The most official water costs $120k a gallon
93 points by surprisetalk 7 hours ago | 70 comments

sudb 5 hours ago
Reminds of me of the "most expensive peanut butter" at ~$2.44/g - sold by NIST, I assume for some sort of peanut butter factory process calibration: https://shop.nist.gov/ccrz__ProductDetails?sku=2387
reply
cyphar 3 hours ago
For more context, sample reference materials from the NIST reference catalogue are not meant to be exceptionally high purity examples of the sample -- SRMs are usually commercially-produced products. The real value is that each sample comes with a datasheet with incredibly detailed and accurate measurements that are (as you said) used for calibration of measuring devices in industrial settings. To quote a Veritasium video about the topic from a few years ago[1], "the thing you're paying for is not the peanut butter, it's the knowledge of exactly what is in the peanut butter".

[1]: https://youtu.be/esQyYGezS7c?t=226

reply
buildsjets 48 minutes ago
Truth. I once paid nearly $500 for a set of 17 certified, calibrated wooden pencils, with documentation showing that each one met their rated hardness levels of 9H thru 6B. And for $500, they don't even come with erasers, that is an added-cost option.

https://www.gardco.com/product/pencil-hardness-test-kit-astm...

reply
foobarian 3 minutes ago
Can I ask why :-) Sounds like an interesting project.
reply
cjs_ac 4 hours ago
SRM 2387 pairs well with a cup of ISO 3103.

https://en.wikipedia.org/wiki/ISO_3103

reply
buildsjets 4 hours ago
Cap your meal off with a NIST 1196a standard issue cigarette? I am told they taste like drywall and taxes.

https://www.nist.gov/news-events/news/2020/02/nist-releases-...

Perhaps you would prefer some standard issue meat.

https://www.sigmaaldrich.com/US/en/product/sial/nist1546a

reply
andy99 11 minutes ago
I had seen the cigarette. But but not the meat which costs $2,690 for four tuna-can size cans and is a mix of pork and chicken

Thinking of a concept for a restaurant meal along the lines of a calibrated fried spam dinner, that costs ~2k for about a can worth, fried like they do in Hawaii.

Come to think of it, would love to see an “iron chef- calibrated foods” edition.

reply
sushisource 4 hours ago
A suggested product is "domestic sludge", lol. This is such a fun rabbit hole:

https://www.sigmaaldrich.com/US/en/product/sial/nist2781

reply
inigyou 2 hours ago
Pretty much every process conducted at scale will have such reference samples.
reply
tshaddox 3 hours ago
reply
selimthegrim 38 seconds ago
Now I wonder what the GOST ones were like
reply
throwup238 3 hours ago
That’s $508 for two cartons for 400 cigarettes total, so $1.27 per cigarette. They’re not characterized like the peanut butter, they’re just manufactured to a specific QC standard.

California/New York prices are easily $0.75/cigg so it’s not that much more expensive.

reply
ryancw 3 hours ago
400 cigarettes actually! Not too bad.
reply
georgeburdell 2 hours ago
Yes, it’s a bit frightening to know what a hidden dependency NIST is on many things. In my industry, all calibrations ultimately derive from physical NIST standards for resistance, frequency, etc.
reply
masklinn 2 hours ago
It’s frightening that… NIST is doing the very thing it was established to do back when it was called the Bureau of Standards?
reply
ajsnigrutin 22 minutes ago
You can still make cookies out of reference and/or lab materials :)

https://www.youtube.com/watch?v=crjxpZHv7Hk

reply
juleiie 4 hours ago
I wouldn’t put it past people in my country to buy it just to show off wealth :/

I have recently realized we imported the western world in the most superficial ways.

reply
spullara 4 hours ago
A US gallon of deuterium water costs roughly $2,600 to $3,800, while a gallon of pure tritium water (super-heavy water) would cost about $44 million.
reply
BizarroLand 3 hours ago
Makes me want to go to the ocean and start filtering out water somehow.
reply
inigyou 2 hours ago
You can centrifuge it. Heavier molecules move closer to the bottom. Repeat until you reach the desired purity.

You need shitloads of ordinary water to extract a liter of heavy water. And I'd be surprised if there's any superheavy water in that, because tritium radioactively decays.

Fun fact: Heavy water is slightly toxic to humans. If you only drank heavy water, you'd die as it shifts the chemical reactions in your body out of balance. A moderate amount is fine though.

reply
danbruc 2 hours ago
Another fun fact - heavy water has a sweet taste.
reply
htrp 5 minutes ago
How did someone ever figure this one out?
reply
delecti 2 hours ago
There are a variety of ways to do it, but you'd want to be careful that you don't attract attention from your local nuclear regulatory agency.

https://en.wikipedia.org/wiki/Heavy_water#Production_methods

reply
rwmj 4 hours ago
Interesting article but left me wondering why they don't use ¹H₂¹⁶O as the "standard water"? Surely you can make that (relatively) easily in a centrifuge by spinning out the lightest fraction of water.
reply
thayne 3 hours ago
At the very least that has the advantage that it's possible to make more of it, so you aren't dependent on a limited sample of standard water.
reply
artisinal 4 hours ago
This is why we should ditch Celsius for Kelvin. I never liked how we coupled 0 and 100 to where water freezes and boils under certain conditions. Kelvin doesn't rely on non-invariant reference points.
reply
Guvante 4 hours ago
You want numbers to be understandable and K produces illogical numbers.

(There is a reason it is common to drop the century when talking about years, humans are less skilled at numbers over 100)

And I know temperatures in Fahrenheit go higher but it is common to refer to Phoenix as "over 100" even when it's high can peak over 120.

reply
zinckiwi 3 hours ago
That’s 02025 thinking.
reply
thaumasiotes 3 hours ago
> (There is a reason it is common to drop the century when talking about years,

Yep.

> humans are less skilled at numbers over 100)

No, the reason is that the century is constant and therefore carries no information. Where the century isn't constant, it never gets dropped.

reply
Guvante 2 hours ago
You can say the same thing about decades

No one says 6 even though it has a very reasonable information density for conversations. It is either a relative year or a two digit one.

reply
warmdarksea 3 hours ago
this is actually why i like fahrenheit the most... 0F is "really cold" and 100F is "really hot", whereas in celsius 0C is "kind of cold" and 100C is "uninhabitable". it makes sense for the 0-100 range to be reserved for temperatures frequently encountered in our habitat as humans i think...
reply
kccqzy 2 hours ago
That’s a biased view that leads to a scale only convenient for people living in particular latitudes. Now talk to someone living near the equator like Singapore, and 0F is unimaginably cold (even 40F is basically unheard of), and 100F is very hot but not extraordinarily so. Then you talk to someone living close to the arctic and the converse is true.

All this talk about 0F to 100F being frequently encountered probably reveals your local climate conditions more than it helps define a useful global standard for all humans. We should accept that humans live in more diverse conditions than we intuit and such a scale probably won’t be useful for all.

reply
cyral 2 hours ago
Now I'm imagining some kind of timezone-like temperature unit that varies around the world, and a "100 things programmers should know about ___" post about all the intricacies and special cases.
reply
wongarsu 3 hours ago
Celsius maps nicely to the temperatures observed in a typical Finnish sauna. 100C is among the highest typical sauna temperatures (very relaxing in dry air), and afterwards you cool off with water around 0C

Also, having 100C as water temp for black tee and 80 for green tee is nice to remember

reply
al_borland 3 hours ago
This has always been my primary argument in favor of Fahrenheit. It makes sense for humans, which is all a majority of the population really cares about. In addition, it's more granular. Sometimes turning the thermostat up or down by 1C could be too much, where 1F is just right.

In science and math, I can see where having round numbers for commonly used temperatures would be useful, but if 0C and 100C are the only places that happens, it's not that hard to memorize 2 numbers.

reply
Timpanzee 2 hours ago
You only think it makes sense for humans because it makes sense for you, where you live the same way you feel changing a thermostat by 1F is perfect but 1C is too much: that’s true for you, not universally.

Fahrenheit functionally requires you to be biased in order for it to seem like a good system.

reply
al_borland 2 hours ago
It provides greater granularity in the range of human comfort, without descending into decimals. How is that not more useful for day-to-day human usage?

0F is cold, 100F is hot. I experience both of these temperatures outside where I live every year.

-20C to 40C seems much more arbitrary as a way to measure the zone humans mostly live in.

I would wager that the outdoor and room temperature are an overwhelming majority of human use for temperature. It seems like it would make sense to optimize for how temperature is more commonly used at a population level.

I would say those who see Celsius as a good system are also biased, because they grew up with it and it’s what they’re used to.

reply
masklinn 2 hours ago
> 0F is cold, 100F is hot. I experience both of these temperatures outside where I live every year.

Most people never experience anywhere near 0F in their lives.

> -20C to -40C seems much more arbitrary as a way to measure the zone humans mostly live in.

The zone most humans live in does not routinely (if ever) get anywhere near -20, to say nothing of -40.

reply
al_borland 10 minutes ago
The -40C was a typo, should have been 40C, to mirror the 0F to 100F, rounding a bit to make it less silly. It’s been fixed now.

Millions do experience 0F, and it’s ok to not use the full range. At or around 0F is cold, as you move closer or beyond it, it’s cold. As you move closer to or beyond 100F, it’s hot. Most people would likely prefer to never deal with <0F or >100F.

reply
vagrantJin 2 hours ago
Celcius is pretty intuitive actually. It takes atmospheric pressure into account, Water is pretty abundant on the surface of the planet, and life kinda depends on it. It's a remarkable substance and quite stable. I dont think theres another substance where we can see all three phases with kitchen appliances and still consume it. Equally good for bathing and drinking. I need steam to clean clothes and cook my food. Using water as a benchmark for how we measure temperature is remarkably useful.
reply
al_borland 2 hours ago
With all of those examples, the actual temperature doesn’t really matter.

For drinking and bathing, people go by feel. In the kitchen, the freezer makes water freeze, and the stovetop will make it boil into steam, but exact temperatures aren’t really relevant here either. The same goes for ironing clothes, the iron just does it.

When I think about where explicit temperatures are needed in the home, the oven and thermostat are the two things that come to mind. Water is pretty irrelevant for both of those. For cooking or frying someone may monitor meat or oil temps, but those aren’t water. A hot tub might have an actual temp on it, but those tend to be around 100F, not 100C, as they are meant for humans, not boiling.

Water is an amazing liquid, but I still struggle to see how measuring all our temperatures based on it makes sense.

reply
arjie 20 minutes ago
Haha, judging by my personal standards, we need to set the zero to 15 C and the 100 to 40 C. My Degrees Arjie are going to be a revolution.
reply
Paracompact 2 hours ago
OF will still kill you faster than 100F, and most folks aren't comfortable at 50F.

It inspires the following idea: A nonlinear scale where ~72F is 0, and +X represents the same level of heat discomfort as -X represents cold discomfort.

reply
ajmurmann 3 hours ago
This makes sense if weather is the main focus. For things like cooking Celsius is great though.
reply
quickthrowman 5 minutes ago
The temperature water boils at is irrelevant when you can just look at it to see if it’s boiling. Any stove will get it hot enough, assuming you have enough energy.
reply
0cf8612b2e1e 3 hours ago
I am failing to understand why there would be a distinction. They all seem arbitrary set points you learn.

