A corporation running a nuclear reactor to train AIs might just be the most cyberpunk news headline I’ve ever seen.
This gave me an idea for some level design I might want to use in a video game.
I’d play that, the fact that you’d include this content sounds dope
I don’t get why a train AI would need so much power, how hard is it to drive a train?
Will the nuclear reactors be on the train with the AI, or will it be some sort of wired transfer?
I assume they can power the train AI by pantograph or third rail - no reason to have nuclear powered trains, this isn’t Factorio.
If your train has a bad conductor, it will require more power.
Yeah tell me about it, I only have bad conductors.
This is how we end up with Blaine the Mono
Not sure if you’re making fun or actually not understanding? To clarify, they need the power for training AI models. No trains are involved, neither passenger nor cargo – though atom powered trains sounds interesting as well!
Thank you, you are very kind. It was meant as a joke.
Nuclear power is actually way cheaper.
You just need to find a geologically safe place to put it and you need to make sure everyone involved follows safety protocols to the letter. And you can’t have anyone cutting corners to save money. You need to spare no expense when it comes to safety.
The only issue is that people don’t stay strict with keeping everything safe sometimes. People are terrified of it because when something goes wrong, everyone can see the very gruesome results very quickly
But I don’t think microsoft or any company should be making an AI at the rate they are if it’s going to take as much resources as it seems.
https://en.wikipedia.org/wiki/Cost_of_electricity_by_source#Capital_costs
No, it isnt.
Safety isnt what makes nuclear expensive. You actually got rawdogged by Nuclear fanatics.Damn. Sucks to still see Natural Gas as the cheapest.
I think that modern nuclear designs have a place in decarbonization but I don’t think it is cheaper and we have a lot of hurdles still.
Wind power at shores is generally cheapest
Wikipedia is owned by someone who’s married to an intelligence agency member.
Lemmy is owned by genocide deniers. Whats your point?
Capital costs is incomplete, you need to look at lifetime costs versus lifetime production to get a more useful average - Levelized cost of energy (LCOE)
https://en.wikipedia.org/wiki/Levelized_cost_of_electricity
https://www.lazard.com/research-insights/2023-levelized-cost-of-energyplus/
https://ourworldindata.org/grapher/levelized-cost-of-energy
https://www.iea.org/reports/projected-costs-of-generating-electricity-2020
Wikipedia is owned by someone who’s married to an intelligence agency member.
I see absolutely no relevance here
Yea this entire clickbait can be summarized as “company looks to spend less on high capacity power”
And you just need to find good non science fiction way to deal with nuclear waste.
And some ethical ways of aquiring uranium.
With all that calculated in, I am certain how much cheaper it is.
Rare lemming reasonable take
I guess the nuclear power people are gonna become Microsoft fans… powered by nuclear power.
There is nothing wrong with nuclear power
Thats what France said.
And France is doing perfectly fine with it except for skimping on maintenance and also them coming up on their end-of-life without replacement
Half of them were off at some point which means they lost over a third of their entire energy production which means they had to import a ton.
Some of them are still not back online.
deleted by creator
That list records 8 fatalities related to nuclear power in the US. All time.
Coal is responsible for more than 40. Per year. Just in my city.
You are also forgetting the years of life lost due to air pollution.
To be fair, though, the top comment said “nothing wrong” and the guy just followed up with the wrong
I still think we have access to this energy source and are basically just ignoring it so oil and coal and continue to soak up tax payer funded subsidies and ruin our planet.
deleted by creator
Almost anything has the potential to negatively affect tens of thousands of people when it’s managed as recklessly and negligently as Chernobyl.
Chernobyl was less a reactor and more a bomb with a very long fuse. Saying we shouldn’t build nuclear reactors today is like saying you shouldn’t take a modern cruise because 14th century sailing ships sank all the time.
deleted by creator
How many years has it been since chernobyl?
deleted by creator
Not the 320 years it will take for the exclusion zone to be habitable again.
people need lots of energy for life as well, now what??
Just fill the Country with Solar, Wind and Water… won’t take 10 years and will be cheaper too.
Nonsense, Microsoft will just put lots of PMs and Scrum masters on the task and they’ll have a working reactor in 1 year max.
/s, just in case any PMs are reading this and think it’s totally reasonable
Power density matters. And nuclear is pretty fucking dense haha
… for some applications. Not most tho. Really like 5. Everything else should be solar/wind/hydro
… And cause a lot of pollution and ecological stress, unless you funnel a LARGE amount of money and time into it.
Do you want to argue, that the construction of a nuclear power plant causes significantly less ecological stress and pollution than solar panels and windturbines?
Think about if you really want to claim that as a thing you actually believe in.
I’m just gonna throw some words in a pool.
concrete, steel, space, deforestation, river, 10+ Years construction time, heavy machinery, dust, natural habitats, fuel, mining, waste, noise, cost, france…
Thank you. i rest my case.Goddamnit France. Classic France.
Half of those aren’t even relevant.
The actual construction takes about 4 years, but legal issues such as rules changing and politics, legal issues, and additional planning tend to push this up to 6-15 years in extreme cases. To draw a parallel: building a 1GW windmill farm, such as the Thorsminde off shore windmill farm is estimated to take 5 years of pure construction time, and politics and legal issues have so far added 4 years to this from the day it was announced, giving a total construction time of about 9 years without delays.
Cost wise, Thorsminde is projected to cost 2.1 billion USD, and that’s without running costs, possible delays, or deconstruction costs at its 30 year end of life. The construction of a nuclear plant usually ( as in the projects that have been finished and we know the total construction costs of) costs anywhere from 6 to 9 billion USD. So yes, nuclear is more expensive, as you said.
Of course windmills don’t just pop out of the ground, so heavy machinery will also be required, and the sound of the hammers building the foundation will likely drive away any sound sensitive life in a 100-200 km radius, such as whales. This can be partly mitigated by running the hammers at lower power, adding about 30-50% (might be more, foundations take a long time to build) additional construction time and driving up the price.
The windmills will also change the life of the area dramatically throughout its life, VS nuclear, which requires mines that cause decent damage, but do not pollute in any significant way at the reactor site (unless you pump the waste water from the usually closed first loop directly out to the rivers and sea, or swear on running the power plant without cooling towers during droughts).
Also the resources needed to make a 1GW wind farm are immense, and contrary to nuclear, we can’t currently reuse the waste from deconstruction, which there also is quite a lot more of. Furthermore, maintenance will be hell, as you have much more moving parts (not per windmill, but per farm, which has multiple windmills) as a nuclear plant.
Do you realise that you can also build windmills… where you would put the Power Plant? On Land? And that would reduce the time and cost of construction?
You could also fill barren fields with solar panels and use space that not even a solar plant could use, this in turn also gives animals shade and helps biodiversity and bug species.
And doesnt have a third of its construction cost as running costs forever.
You can also scale wind turbines in minutes. Look at France how much it costs to have nuclear plants not running.In what way can we reuse the nuclear waste?
You do realise how much space windmills would need to produce as much power as a single nuclear plant, right? That is also the reason we try to build them in the water.
And when did I write anything about nuclear waste? I specifically pointed out that I was talking about deconstruction waste, where cooling towers turbines, and general facilities can be reused, and only the core shielding of the nuclear reactor has to be specially disposed of, versus the wings and foundation of windmills, which we don’t really know what to do with right now, so we kinda just bury them wherever and hope it doesn’t come back to bite us later.
You didn’t. I did. What about nuclear waste? It doesnt go away and if we build so much nuclear we also have so much more waste.
The blades can be recycled btw. we just dont do it because we dont have capacity for them.
Which brings me back to the nuclear waste. Oh and Fukushima. Chernobyl. When are we getting rid of those?The amount of waste produced is extremely small for how much power you get, and is dealt with in exactly the same we we deal with literally all of our garbage: put it under ground and call it a day.
Those wind turbines and solar panels also get constructed, and affect a much larger area. It’s not an obvious comparison
Duh, Yes things have to be built. A Windmill is built in a few weeks by way less people and has no risk of exploding into a huge cloud of death.
Obviously building one wind turbine is less disruptive, but you need hundreds to get the same output, and they only work when it’s windy.
It’s always windy. We live on a spinning planet.
Solar needs sun. Nuclear needs water to cool. Hydro needs water.
If you combine solar and wind you can replace many nuclear plants by just using the space we are already using.There are a lot of good arguments for wind, and I’m not arguing against it, but density and consistency are well known issues. You absolutely cannot replace a nuclear plant with a wind farm of the same size and get the same output. That’s not necessarily a bad thing, wind farms can often coexist with other land uses, but that’s still a disruptive environment.
