Hacker Newsnew | past | comments | ask | show | jobs | submit | more avalys's commentslogin

Of course it was not unpreventable, though it might turn out that preventing it would have been unreasonably expensive.

But, the FAA inspectors are not responsible for making sure planes are safe to fly. They are responsible for making sure the people whose job that actually is, are doing their jobs effectively. That’s a critical difference.

It’s UPS maintenance personnel who are responsible for making sure that UPS planes are safe to fly. Yes, it’s possible that there is some institutional failure at UPS, that could have been caught if FAA inspectors were working in the past 30 days, but this isn’t the most likely scenario, and the root cause and responsibility (in this hypothetical) would still lie with UPS and not the FAA and the shutdown.


Every multi-engine airliner is designed to be able to take off safely even if an engine fails at a critical moment. What might have happened in this case is that the mechanism of failure of one engine caused damaged or interfered with the operation of another engine (via smoke, debris, etc.), and taking off with two engines degraded is not part of the design criteria.


I do think 'engine fails' and 'engine has left the building' are two different categories of problems. Even if the rear engine was working I'm going to assume this craft would have crashed, probably just farther down range.


Some engine failures can't be contained within the cowling, like turbine disc rupture. Probably something like this happened where fragments punctured the surrounding wing structure and/or fuel tanks.


Zero. This almost certainly has nothing to do with the shutdown.


There's just no way that's actually true though in a complex environment like airports and airplanes.


[flagged]


This is a silly take because having your ATC workers unpaid for over 30 days is going to increase the risk of catastrophic plane crashes. Even if this had nothing to do with this.

Footage of plane crashes are certainly important to know _this could start happening to passenger planes_


*"something like this", perhaps. Rather than this specifically.


“Almond milk” is not milk. You know what else is less carbon-intensive than milk? Candy corn. But that is also _not milk_, and so equally irrelevant!


Almond milk is an economic substitute for dairy milk, making the comparison appropriate. No need to be dense about it.


Almond milk is not dairy milk, but it is absolutely "milk", in the sense of a white liquid derived from plants - a definition that has existed in English for hundreds of years.

The name "almond milk" has been used since at least the 1500s.


I don’t drink ‘almond beverage’ but given the amount of uses it has substituting milk (and the amount of people that accept them) it seems like a very relevant comparison. Maybe I’m not sofisticated enough but I’m yet to see a candy corn mlik latte be ordered.


> I’m yet to see a candy corn mlik latte be ordered.

Not exactly the same, but can I interest you in a caramel-waffle-oat-milk latte?

https://mightydrinks.com/cdn/shop/files/Barista-CW.png


You certainly can sir!


But that's such a good idea though


> “Almond milk” is not milk. You know what else is less carbon-intensive than milk? Candy corn. But that is also _not milk_, and so equally irrelevant!

While almond milk is an incomplete substitute for bovine mammary secretions, it is so much closer than candy corn that it has been used as a substitute for the last 800 or so years, and shared the "milk"-ness in the name before we had an English language:

  The word “milk” has been used since around 1200 AD to refer to plant juices.
- https://www.ift.org/news-and-publications/food-technology-ma...

This makes this use of the word older than English people spelling the thing chickens lay as "eggs" rather than "eyren": https://www.youtube.com/watch?v=rZ5znvym68k

The Romans called lettuce "lactūca", derived from lac (“milk”), because of the milky fluid in its stalks: https://en.wiktionary.org/wiki/lettuce

Similar examples abound.

For example, I grew up in the UK, where a standard Christmas seasonal food is the "mince pie", which is filled with "mincemeat". While this can be (and traditionally was) done with meat derivatives, in practice those sold in my lifetime have been almost entirely vegetarian. The etymology being when "meat" was the broad concept of food in general: https://en.wikipedia.org/wiki/Mincemeat

Further examples of this: today we speak of the "flesh" and "skin" of an apple.

Personally, I don't like almond milk. But denying that something which got "milk" in its name due to it's use as a milk-like-thing, before our language evolved from cross-breeding medieval German with medieval French, to argue against someone who said "Insofar as it is used for milk", is a very small nit to be picking.


I mean I'm drinking a coffee with almond milk right this second.. which coincidentally replaced the dairy milk

Candy corn in my coffee wouldn't taste anywhere near as good


Is this water even “used” in the same sense that water is used for bathing or agriculture?

Where does the water go? If they simply take in cold water and release hot water, that water is still available for other uses.

Or do data centers use evaporative cooling just like power plants?


> Or do data centers use evaporative cooling just like power plants?

Yes. Not always, but evaporative cooling is much more energy efficient than heat exchange to outside air.

That said, stories about data center water use are a distraction from much bigger water consumers like golf courses and agriculture (e.g. to ship alfalfa to the middle east).


The problem with data center water usage is that it is unnecessary from the PR point of view. Data centers can run on air cooling just as good, but more expensive. For all I know, we could also do just as good without data centers, like we did 20 years ago.

With agriculture, water usage is necessary as eating is not something optional and everyone needs to eat to survive. From the PR point of view, of course. We couldn't live without agriculture, as we had agriculture 20 years ago too.

Golf courses are unrelated as they don't use nearly as much water as agriculture or data centers.

PR is everything, the narrative is what makes the difference. There is a lot of hypocrisy in this field, which is why I try to avoid it, but there is also some truth in it - we really didn't need that many data centers 20 years ago.