450F vs 230C does not feel like anything gets easier there. Maybe in some highly technical baking there are some nice deltas that appear?

reply
clickety_clack 3 hours ago
Aren’t you just remembering specific numbers though? Unless you’re trying to boil water unseen so that you can only see a thermometer, the specific 100°C mark isn't that useful.
reply
clickety_clack 3 hours ago
I actually came around to Fahrenheit for the exact same reason. 0°F and 100°F are kind of the top and bottom of the temperature range I’m going to be experiencing and anything outside that is waaay too hot or cold. In Celsius a large portion of people have to flip to negative numbers regularly during the normal year, and everything between 45°C and 100°C is completely unused.
reply
mmooss 2 hours ago
> 0°F and 100°F are kind of the top and bottom of the temperature range I’m going to be experiencing and anything outside that is waaay too hot or cold.

Food, saunas, the weather, a fever, and much more are often over 100F.

reply
Timpanzee 2 hours ago
Except 0F being “really cold” and 100F being “really hot” is a personal bias of yours based on where you’ve lived. It’s not true for everyone or everywhere.

And even if it were true, it still wouldn’t work. 100F in a dry environment is nice while, in a humid environment, 100F is awful. Same for 0F, a dry and static 0F is fine while a wet and windy 0F is killer.

reply
TylerE 3 hours ago
Also for things like hvac temps Celsius is rather too coarse where as 1F is right around the perceivable difference.
reply
SketchySeaBeast 3 hours ago
There's nothing stopping the use of decimal points for Celsius. Though there's also nothing stopping decimal points in Fahrenheit, the rare instance where an imperial sub-unit doesn't make things go fucky.
reply
analog31 4 hours ago
I think that both the Celsius and Fahrenheit scales are now defined in terms of Kelvin and a formula. Most if not all of the units of measure are no longer related to their history except for an agreement to make the new units match the old ones at some point.
reply
0cf8612b2e1e 3 hours ago
The US is technically metric in that the units are defined in terms of the metric equivalent. A US inch is exactly 2.54cm.
reply
throwatdem12311 3 hours ago
Now do metric vs imperial.
reply
artisinal 3 hours ago
Nothing screams freedom more than using ancient terms like horsepower and measuring tires in inches and percentages.

The other day I heard an American talk about a TV being 65 inches. Took me a while to understand they had a 165 centimeter diagonal television.

reply
tancop 2 hours ago
most tv makers and electronics stores in europe give you the screen size in inches first and cm second. they only come in a few common sizes so the specific units dont really matter.
reply
clickety_clack 3 hours ago
I bet you’d use lightyears over petametres or exametres though.
reply
Alpha3031 2 hours ago
It's more typical to use parsecs in practice for astronomy.
reply
vel0city 2 hours ago
Speaking of tires, its actually a crazy mixture of units here in the US. You might have a size like P215/65R15 which translates to a radial passenger tire of a 15" rim, 215mm wide tread (sidewall to sidewall), and an aspect ratio of 65 which is a ratio of the sidewall to tread width.

So a tire size is expressed in inches, millimeters, and ratios.

reply
alehlopeh 3 hours ago
Surely you mean tyres? It’s all arbitrary. Get over yourself.
reply
xgulfie 4 hours ago
Kelvin does still rely on each degree being 1% of the way between freezing and boiling so it kinda does rely on a variant reference point
reply
chermi 4 hours ago
The reference point is the triple point, and invariant.
reply
xgulfie 2 hours ago
Sure technically yes, but also celsius is now technically defined as just Kelvin plus 273.15. Degrees have technically been redefined as exactly 1.380649×10^−23 joules
reply
chermi 44 minutes ago
Why is that only "technically yes"? It's precisely how the thermodynamic and empirical temperature scales are matched.
reply
cactusplant7374 3 hours ago
How does the water taste?
reply
pack_stimulus 5 hours ago
Loved reading this. Thank you!
reply
dorianmariecom 4 hours ago
gallon is not a si unit ;)
reply
sandworm101 4 hours ago
As specialized world-class lab supplies and reagents go, that is not very expensive. Stuff like pure Xe (for space thrusters) can be more by mass.
reply
alanfalcon 5 hours ago
[dead]
reply