Drewbahr's recent activity
-
Comment on Was this English queen a 'predatory' collector of beautiful things? in ~humanities.history
-
Comment on What is a business/org that's so terrible no one should use if possible? in ~life
Drewbahr LinkFacebook, what with the enabling fascism (e.g., Cambridge Analytica) and the enabling genocide (e.g., Myanmar) and the rampant misuse of AI (e.g., fake profiles and faking AI chatbots of...Facebook, what with the enabling fascism (e.g., Cambridge Analytica) and the enabling genocide (e.g., Myanmar) and the rampant misuse of AI (e.g., fake profiles and faking AI chatbots of celebrities) and the fact that Mark Zuckerberg is buying up huge swathes of Hawaii specifically to make a "fuck you I got mine" bunker.
-
Comment on Data centers don't raise people's water bills in ~tech
Drewbahr Link ParentEvaporation would almost certainly never reach "true" equilibrium for all of the reasons I mentioned in my previous, long post. Vapor will find ways to get out of the system, which will cause more...Evaporation would almost certainly never reach "true" equilibrium for all of the reasons I mentioned in my previous, long post. Vapor will find ways to get out of the system, which will cause more vapor to evolve out of the liquid (see: vapor pressure). A system that allows for evaporation to the atmosphere is inherently not closed-loop, because it is losing material to the air at all times. That is, in fact, how arguably the most popular form of open-loop cooling works (evaporative cooling).
Evaporation within a closed loop needs to be limited either by air bladder/nitrogen blankets with some form of water reclamation system, or by additional cooling being applied to inhibit the formation of vapors in the first place. Either way, you're adding or removing energy in some form that needs to be accounted for. It's not impossible, it's just complex.
Regarding the datacenter cooling you mentioned previously ... I'd have more of an opinion if I actually knew what they were doing. The first article - "Sustainable by design: Next-generation datacenters consume zero water for cooling" - does not describe what they are actually doing, only that they are using less water (via a chart with no sources listed). That article reads more like an advertisement than an article describing actual measures being taken to reduce water usage.
The other article/page - "Reducing Data Centre Environmental Impact with Waterless Cooling" - only sparingly mentions ways that they would reduce water consumption, via air coolers with "integrated compressors and condensers". I have a good idea about what they mean - it's similar to the lube and seal oil systems used by machines all over the world, which are "closed loops" that do not require the constant addition of oil. What I can say is, those systems suffer from all of the issues I mentioned previously - machines break down over time, corrosive elements build up, etc etc.
In particular, the second article mentions that "Little maintenance is required for the closed loop, simply the occasional addition of water treatment to control oxidation and bacteria growth." If they're controlling oxidation, that means they're controlling rust - which likely means the addition of acids or bases to scavenge oxygen and other oxidizers that accumulate over time. Adding acids or bases means adding corrosive chemicals to your water supply; they need to be controlled with a gentle touch, and the physical orientation of the piping is going to matter a lot.
As an example of this, a facility that I worked for, years ago, had an acid vapor recovery system that needed to be replaced. They selected a steel alloy designed specifically to withstand the kind of corrosion they were experiencing in the existing piping. The problem was, their pipe racks - the supporting system keeping all of the pipes elevated off the ground, so people could work in the area - were quite crowded. Of course, the obvious answer would be to expand the pipe racks; however, that costs a lot of money and requires a lot of infrastructure to be built. So instead, they opted to remove the corroded piping and install the new alloy piping in its place.
Problem was, the piping needed to be installed as free-draining - meaning, it had to have a slope back to the acid tanks, so that any liquid that precipitated or condensed would automatically drain back into the acid tanks via gravity. That way, nothing acidic would accumulate in the piping itself, and only the vapor would flow out into the recovery system. There was one little stretch of piping that couldn't quite meet the typical 1" per 100ft of slope required for free-draining; it had to be installed dead-level. Normally, a little stretch of dead-level piping wouldn't be an issue - there's enough vapor traffic and liquid drainage that it would have no real negative effect.
So, the engineers carefully measured everything and got the piping installed. Unfortunately, it didn't quite get installed dead-level; it had a very, very slight slope away from the acid tanks, rather than toward them. The result? There was a slight accumulation of acidic precipitate in the acid gas recovery piping, and the piping - built from an alloy specifically intended to withstand the kinds of corrosion it was going to experience - suffered through-wall corrosion in less than 6 months. The whole effort had to be done again with a full replacement of the already-replaced piping, which effectively doubled the cost of the project.
Now, the kinds of acidic liquid that system was dealing with were way, way more severe than what a cooling system like we're discussing here would experience. I mention that example, though, as a way to point out that you can design a system, on paper, to be perfect for its intended use - but physical reality and constraints can conspire to ruin it quite quickly.
-
Comment on Why do so many people think US President Donald Trump is good? in ~society
Drewbahr (edited )Link ParentWhite is a spectrum. You're not white until you are. Greek immigrants weren't considered white, until they were. Same with Italian, same with Irish, same with any other myriad ethnic groups. Hell,...White is a spectrum. You're not white until you are.
Greek immigrants weren't considered white, until they were. Same with Italian, same with Irish, same with any other myriad ethnic groups.
Hell, most census forms I've seen ask you, separately, if you are white and if you are Hispanic - because according to our government, you can be Hispanic AND white.
-
Comment on Why do so many people think US President Donald Trump is good? in ~society
Drewbahr (edited )Link ParentOh, Trump supporters can be, and often are, very sexist too. They contain multitudes. Am I being reductive? Maybe, but I certainly don't think so. Trump is a fascist, everything he's said and done...Oh, Trump supporters can be, and often are, very sexist too. They contain multitudes.
Am I being reductive? Maybe, but I certainly don't think so. Trump is a fascist, everything he's said and done indicates he's a fascist, he's openly calling himself a dictator, and people still voted for him. He's been openly bigoted for decades, and at least pedophile adjacent for that long as well.
What does that say, other than that his supporters actually support him and what he says and does?
EDIT TO ADD: I should also add, pointing to the 15ish percent of Black Americans that voted for Trump as some sort of sign that Trump voters aren't racist, ignores the fact that he won the majority of white votes by a significant margin.
Again, combine that fact with his rhetoric, and tell me that race being a significant factor (in my opinion, THE significant factor) isn't accurate.
-
Comment on Why do so many people think US President Donald Trump is good? in ~society
Drewbahr LinkBecause there's a lot of racists in this country, whether they know it or not.Because there's a lot of racists in this country, whether they know it or not.
-
Comment on Data centers don't raise people's water bills in ~tech
Drewbahr Link ParentI'm not sure why you're taking my argument here personally, because I assure you it's not. I'm not intending to offend you; I'm just talking about physics here. My initial comment is talking about...- Exemplary
I'm not sure why you're taking my argument here personally, because I assure you it's not. I'm not intending to offend you; I'm just talking about physics here.
My initial comment is talking about closed-loop systems - ones where you don't need a continuous inflow/outflow of coolant. Modern air conditioning systems are "closed loop" in that sense, as are refrigerators and other common appliances. I'm a chemical engineer, and I've worked in industry for nearly 20 years at this point. My education and professional experience both lean quite heavily on mass and energy balance - if there's one thing a chemical engineer needs to understand, it's how energy and material moves in systems.
So when I talk about things like "evaporation loss", it's not random. Evaporation loss does happen in closed systems. Anytime you're running a coolant or refrigeration cycle, you're dealing with thermal expansion - and if you have thermal expansion, you need to have vapor space to permit the contained coolant to expand and contract. This is usually something like a reservoir, a tank, or some other "wide spot in the line". When you have a tank/reservoir, there is always some amount of coolant vapor in that space; physical laws of vapor pressure apply. When you have coolant vapor in a space like that, that is evaporation.
Now, if you want to reduce or eliminate that evaporation, you can do so - but it requires additional inputs of vapor barriers like nitrogen or air. So you can put in something like an air bladder or nitrogen blanket to inhibit the formation of this water vapor (which will never be perfect; it cannot be perfect, because physics still applies) but that creates new points where water can escape. Nitrogen blankets typically need a constant in/out flow, and if it's flowing out - so too will incremental amounts of water, which requires topping off.
Salts can develop because of incidental exposure to people - sweat, bodily fluids, that sort of thing. People need to work on these systems, and people aren't perfectly clean. If the cooling system is in a cleanroom environment you can eliminate most of these issues, but never all of them. But the concern about salts isn't limited to the interior of the cooling system; they can accumulate on the exterior as well, which can lead to corrosion and damage to piping. Again, this can be caused by people and bodily fluids, but also from animals (which probably aren't going to be a major problem, but who knows), the ambient environment, that sort of thing.
Anywhere that there is water, there are microbes. If you want to disinfect the water to limit microbial growth, you can do so - but it requires chemical addition to your water, which means adding things like bleaching agents and other halogen-containing compounds. When you add bleaching agents to systems containing metals, you'll create salts - which brings us back to the problem mentioned above. Now, you can create your cooling system out of plastic piping (PVC, HDPE, etc), but plastics and common bleaching agents don't always play well together.
Beyond all of this, there's also the problem of moving parts. For a typical cooling system, you'll need your reservoir, some pumps, and some heat exchangers (probably air-cooled). All of these - the pumps in particular - have moving parts, probably made out of metal. These parts wear out. In the case of pumps, "wearing out" will include the gradual grinding down of the gears within the pump, which will start sending tiny metal fragments throughout the system. These will promote rusting, which will not only cause damage to the piping but also create accumulation of corrosive byproducts which will lead to further corrosion, as well as affecting water purity.
As for re-chilling, when I said that physical laws apply, I'm also referring to Newton's Laws of Thermodynamics. You cannot create a perfect closed-loop that loses zero energy, it is physically impossible. You always lose some of your energy to entropy and waste heat generation from things like friction (2nd Law). Over time, your cooling system will become less cool, and will require either an influx of new, cooler coolant ... or will require re-application of external cooling to bring it back to within your pre-determined requirements.
Leaks will always, always happen - anyplace you have a weld, a joint, a flange, a change in direction, or a connection, there is a point that a leak can occur. Anyplace you have a valve for filling up, draining, installation of an instrument, that sort of thing - all of those are potential leak points.
All of these factors, and more, are why I say you cannot create a truly closed-loop system. It is simply not possible. And even beyond the impossibility of it, there's also the financial reality that while you can create really, really tight "closed-loop systems" that have near-zero leakage and loss, it is never truly zero ... and the financial costs associated with trying to make that perfect closed-loop system may be exorbitantly high. It is often far less expensive to create an open-loop, particularly for a major installation like a data center. For something small like a home air conditioning system, it's a different story; the scale is vastly different.
-
Comment on Data centers don't raise people's water bills in ~tech
Drewbahr Link ParentA desktop PC uses several orders of magnitude less water than a data center. It's not comparable, even at scale - in no small part because the number of points of failure are substantially less in...A desktop PC uses several orders of magnitude less water than a data center. It's not comparable, even at scale - in no small part because the number of points of failure are substantially less in a PC than for a large-scale installation.
-
Comment on Data centers don't raise people's water bills in ~tech
Drewbahr Link ParentIt's not cynicism. It's physical reality.It's not cynicism. It's physical reality.
-
Comment on Data centers don't raise people's water bills in ~tech
Drewbahr Link ParentAny system that insists it will only be filled once, is lying to you. Basic laws of physics apply to everything.Any system that insists it will only be filled once, is lying to you. Basic laws of physics apply to everything.
-
Comment on Data centers don't raise people's water bills in ~tech
Drewbahr Link ParentThere is no such thing as a truly closed loop system. Cooling water has to maintain a minimum degree of purity, and works best within a certain range of temperatures. Over time the water in the...There is no such thing as a truly closed loop system. Cooling water has to maintain a minimum degree of purity, and works best within a certain range of temperatures. Over time the water in the loop will get contaminated (with microbes, salts, etc), and over time it'll also experience temperature creep (because entropy exists). So it'll have to be replaced, re-cooled, etc. There's also water loss due to evaporation, leaks, that sort of thing.
Cooling water systems at the refineries where I've worked operate in the thousands, if not millions of gallons per minute flow rates. It's not closed loop, but I'd imagine data centers probably require comparable flow rates. Let's pretend that your data center operates at a paltry 10,000 gallons per minute. If you only ever have to make up 1/100th of a percent (number out of my ass there) to maintain water balance, that's still one gallon per minute you have to maintain. Over a day, that's 1,440 gallons per day.
Let's say the typical office worker uses the toilet three times a shift, and drinks one gallon of water while at work (which would be a lot). One flush is typically 1.6 gallons of water, so per person that's roughly 6 gallons per day at the office. In order for the office to come close to the data center, it would need to have 240 people in attendance per day.
My office buildings at work don't have 240 people, so by these wild ass numbers a data center consumes more water, per day, than my office buildings do.
But again, these are rough numbers, just for food for thought.
-
Comment on Data centers don't raise people's water bills in ~tech
Drewbahr Link ParentI would imagine flushing toilets and using sinks costumes substantially less water than a data center.I would imagine flushing toilets and using sinks costumes substantially less water than a data center.
-
Comment on A Palo Alto scientist's $10M plan to kill California redistricting in ~society
-
Comment on What's your current PC wallpaper? in ~tech
Drewbahr LinkA picture of my wife, my dog, and my son reading a book together, as it has been for the past several years.A picture of my wife, my dog, and my son reading a book together, as it has been for the past several years.
-
Comment on Actually, slavery was very bad in ~society
Drewbahr Link ParentTrump is a widely known racist. He absolutely was trying to downplay the horrors of slavery. He does not deserve any benefit of the doubt here.Trump is a widely known racist. He absolutely was trying to downplay the horrors of slavery. He does not deserve any benefit of the doubt here.
-
Comment on Actually, slavery was very bad in ~society
Drewbahr Link ParentI hate to be the bearer of bad news, but the notion that The United States has truly been a "tapestry of pluralism" ... is just not supported by its own historical record. The nation was founded...I hate to be the bearer of bad news, but the notion that The United States has truly been a "tapestry of pluralism" ... is just not supported by its own historical record. The nation was founded by slaveholders, for slaveholders. We were not among the first European nations to criminalize slavery, and when we did, it was only the import of slaves. Black Americans have only had the ability to attend integrated colleges for two generations. Two.
McCarthyism was not long ago. It is not the product of some bygone age. The kind of rank racism and hatred that birthed McCarthyism is still going strong in US politics to this day. It never went away. It barely even went undercover. Look up what Reagan's policies and views were and tell me they weren't just a continuation of McCarthyism.
If you're a white American male, particularly one fitting the stereotypically cis-hetero archetype, and if you're some flavor of Protestant Christian ... you've probably never really been affected by the ways that the US treats anyone else. That was the case for me, for most of my life. But if you're anything outside of that narrow demographic, then your experience in this country has been very, very different. Particularly if you're Black.
As for how Americans can be living through this and cheer it on? That part's easy. It's the same reason they've always been cheering it on. Because there's a significant - not a majority per se, but still significant - chunk of the population that agrees with the racism, because they are racist. The Klan never went away. Neither did the Nazis - in fact, the USA not only imported many, many Nazis from Germany after WW2, but things like Jim Crow laws and "The Trail of Tears" directly influenced and led to Nazi thought.
What we're seeing now, is what generations of non-white Americans have been experiencing for generations. Just, now it's starting to affect some white people too.
-
Comment on AI eroded doctors’ ability to spot cancer within months in study in ~tech
Drewbahr Link ParentPeople can have disagreements and misunderstandings without being called neuroatypical. I'm going to walk away, because yet another person is coming in here trying to explain how my observation is...People can have disagreements and misunderstandings without being called neuroatypical.
I'm going to walk away, because yet another person is coming in here trying to explain how my observation is incorrect.
-
Comment on AI eroded doctors’ ability to spot cancer within months in study in ~tech
Drewbahr Link ParentI get that. I understand that LLMs only occasionally end up saying factual things, and even then only by chance or coincidence. I understand statistical probabilities, I really truly do. I don't...I get that. I understand that LLMs only occasionally end up saying factual things, and even then only by chance or coincidence. I understand statistical probabilities, I really truly do.
I don't care if the system is working correctly. I'm focused entirely on the statement I quoted.
I must be taking crazy pills
-
Comment on AI eroded doctors’ ability to spot cancer within months in study in ~tech
Drewbahr Link ParentA tool, even when used for the purpose it wasn't intended for, can still just be wrong. Please, I beg you, understand that I'm solely arguing that an LLM can be wrong. New York City is not the...A tool, even when used for the purpose it wasn't intended for, can still just be wrong.
Please, I beg you, understand that I'm solely arguing that an LLM can be wrong. New York City is not the capital of the United States, no matter who or what says it is.
I feel like I'm taking crazy pills here.
-
Comment on AI eroded doctors’ ability to spot cancer within months in study in ~tech
Drewbahr Link ParentMy whole disagreement is here: It is wrong. Full stop.My whole disagreement is here:
When a chatbot says that "nyc is the capital of the united states" is that response wrong? It's not, really.
It is wrong. Full stop.
She was British and a queen, so the answer's going to be yes, right?
Predatory collection of beautiful things is what The British Museum is all about, and I'd imagine the monarchs are no different.