Golf course water usage vastly dwarfs data center water usage. Google used something like 1 billion gallons a year for their DCs. Single golf courses in arid regions can use upwards of a hundred million gallons a year, and in those areas there can be dozens of courses. It's not even close in terms of water usage.


Golf courses in the USA used about 2.1 billion gallons of water per day circa 2004 [1]. In other words, the annual usage of Amazon's datacenters per the article, 7.7 billion gallons, is less than the amount of water used on just American golf courses in four days.

[1] https://www.usga.org/content/dam/usga/pdf/Water%20Resource%2...


> Data centers can run on air cooling just as good, but more expensive

"More expensive" means spending more on air conditioning. Ergo more electricity used, higher electricity demand, more natural gas burned and carbon emissions, higher consumer power prices. So a different kind of PR disaster.


The difference in energy usage won't be noticeably higher for PR purposes. Of course, the difference comes at a price, cutting which is the main incentive for water usage.


OK, so we need to include the effect of introducing a green house gas, water vapor, into the environment.


> That said stories about data center water use are a distraction from much bigger water consumers

That's something of a fallacy of relative privation. When water is scarce, all frivolous uses should be under scrutiny. The others you mention have been well-known for a long time. The current stories simply highlight a new consumer people haven't thought of before.


I don't think it's a fallacy, it's much easier to optimize water usage for something that much larger.

Agriculture uses about 70% of all freshwater while datacenters are less than 0.5%

Some leaky channel will cost more than all the datacenters.


Also - Will data center water usage remain "negligible" if AI succeeds at wide adoption and scales to 100x current deployment? If 100x current usage levels become a concern, I don't know why people pretend that current usage is not a concern for a tech that many of those same people are projecting to scale.


>Also - Will data center water usage remain "negligible" if AI succeeds at wide adoption and scales to 100x current deployment?

Yes.


Even in case of evaporative cooling the water is not used. It's returned to the environment.


While technically true, if your datacenter is in Phoenix and you just consumed a few acre feet of water to raise the relative humidity by 0.000001%, for all intents and purposes that was a massive waste of water.


If the data-center depletes the water table used for surrounding wells, whether the "released water" is still "available" is irrelevant to those dependent on wells for water.

e.g. https://www.nytimes.com/2025/07/14/technology/meta-data-cent... https://www.bbc.com/news/articles/cy8gy7lv448o


> Where does the water go? If they simply take in cold water and release hot water, that water is still available for other uses.

Well, first you need to cool it down in a way that's good for the environment, I presume. You should not pour down hot water in a cold river with all its fishes and plants.


Datacenters can use a few different kinds of liquid cooling, including the one you describe. It varies quite a bit by geography.


Even if it's evaporative cooling, couldn't the water vapor just be... condensed back into water?


Yes - by cooling it. See the problem?


You don't have to use energy to cool it though do you?

Couldn't it just flow into a big passive outdoor radiator?


Maybe it’s confusing because it’s misnamed?


This is like saying the non-negative integers under addition, lists under append, and strings under concatenation are all just misnamings of the semigroup operator.

https://hackage.haskell.org/package/base-4.21.0.0/docs/Data-...


Is it? Two things are equal exactly when they aren’t exclusive.


This is exactly how it works in the US, too.

The reason this topic continually comes up is that people in the US are stupid and bad at math, and the IRS is very heavy-handed and issues penalties for minor tax errors, so people are afraid to interact with the process without a trusted intermediary.


Literally none of this is true.

The irs is neither heavy handed nor particularly quick to issue penalties.

There is an extremely effective and powerful alliance between certain republican politicians and tax industry corporations that work to convince people taxes are hard and the gov can't do them and they need an agent.

It works.


Okay, the official IRS policy is that you don't have to file taxes if you don't owe anything. What happens if you do not file your taxes, but the IRS believes you owe them money?


I mean, it depends if the amount is $10billion or $10, but generally they start by sending you a letter at wherever they think you live saying "hello, please write us a check for $x, thanks".

Then they do that... again. At some point they probably put your name on some kind of list of Bad Taxpayers but unless we're talking millions here they probably aren't sending agents after you in specific.


I mean, I don't file anything. For my car tax, I go to the site, enter my license plate, and a couple of other details, and the number shows up.

I enter my credit card number, and pay. That's all.

Same for other stuff like housing tax, too.


That's how it works in the US also, though personal property and real estate taxes are collected at the state and local level (if they exist, which is dependent on the state and local government).

For most people in the US, filing their taxes is a very simple process, which is why it's so annoying that Intuit has successfully lobbied to integrate themselves into the process.


Yes, and this includes the US!

When people say they are “paying their taxes”, really what they’re doing is checking whether the automatic tax deduction out of each paycheck was properly calculated over the whole year, and whether any special circumstances make them eligible for a refund (or whether they’ve had other income they need to pay tax on).


Yes, and what do you think it is like in the US? It works exactly the same way.


All the same criticisms are true about hiring humans. You don’t really know what they’re thinking, you don’t really know what their values and morals are, you can’t trust that they’ll never make a mistake, etc.


I think you're misreading the article; the point here is not "LLMs are bad and can't replace humans," the point is that many non-technical people have the expectation that LLMs can replace humans _but still behave like regular software_ with regard to reliability and operability.

When a CEO sees their customer chatbot call a customer a slur, they don't see "oh my chatbot runs on a stochastic model of human language and OpenAI can't guarantee that it will behave in an acceptable way 100% of the time", they see "ChatGPT called my customer a slur, why did you program it to do that?"


You can teach a human when they make a mistake. Can you do the same for an LLM?


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: