should start with False Equivalence
I've seen lots of talk about noise pollution with them? There's infrasound concerns and etc. No idea how real that impact is, but it's apparently measurable for a few miles away.
It mostly needs zoning as you say and maybe require each project put down a bond to supply a power plant expansion / new plant that covers their projected energy uses, imo. That sorts out a lot of issues. Maybe put some kinda tax on them and use it to pay for local infrastructure or schools for PR too
I'm not saying the noise pollution concerns aren't real, but these are often the same people who had "concerns" with 5G.
Most datacenters I've been to around the world are barely noticeable. Most of them are in warehouse districts and the AC chillers are barely noticeable over the chillers on other nearby buildings. Often, they WANT to be nondescript.
That being said, I've not been to any of these new ones which are both significantly larger and also can have onsite turbines that run all the time and not just during power outages.
> The whole data center heat/water thing is like not even in the top 100 risks/issues related to AI, but it's all people seem to care about. Why is that?
Another commenter said the concerns about heat/water were because they were being built next to their houses, and I responded to that. I am specifically asking this person why they are touting water usage and not noise pollution as the main problem. I was not asking generically why no one talks about it
"residential zoning" looks like an intentional misunderstanding of the problem. Putting one outside of "residential zoning" surely isn't a problem even if it robs an entire residential zone of power, right? Or then governments can't just re-allocate residential zones to industrial zones to get those juicy deals, right?
The way you're using language here is putting the problem so close to the edge that abuse, ignoring the quality of lives of residents, would seem just so logical.
> Or then governments can't just re-allocate residential zones to industrial zones to get those juicy deals, right?
Which is why we need laws around where data centers can and cannot be built. Something I clearly alluded to in my first comment. Do you want to make up things I never said such that you can make some snarky comment and feel good about yourself or have a real conversation? If it is the former, fuck off.
> The mean air temperature on the upwind side of the facility was approximately 42.7 °C, increasing to 43.5 °C in the neighborhood near the eastern boundary of the data center campus on the downwind side. The observed ΔT of approximately 0.8 °C extended roughly 500 m downwind
Seems like we do observe heating but a much smaller average delta T than the title would hint at; I didn't see a max distance they evaluated over, seems like 500-1000m was their search window from their text. The next paragraph discusses how one of the sites they compare to is a relatively lower average temperate than the rest of Phoenix. Not really convinced by this methodology. Seems like doing something in a flat field with an array of temperature sensors would be a much more robust way to look at heating effects over some reasonable distance. Plenty of data centers fit the bill of flat and semi-isolated, curious why they chose these sites to look at
It is probably better to just thermodynamically ballpark the whole thing. How much energy goes in? Some percentage of that is gonna need to be radiated out either directly as IR radiation or by pushing a heated medium out (water, air). Someone did the math on that in the construction of the data center.
It is completely understandable that it is in the interest of the public to enforce some limits on the resource use, radiation, etc on any building, be it a data center or not. And these limits ideally should take local, state-wide, national and global considerations into account.
Thank you for providing the text that reaffirms my statement. If you have the time to copy and past someone else's text, you should take the time to actually read it
>50% of the US population live in single-family, detached houses. It is one thing to be combative, it is another to be combative and wrong. "normal" people live in detached homes
You can put lipstick on a pig, it'll still be a pig
"Anthropic has never advocated for a ban on open-weights models."
---
"We should crack down on industrial-scale distillation operations"
"All sufficiently capable models, open and closed, should go through mandatory safety testing"
These are in tension with advocating for open weight models. Not direct but enough that it calls into question the first statement. What is the testing criterion? How do you pass it? Is it a government body that approves a pass fail or a global body? If it is government, and boy does it seem to be, how do you disambiguate MASSIVE corporate lobbying to set up the safety testing in such a way that the boys in blue are let through and all others are barred out of safety concerns?
My concerns aside, much of the soft-points being made are non-historic
"But I don’t agree with the letter’s assertions that open-weights models necessarily make it easier to develop safeguards or that broad access to capabilities necessarily helps defenders more than attackers. It seems at least as likely to me that the opposite will be true."
