Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

How AI labs are solving the power problem

newsletter.semianalysis.com

Post reply on HN
Print view View original

251–260 of 270 posts

Re: How AI labs are solving the power problem

#251

Post by jasonwatkinspdx »

0xDEAFBEAD wrote: post #217
tialaramex wrote: post #206

Earlier quoted context omitted.

So the thing about "Ya'll should ..." is that it's often mistaking "Group A and Group B hold conflicting beliefs but share a characteristic" with "Group C, all the people with this characteristic, all exhibit an incoherent belief structure". For example suppose you apply this to the US Senate. So instead of Group A (Democrats and a those who caucus with them) and Group B (Republicans) we instead think there's a singl…

You might as well argue: "Part of my brain thought A at time Y. A different part of my brain had a different thought B, at a different time Z. Why the accusations of hypocrisy?" The problem arises when an individual or group tries to represent themselves as more credible/consistent/coherent than they really are. If you freely admit that you have multiple personality disorder, hypocrisy is to be expected from you as a…

You're making a sweeping bad faith representation of progressives yourself.

It can simply be the case that your accusations of hypocrisy are wrong, because your entire characterization of the discussion/position is wrong.

I'm not going to say "you made a good point" because you flatly have not.

Re: How AI labs are solving the power problem

#252

Post by all_factz »

Lerc wrote: post #249
all_factz wrote: post #247

Earlier quoted context omitted.

“I reject demonisation of the wealthy” is quite an odd thing for someone identifying as “far left” to say. But then you go on to identify as a “left liberal” - canonically not considered far left - so maybe I shouldn’t be surprised. Whether it’s worth demonizing anyone or not, we can condemn actions that hurt innocent people and we can maintain skepticism of the ultra-wealthy and their motives without “bullying”. It…

Therein lies the rub, when people are surprised to see a left wing person criticising the idea of othering, you have to wonder what principles they have left to call left wing. >we can condemn actions that hurt innocent people and we can maintain skepticism of the ultra-wealthy and their motives without “bullying”. Indeed and I do condemn actions. What I don't do is conddemn people. I am for robust regulation, free e…

Sure, I agree. Kneejerk condemnation and othering is bad.

But there’s a need to balance even-handedness with a healthy skepticism of those in power. Otherwise you risk becoming an apologist. No one is saying not to do your homework or not to think critically, but we’re also saying not to come in guns blazing in defense of moneyed interests. That’s what the person who brought up the hidden agenda stuff seemed to be doing - making assumptions that favor capital without even taking the time to read the article that addressed those assumptions. That’s not even-handed, it’s biased against labor.

Re: How AI labs are solving the power problem

#253

Post by badc0ffee »

immibis wrote: post #243
badc0ffee wrote: post #199

Earlier quoted context omitted.

I just looked, and you're right. It's in an industrial area several km from homes, and near existing NG facilities, including one where they flare gas. https://www.facebook.com/abacustrategic/posts/pfbid02rrUwoWM... https://maps.app.goo.gl/fYwcSi8vfPBnsYeK7 I don't doubt that it is a source of pollution, but I agree that this is overblown in the same was as the claims that datacentres are using up all the fresh water…

It's not overblown. That plant has sufficient pollution filters installed. You can find more info elsewhere in this comment section.

But it's still not localized to the one neighbourhood in the Time article any more than any other in Memphis (or even West Memphis across the river).

Re: How AI labs are solving the power problem

#254

Post by Arn_Thor »

renewiltord wrote: post #157
Arn_Thor wrote: post #151

Earlier quoted context omitted.

> It feels racist to expect people to assume a neighborhood is 'resource poor' just because it is 'historically black'. Statistically poverty is correlated with race. For reasons to do with (quite recent) history.

[flagged]

I don't know, what are we doing with these assumptions? Are we trying to lift disadvantaged communities by providing extra resources and help people get on a better footing in life, or harassing individuals on the street because they have a certain skin color? I'm very eager to support the former and protest the latter.

Re: How AI labs are solving the power problem

#256

Post by Arn_Thor »

mucle6 wrote: post #155
Arn_Thor wrote: post #151

Earlier quoted context omitted.

> It feels racist to expect people to assume a neighborhood is 'resource poor' just because it is 'historically black'. Statistically poverty is correlated with race. For reasons to do with (quite recent) history.

Statistics are not a license to assume. Crime rates also statistically correlate with demographics, but if I assume a specific person is a criminal based on that stat, I would (rightly) be called racist. Expecting people to assume 'historically black' == 'poor' similarly feels racist.

The question is how you check, qualify and--last but not least--apply the statistical findings. Are we trying to lift disadvantaged communities by providing extra resources and help people get on a better footing in life, or harassing individuals on the street because they have a certain skin color? I'm very eager to support the former and protest the latter.