It’s good to put pressure on nuclear, the reason it’s so incredibly safe is because it’s highly regulated, but to completely ignore it is throwing the baby out with the bath water.
The question isn’t “are nuclear plants perfectly safe”, the question is “will adding nuclear plants to our energy portfolio reduce the risks from climate change enough to offset the risks they introduce.”
I think, in that framework, replacing existing coal power plants with modern nuclear reactors is a huge overall benefit.
Wind and solar are great but there’s still a lot of work needed on storage and transmission before they can be viable grid scale. Realistically, saying no to nuclear doesn’t mean more wind, it means more natural gas. And those LNG tankers really are floating bombs.
The ultimate “AI product” will be a videogame that keeps you playing as long as possible. Indefinitely even.
I think we are playing it right now omg
Have you tried dwarf fortress? XD
Its going to be all for nothing at the end of the day.
Elaborate?
Well, there’s only so much training you can do before it has ingested all of Garfield lore and fan creations, so you start over fitting and can’t generate new Garfield material anymore, and at that point, what’s even the point of going on?
We either move on to Sonic, Heathcliff, or let Odie have his day in the sun.
This seems kind of ideal though, computers provide a near constant load (relatively speaking) that combines very well with nuclear energy.
Perhaps we should be asking why we haven’t already been doing this for the past decade?
Anything that gets us closer to thorium/LFTR energy is fine with me.
I have read scifi (Bear : Blood Music, Egan : I forget) where packing enough computation into a small enough space caused reality to warp. So look out!
Small Modular Reactor technology is the future, and it’s really promising.
Self-contained (no onsite refueling), mass produced (cheap, higher quality), and modular (add more for more power, or small enough to power a data center).
Here’s some quick videos from a professor of Nuclear Energy covering topic:
Small Modular Tractors:
Micro-Modular Rectors:
This is the best summary I could come up with:
Training large language models is an incredibly power-intensive process that has an immense carbon footprint.
Now, The Verge reports, Microsoft is betting so big on AI that its pushing forward with a plan to power them using nuclear reactors.
Yes, you read that right; a recent job listing suggests the company is planning to grow its energy infrastructure with the use of small modular reactors (SMR.)
But before Microsoft can start relying on nuclear power to train its AIs, it’ll have plenty of other hurdles to overcome.
Then, it’ll have to figure out how to get its hands on a highly enriched uranium fuel that these small reactors typically require, as The Verge points out.
Nevertheless, the company signed a power purchase agreement with Helion, a fusion startup founded by OpenAI CEO Sam Altman earlier this year, with the hopes of buying electricity from it as soon as 2028.
The original article contains 346 words, the summary contains 147 words. Saved 58%. I’m a bot and I’m open source!
The reason is ultimately irrelevant, but I welcome more nuclear energy.
They could just invest in a solar farm or something, they are just a lot more economical.
Nuclear is okay, but the costs compared to renewables are very high, and you have to put a lot of effort and security into building a reactor, compared to a solar panel that you can basically just put up and replace if it snaps.
You probably know this discussion already through.
Edit: Glad to see a nice instance of the discussion going here.
In their specific use case that won’t really work.
They want to use all of their available property for server racks. Covering the roof with solar won’t give enough power/area for them. A small reactor would use a tiny fraction of the space, and generate several times the power. That’s why it’d be worth the extra cost.
For those who haven’t seen this discussion before, I feel like doing the next step in the dance. Cheers Plex.
It’s important to note that nuclear is capable of satisfying baseload demand, which is particularly important for things like a commercial AI model training facility, which will be scheduled to run at full blast for multiple nines.
Solar+storage is considerably more unreliable than a local power plant (be it coal, gas, hydro, or nuclear). I have solar panels in an area that gets wildfire smoke (i.e. soon to be the entire planet), and visible smoke in the air effectively nullifies solar.
Solar is fantastic for covering the amount of load that is correlated with insolation: for example colocated with facilities that use air-conditioning (which do include data centers, but the processing is driving the power there).
are you arguing solar is more economical than nucleae cause if so youre wrong by a longshot
The people who actually put money into energy projects are signalling their preferences quite clearly. They took a look at nuclear’s long history of cost and schedule overruns, and then invested in the one that can be up and running in six months. The US government has been willing to issue licenses for new nuclear if companies have their shit in order. Nobody is buying.
That was true 20 years ago. You are working off extremely outdated information.