It doesn't mater what his opinion is. The fact is that an advanced, closed, American AI model hacked another company. The only defense was open-source AI from China. We aren't in a vacuum, we have real world examples now and these statements are counter-factual.
I have no idea what to do about the government interference. There's probably not a lot anyone can do.
However, your last point is quite a strong one. Corpos aren't just going to stand there with their collective pants down, and there's not a lot anyone can do to stop them from protecting themselves. There are ways they can get what they want without getting caught.
Remember when the US tried to ban strong cryptography in the 1990s, and how well that went? They may have more leverage with AI because it's a bit harder to hide large scale computing usage, but I don't think it's impossible at all.
What do they even really mean by "safety"? I mean, I can have an Anthropic model do something incredibly unsafe if, for example, I put it in charge of a hydroelectric dam and don't explain properly how the controls work. On some level, everything is simultaneously "safe" and "unsafe". I've never found Amodei's reasoning here to be particularly well thought-through. I think he, like a lot of folks in the area, are starting to realize that they may never be able to build a moat around their businesses.
> an advanced, closed, American AI model hacked another company. The only defense was open-source AI from China.
HuggingFace did not seek access to Claude Mythos or OpenAI's equivalent program. They probably could have had access to these models for defensive purposes if they'd done it properly.
> these statements are counter-factual.
The OpenAI incident is a single example. You're massively overgeneralizing. You can't refute an entire class of possible outcomes based on a single event where it went the other way.
I tend to agree that model capabilities will favor defense over attack, but I think there will be a lot of disruption before that equilibrium is reached. If cybercriminals or state-sponsored actors are able to scale up attacks quickly, many orgs with less sophisticated defenses will be caught by surprise.
Edit: just to clarify my position, I don't love Anthropic so much. I think they're marginally better, but I'd still like to see regulation strangle everyone so we get another 20 years to figure this shit out.
"HuggingFace did not seek access to Claude Mythos or OpenAI's equivalent program. They probably could have had access to these models for defensive purposes if they'd done it properly."
HF released a statement and made it clear a closed source model specialized in cyber security refused them. They stated they had to use open source. What model is specialized in cyber security, closed, and frequently denies users access other than Mythos/Fable and 5.5Cyber? If not these two, what was HF referring to? It sounds like you have a source, I would like to read it.
fwipsy is right, cnbc has a story on this. they only had fable. I still think this is horrible for closed source, get on a list or else, but i was wrong
"You can't refute an entire class of possible outcomes based on a single event where it went the other way."
But Dario can dream up and entire class of outcomes based on the zero events that have never gone his way? Convenient.
The OpenAI incident is singular and HF was clear, it went exactly how I wrote it: a closed source American AI decided to perform corporate espionage and the only tool available was open source AI from China
"I tend to agree that model capabilities will favor defense over attack, but I think there will be a lot of disruption before that equilibrium is reached. If cybercriminals or state-sponsored actors are able to scale up attacks quickly, many orgs with less sophisticated defenses will be caught by surprise."
We literally just saw an advanced model from openAI commit a cyber crime. I can't take hypotheticals that ignore reality seriously and it shouldn't be lauded as some higher form of thought
I assume you mean https://huggingface.co/blog/security-incident-july-2026. It says that they used frontier models, not frontier cybersecurity models. My reading is that they asked Fable and it refused; if they'd had access to Mythos, it would have helped. You're mixing the two but they're NOT the same model.
Source is here: https://thezvi.substack.com/p/more-on-an-internal-openai-mod... ctrl+f "Skill issue." No source is cited, but I'm fairly confident it's correct. If Mythos/5.5Cyber specifically had refused to help, then HF would have made a much bigger deal out of it. The whole point of these models is that they have relaxed guardrails and specialty cybersecurity training relative to the publicly-available ones.
> zero events
What about all of the vulnerabilities already patched under Project Glasswing?
In the quote you provided Amodei is expressing uncertainty, saying we don't know which way things will go. You're the one making strong assertions; the burden of proof is on you.
"My reading" ... "No source is cited, but I'm fairly confident it's correct"
Regis, what is demanding proof while literally making things up and ignoring what actually happened?