Re: How AI labs are solving the power problem

#257

Post by skywhopper »

badc0ffee wrote: post #124
skywhopper wrote: post #119

Earlier quoted context omitted.

Gas furnaces and stoves are known polluters of indoor air: https://www.psehealthyenergy.org/gas-stoves-and-indoor-air-p... Large gas plants are probably relatively clean overall, but the temporary, portable gas generators used by eg the xAI datacenter are not as tightly regulated and aren’t inspected or controlled in the same way. Given the particular corporate agent involved, I’d be surprised if any care at all were…

That is true of gas stoves, but gas furnaces don't exhaust into the house.

But gas turbines do vent to their neighbors’ air, which is the main point here.

Re: How AI labs are solving the power problem

#258

Post by Lerc »

all_factz wrote: post #252
Lerc wrote: post #249

Earlier quoted context omitted.

Therein lies the rub, when people are surprised to see a left wing person criticising the idea of othering, you have to wonder what principles they have left to call left wing. >we can condemn actions that hurt innocent people and we can maintain skepticism of the ultra-wealthy and their motives without “bullying”. Indeed and I do condemn actions. What I don't do is conddemn people. I am for robust regulation, free e…

Sure, I agree. Kneejerk condemnation and othering is bad. But there’s a need to balance even-handedness with a healthy skepticism of those in power. Otherwise you risk becoming an apologist. No one is saying not to do your homework or not to think critically, but we’re also saying not to come in guns blazing in defense of moneyed interests. That’s what the person who brought up the hidden agenda stuff seemed to be do…

Consider the original post I responded to. It asked two questions.

>are you skeptical xai would wiggle around regulations and pollute a city?

But they were responding to I'm incredibly skeptical of any claim that xAI's power use is putting a dent in the local environment which makes no claim as to whether they might obey or disobey regulations, the words "putting a dent" in the local envionnent*

The data from the article does not sufficiently address this, it uses satellite data and a short time frame. Without specifying the resolution of their data (which could be kilometer sized pixels) their claims about locality is in doubt. In short term measures, trends are harder to spot, a rise over months could just mean it is less windy in the nollowing season. Without a ground level meadurement of the air quality and a evaluation of the total local emission from all sources, you cannot hope to measure the health impact of a single cause.

None of that says that they are not polluting. What it says is that this is not evidence of it. Someone expressed skepticism based on the proportional emission of one of many of their ability to move the dial, and was challenged based upon the likelihood of what they might do. Claiming skeptasism that a thief could rob Fort Knox, is not a claim that the thief is honest

>by “agenda pushing” do you mean those who have an agenda to have breathable air?

I simply cannot believe that this is a reasonable interpretation of what they said.

That's the thing that motivated me to post on this thread. That the first post I responded to here was attacking the player, not the ball.

I continue to post replies here out of my own sense of duty to fully explain my position to promote understanding, I'm not trying to win anything here, I only want people to see a honestly held perspective.

Re: How AI labs are solving the power problem

#259

Post by Xss3 »

goku12 wrote: post #226
ACCount37 wrote: post #176

Earlier quoted context omitted.

These "gas turbines" are located next door to the Allen Combined Cycle Plant, a grid scale natural gas power plant with 1.1GW capacity. It's there to power a nearby steel mill. That's the kind of neighborhood xAI has put its cluster in. I'm incredibly skeptical of any claim that xAI's power use is putting a dent in the local environment, and "environmental racism" just reeks of the usual agenda pushing.

Here are some quotes from an article [1] that directly addresses your point: > The turbines spew nitrogen oxides, also known as NOx, at an estimated rate of 1,200 to 2,000 tons a year — far more than the gas-fired power plant across the street or the oil refinery down the road. > The turbines are only temporary and don’t require federal permits for their emissions of NOx and other hazardous air pollutants like formal…

Thank you for taking the time to respond.

A thousand bots can spread a thousand lies faster than we can dispell them, but giving up the public square to them is worse.

What a situation...

Re: How AI labs are solving the power problem

#260

Post by Xss3 »

black_puppydog wrote: post #234
rossjudson wrote: post #227

Earlier quoted context omitted.

Once, when I was a child, I remember the carefully engineered smoke stacks in Sudbury Ontario spilling out acrid smoke sideways, and then straight down into the town. Back then, Sudbury looked like Mars.

So you're saying when not done properly, emissions have very direct negative impact on the surrounding community. Did I get that right?

No theyre saying that since that day everyone has given up and nothing matters anymore. We all collectively decided that it is OKAY and didnt change a single thing since.
Post reply on HN