I'll never understand why tech companies choose some of the locations for their data centers. Considering a big thing with data centers is "keeping stuff cool", you would think they would build them in the northern states, closer to Canada versus the hot sticky swamp.
phyrex · 25m ago
I'm pretty sure Meta's team has written about that at length. It's about many things, such as (power/transportation/internet/energy) infrastructure, political situation, available workforce, vicinity to population centers, property prices, and a whole lot more
gen3 · 42m ago
The speed of light is incredibly slow and data through a wire is even slower. Proximity is worth something
aeve890 · 24m ago
>The speed of light is incredibly slow
I get where you're coming from but still I find funny in so many levels that the literal speed limit of the universe is too slow for our mundane (or even banal in FB case) needs. the universe isn't good enough to our need to move bullshit across the globe. surreal.
In the same vein it would be awesome if this _need for speed_ would materialize in infinite funding of neutrino based communication research.
coolspot · 29m ago
Doesn’t matter for training, as all GPUs are colocated in the same DC.
gowld · 32m ago
So why put a datacenter in Louisiana, far from the vast majority of people in the Americas?
gen3 · 18m ago
You’re pretty close to Texas tech hubs, plus Meta was able to convince them to pass Louisiana Act No 730 so they save a ton on capex
khuey · 25m ago
Cheap land and cheap energy.
righthand · 26m ago
Louisiana and New Orleans have been pushing to make the city a “tech hub” for the past 10 years (why would you build data centers in a flood-prone basin below sea level? I don’t know). I imagine most of it is striking a sweet heart deal with the municipalities that want the business.
jimt1234 · 1h ago
I'm not super-familiar with Louisiana, but my general impression is there's a lot of climate/weather events that are gonna impact power reliability. Hmmm.
mritterhoff · 2h ago
While Meta has a non-binding promise to build more renewable energy, the Louisiana Legislature passed a new law that adds natural gas to the definition of green energy, allowing Zuckerberg and others to count Entergy’s gas turbines as “green.”
As much as I prefer burning gas over coal, conflating it with zero(-ish) emission energy sources like wind, solar, and nuclear is bad.
juujian · 1h ago
Due to all the methane leaks, gas isn't even as much cleaner than coal as it was purported to be... But hey monitoring programs got cut so I guess that solves the problem...
potato3732842 · 37m ago
From a purely greenhouse gas accounting, sure.
Anyone who has to live in a fairly closed system (i.e. this planet) in which fossil fuels are burned for power would be beyond a fool to not strongly prefer gas over coal seeing as their greenhouse emissions are close enough to be within arguing distance. It's all the other stuff coming out that's the problem with coal.
PaulStatezny · 24m ago
I think you might have a typo. Reading your comment literally, it doesn't make sense.
Summarized: Anyone would be a fool not to prefer gas or coal, because their emissions are nearly equal.
One doesn't follow from the other, can you correct/elaborate?
marcusb · 17m ago
They said gas over coal. If you accept the claim that GHG emissions from gas and coal are roughly equal, their claim is the other pollutants from burning coal make gas far more preferable.
potato3732842 · 19m ago
If their greenhouse emissions are even close only a moron would not pick gas over coal because the former's emissions lack all the other nasty byproducts that are present in the latter's emissions.
rcxdude · 17m ago
I think the point is: "you'd be a fool not to prefer gas, because while the greenhouse emissions are about the same, for everything else coal is much worse"
mritterhoff · 1h ago
I agree methane leaks (and monitoring programs cuts) are a problem. But even with them, methane burns much more cleanly than coal. The former primarily emits CO2 and H2O, while the latter emits SO2, NOx, heavy metals and more.
mikeyouse · 55m ago
These definitions always get muddled when flipping between CO2 emissions or pollution... coal is definitely worse from a pollution standpoint, is likely worse from a carbon standpoint, but much of the methane produced from natural gas production is just released into the atmosphere and has a dramatically higher warming effect compared to CO2 -- on the order of 80x more warming potential over 20 years and at least 20x over 100 years.
So only looking at the byproducts of methane combustion is also misleading since nat. gas plants largely aren't burning methane - and blanket statements for all natural gas are also misleading since e.g. the gas from Canada is extremely 'Sour' and releases a ton of sulfur compounds when burned, often with fewer scrubbers than coal plants.
chaos_emergent · 56m ago
I think the problem is that methane is 20x more powerful a GHG than CO2
chris_va · 1h ago
As an aside, methane leaks from coal mines can be worse than upstream leaks from O&G.
h1fra · 1h ago
burning fossil fuel and depleting the local water aquifer, I'm starting to miss the greenwashing era
estearum · 1h ago
Behaving a certain way to pretend being virtuous, it turns out, is almost as good as actually being virtuous.
matthewdgreen · 1h ago
Who is this non-binding promise being made to, and why make one?
JKCalhoun · 1h ago
"I'll gladly pay you Tuesday for a hamburger today…" Seems to be pretty common these days when corporate make deals with cities/counties/states.
juujian · 1h ago
Classic Mark "trust me bro" Zuckerberg
blitzar · 1h ago
Dumb fucks
maxehmookau · 1h ago
Adding natural gas to the definition of green energy is absolutely wild. How on earth did that pass?
dublinben · 1h ago
Louisiana has a long history of political corruption, and the petrochemical industry is a major part of their economy.
jjice · 1h ago
I have to imagine it's just a complete lack of care and classifying it as "green" helps push through something that they're being lobbied to push. I can't imagine this is anything but nonsense.
yoyohello13 · 34m ago
We all know how it passed. Legislators have lots of money in natural gas I’m sure.
babypuncher · 30m ago
These are the same morons who now reject germ theory and think vaccines cause autism, so we shouldn't be surprised that they are functionally illiterate in other scientific fields as well.
I always told myself if I ever became a "tech billionaire" I'd buy out a random abandoned town somewhere, setup high speed internet, and turn a ghost town into a high tech town, cause why not? You could easily become mayor and approve some reasonable projects. Sell extremely affordable housing for the buck (close to actual cost).
I do often wonder if it might be worthwhile to shove a bunch of server farms into a few abandoned mines, if you setup the appropriate infrastructure in said mines to protect your data centers.
idiotsecant · 1h ago
The major tech companies are all scrambling to snap up cheap energy right now. The result is that we are dumping a whole lot of additional carbon in red states and adding a while lot of additional extremely expensive per MWh sources in blue states. In both cases, the winners will be tech company shareholders and the losers will be the people who actually live in these communities who will end up with dirtier, more expensive power.
matthewdgreen · 1h ago
The losers are going to be the energy companies who think they’re getting long-term energy sales but probably won’t be, since these techniques will get more efficient.
The demand for energy will never go down, the more we can produce the more we will use.
lotsofpulp · 14m ago
The article says
> Electricity demand in the U.S. held steady for 15 years but, last year, it increased by 3%— marking the fifth-highest rise this century. More jumps are projected for years to come.
Total electricity generated has been relatively flat for a couple decades.
idiotsecant · 1h ago
The techniques will get more efficient, but the quantity of training will increase monotonically. We aren't going to use less energy overall. The ratepayers are absolutely the ones who will lose out on this.
JKCalhoun · 1h ago
> The project entails more than 2 gigawatts of computing capacity—Zuckerberg said it could eventually expand to 5 gigawatts—programmed to train open-source large language models.
Given that the human brain takes much longer to "train", I wonder how the energy efficiency pans out — comparing the two.
idiotsecant · 1h ago
Biological systems are wildly energy efficient, that's kind of their whole thing. The average human will consume approximately 75kwh worth of calories in their lifetime. There are electric cars with bigger batteries.
[Edit] ok, yes, please. I get that i missed the k in kcal. The point stands. Biological training is massively more efficient, even when you forget to multiply by 1000
ak217 · 1h ago
This is wrong by at least three orders of magnitude. Very roughly, a human requires 2000 kcal a day = 2 kWh a day so 75 kWh is enough to cover about a month, putting aside the upstream losses in the energy supply chain (which are far greater for humans).
In general, saying that biological systems are "wildly efficient" is... wildly wrong. Some biological processes are optimized by evolution... most are not. There are no bicycles in nature.
ctoth · 1h ago
You're off by about three orders of magnitude.
A human consuming 2000 kcal/day (conservative estimate) uses about 2.32 kWh per day. Over 75 years, that's roughly 64,000 kWh.
idiotsecant · 1h ago
Oh, right i did a conversion wrong. Woops. In any case, a rounding error when talking about gigawatts of generation capacity
trylist · 1h ago
We're efficient once we have the energy, sure. How much energy does it take to go from raw sunlight to a calorie your body is actually able to use, and finally to your dinner table?
mushroomba · 1h ago
All of our food was alive before we ate it. All calories used by living things are efficient.
Life is an end unto itself. It does not need to justify its existence by the moral code of technocrat materialism. The fact that this discussion is being had on this board in good faith is morally condemning of our worldview.
gowld · 27m ago
Your forget that a biological system has approximately 0 throughput in work done.
Nearly everything a biological system accomplishes depends on massive external machinery.
Humans are only intellectually interesting because of their use of tools.
I get where you're coming from but still I find funny in so many levels that the literal speed limit of the universe is too slow for our mundane (or even banal in FB case) needs. the universe isn't good enough to our need to move bullshit across the globe. surreal.
In the same vein it would be awesome if this _need for speed_ would materialize in infinite funding of neutrino based communication research.
As much as I prefer burning gas over coal, conflating it with zero(-ish) emission energy sources like wind, solar, and nuclear is bad.
Anyone who has to live in a fairly closed system (i.e. this planet) in which fossil fuels are burned for power would be beyond a fool to not strongly prefer gas over coal seeing as their greenhouse emissions are close enough to be within arguing distance. It's all the other stuff coming out that's the problem with coal.
Summarized: Anyone would be a fool not to prefer gas or coal, because their emissions are nearly equal.
One doesn't follow from the other, can you correct/elaborate?
So only looking at the byproducts of methane combustion is also misleading since nat. gas plants largely aren't burning methane - and blanket statements for all natural gas are also misleading since e.g. the gas from Canada is extremely 'Sour' and releases a ton of sulfur compounds when burned, often with fewer scrubbers than coal plants.
I do often wonder if it might be worthwhile to shove a bunch of server farms into a few abandoned mines, if you setup the appropriate infrastructure in said mines to protect your data centers.
> Electricity demand in the U.S. held steady for 15 years but, last year, it increased by 3%— marking the fifth-highest rise this century. More jumps are projected for years to come.
https://www.eia.gov/energyexplained/electricity/electricity-...
Total electricity generated has been relatively flat for a couple decades.
Given that the human brain takes much longer to "train", I wonder how the energy efficiency pans out — comparing the two.
[Edit] ok, yes, please. I get that i missed the k in kcal. The point stands. Biological training is massively more efficient, even when you forget to multiply by 1000
In general, saying that biological systems are "wildly efficient" is... wildly wrong. Some biological processes are optimized by evolution... most are not. There are no bicycles in nature.
A human consuming 2000 kcal/day (conservative estimate) uses about 2.32 kWh per day. Over 75 years, that's roughly 64,000 kWh.
Nearly everything a biological system accomplishes depends on massive external machinery.
Humans are only intellectually interesting because of their use of tools.