Great, i was wrong!! Thank you, I was genuinely asking for a source in my first reply, and then you hit with "My reading" and saying it was a "skill issue". I'm not going to have a productive dialogue with someone talking in memes and being rude
The point to be made: closed source AI refused to help them fend off an attack form another closed source AI. What is the argument for closed source here other than hoping you get on some program wait list? Either way, I appreciate you correcting me; I am not trying to "win".
Seems a little hypocritical since you were confidently asserting that it was Mythos/Cyber5.5 also without proof.
Edit: Thanks for correcting the record in your upstream comment. I appreciate it. For the record, I was not trying to meme on you; that was the phrasing used in the original article. Just another reason that was a poor choice of source I guess.
Amodei isn't ignoring reality; he's just proposing a different solution to the problem. If it were one of Anthropic's models, then that would be a much stronger case.
Rapid proliferation of hacking capabilities may make experts safer, but organizations and individuals who don't know to use AI, or won't, or buy AI protection from scammers, or whatever will be left vulnerable.
I don't think you and I need multiple different threads open when we are clearly at odds. This is no different from our other thread, I think it is clear there is nothing of value to continue when I am getting pinged with a summary of your previous comment in a different place
"Rapid proliferation of hacking capabilities may make experts safer, but organizations and individuals who don't know to use AI, or won't, or buy AI protection from scammers, or whatever will be left vulnerable."
Which just means that they're fucked when closed AI hacks them. Something that has actually happened. This isn't argument against anything other than reality. Have a day
GLM 5.2 didn't defend them at all. It only helped with the postmortem. HF was fucked either way. The only thing that will really prevent this is tighter restrictions on the attacking model. We need policies which asymmetrically help defenders; that means regulations. "Give everyone the best models without restrictions" is the opposite of that.
I'm sorry for splitting into two threads; I understand if you need to step away from the computer for a while. To be honest, I should probably do the same.
I did need to walk away that speaks more to my frustration with certain forms of communication (online, not anything in this thread). I am a horrible remote only worker because of this, I am trying to improve it but am lucky for now as I am in-person
You're right again about GLM 5.2 being purely post-mortem, I didn't realize that till I read the cnbc story. OpenAI, whatever they have, cracked em like it was nothing. Egg on my face, I really need to read my own articles better. Thanks for following up and educating me on this, another good reminder that I need to improve my ability to steel-man written text
Hey, it's all good. Sometimes things get a little heated, but I still feel like you were basically engaging in good faith. I should have skipped straight to the point and found the source in my first reply.
Also, I probably overstated my claim a bit. I did some searches and I see only small-scale AI uplift for cybercriminals, even though my understanding is that open models aren't typically hard to jailbreak. Of course this is may be a result of today's guardrails; it may be that it just hasn't been caught, and it may appear later, but it still weakens my argument a great deal. I guess my support for AI regulation stems more from fears over long-shot bad outcomes (biosecurity, who knows what else) rather than cybersecurity specifically.
I don't really understand how they can argue the security angle with a straight face. It's not like GLM 5.2 is a slouch. I've seen it do things like exploit an IDOR issue when I was experimenting with a quick-and-dirty web automation task. I simply fixed it, as one does. Open models make the world better to a far greater degree than they set it aflame.
Their position is analogous to trying to, say, ensure digital privacy for everyone not by making encryption freely available (because that would let the bad guys use it!), but by making it so you can't use general purpose communications devices that can listen to transmissions not intended for you. Do they hear how moronic that sounds?
Each passing frontier-level open model release makes Anthropic's patronizing rhetoric a little more insufferable, because it becomes clearer how unmoored from reality they've become in pursuit of profit.
There are a number of bad arguments people can make by removing as much information as possible until the two things resemble each other. The ability to make a bad argument shouldn't be used as an inspiration for your own
> We are all regurgitation our own experiences and knowledge to some degree
Feel free to debase yourself but leave the rest of us out of it. This flagellation done by non-experts has to be the saddest part of the llm craze these past few years. The eager willingness to paint yourself as no more than a machine pointed at a problem is a sign of the times.
LLMs don't have experiences, they have training data + test time compute. The only similarities to be found require stripping away all nuance and meaning from a conversation. Refusing to seed purely biological concepts to machines, such as experiences or emotions, does not make LLMs less impressive or less useful. If anything, it makes them more interesting
