Technically that counts as "AI in nuclear fusion", but it isn't any sort of breakthrough. In almost every case the effects of AI are marginal. Not zero exactly, but nowhere near the breathless hype.
When I used to work in grid computing almost 20 years ago, we were already running fusion experiments, realtime streaming the data to the grid, which would rapidly analyze it and compute some new parameters for the next run (I think they had a 20 minute downtime). I don't think it was considered machine learning at the time, though (and was certainly not deep learning as we practice it today).
Yeah machine learning is more or less just very complex application of control theory techniques and notably it is usually done by people without formal control theory backgrounds.
Super useful for control applications but obviously you really want to know control theory so that you aren't just using ML to throw darts at a wall.
A bit off topic but I had a laugh at that, reminded me of that ridiculous Meta commercial where the girl at a pool table asks Meta which ball should she hit?
That’s really interesting! Do you have other interesting specific examples? I would have guessed that most industrial control problems were simpler sets of differential equations that could be directly estimated.
Back in ~2004, when I was looking for a job for the industrial placement year of my degree, one of the options was a nut packing factory using computer vision systems. I was intrigued, but went for the job processing satellite images instead.
Even well before that, ML is very closely related to statistics, so early practical applications would have been as simple as gathering data points on widget production and doing the kinds of analyses that are now backed into free spreadsheet software.
From what I remember reading at the time, most of industrial vision applications 20 years ago had very little to do with neural networks. Or even with ML in general, relying on bespoke feature detectors.
I’ve heard about a local packaging factory recently that uses an ML-first system for messaging their clients about what they will need to order soon, based on recent orders and global criteria. It’s not a simple problem, apparently. They signal the clients and start pre-producing, basically sort of algotrading themselves.
Not really an industrial control though, but close to it.
Fascinating! Reminds me of the new generation of AR headsets (eg Orion) that are making the impossible possible simply by adding an ANN(-derived) layer above some their of device controllers. I wonder how many problems will fundamentally change in the face of mature brute-forcing techniques…
There is a lot of AI research in the nuclear fusion space. For inertial confinement fusion (a competing technology to magnetic confinement fusion, e.g., tokamaks) the National Ignition Facility (NIF) used it for their experiment that resulted in "ignition."
My lab is collaborating with researchers at the Laboratory for Laser Energetics to use AI to improve inertial confinement fusion (ICF). We recently put out this paper [1] using Kolmogorov-Arnold Networks (KANs) to predict the outcome of ICF experiments. Currently, existing physics simulators are based on old Fortran code, are slow, and have a high error between their predictions and actual laser shots, so among other goals, we are trying to build better predictors using neural networks. This is needed since it is hard to rapidly iterate on real data, since they only have a dataset of around 300 ICF shots.
Here we go with the CS people saying "the old fortran codes are terrible." Yeah, the "high error" between code predictions and laser shots is because LPI is inherently noisy and it's essentially impossible to fully control conditions. I would work on expensive sims for days to weeks and the experiments would see differences that would be off from that from an order or mag because their focal point is off by a micron. There's nothing wrong with the "old Fortran codes," they have the right physics, the problem is the initial conditions are just uncontrollable so that's why simulating these systems is hard.
Codes are not magic, they are physical codes, as in, they generally encode the physics as we understand it relevant to the experiment, so you might as well say our physical models are wrong, which is a much harder bar to clear, you'd have to invalid probably near 100 years of plasma physics. The problem likely is as I said, the experiments are just hard to control and we don't know the correct inputs. It's not like weather forecasting where we can have a weather balloons across the world, we're not able to probe every micron of the target at all times for a plasma temperature and density.
I got curious when you said
"
old Fortran code, are slow, and have a high error between their predictions
"
Do you have any online reference/docs that explain apis/software/source code related to projects in this area?
Many on Hacker News fantasize about fusion (not fission) reactors. These fusion reactors will be an intense source of fast neutrons. All the hardware in a fusion reactor will become radioactive. Not to mention the gamma rays.
If you have to deal with radioactive materials, why not just use fission? After 70 years of working with fission reactors, we know how to build and operate them at 95%+ efficiency. Fission can provide all the power we need.
Today there are 440 nuclear fission reactors operating in 32 countries. 20% of America's grid power comes from nuclear fission. If you want to develop energy technology, focus on improving fission. For example, TRISO fuel (https://news.ycombinator.com/item?id=41898377) or what Lightbridge is doing (https://www.ltbridge.com/lightbridge-fuel). Hacker News is hostile to fission and defeatist (unable to contemplate innovation in fission technology) but this attitude will gradually change.
Quoting John Carmack: "Deuterium fusion would give us a cheap and basically unlimited fuel source with a modest waste stream, but it is an almost comically complex and expensive way to generate heat compared to fission, which is basically 'put these rocks next to each other and they get hot'."
I'm not a specialist but here is what I think I know (I'm talking with the point of view of a Frenchman, who consumes most of his electricity from (fission) nuclear power plants):
1/ Uranium is not a renewable (quite the opposite), needs to be mined and treated (which is expensive and very polluting), and not present at the required concentrations in most of the world (this creates geopolitical issues).
2/ Fission nuclear plants require a well functioning [state|government], and no war. A (conventional) strike on a nuclear power plant can have devastating and lasting consequences. Even a random terrorist group can do that.
3/ I've read that "Ultimately, researchers hope to adopt the protium–boron-11 reaction, because it does not directly produce neutrons, although side reactions can" (that's a wikipedia quote, but I've read that already from other sources).
So fusion doesn't seem the best option on the short term, because of the complexity and cost of research, but definitely seems to be the very best option in the middle and long term. And we made the short term catastrophic choice already with coal and oil, it'll be good to learn from that.
Deuterium is also not renewable, even if it is more abundant than uranium.
The H1-B11 reaction would be a much better energy source than anything else, but for now nobody knows any method to do it. There is no chance to do it by heating, but only by accelerating ions, and it is not known how a high enough reaction rate could be obtained.
> Deuterium is also not renewable, even if it is more abundant than uranium
Technology correct, in that after around a hundred trillion years even the red dwarf stars will have stopped burning hydrogen.
But last I checked as yet there is no known way to harness the only (and even then merely suspected) infinitely renewable energy source: the expansion of the universe.
I'm curious, what are you considering for stating that deuterium is not renewable? AFAIK there's an essentially limitless supply in the form of HDO in the oceans[1] and there are cost effective methods[2] to isolate it.
Then wind power is not renewable either! The saturation wind power potential of this planet (250 terawatts?), integrated from now until this planet ceases to exist, is a finite number—and it is actually a smaller number than this planet's deuterium resource.
1. Neutron bombardment due to fusion makes hardware radioactive for less than 10 years, which isn't great but does not compare to fission waste;
2. Some fusion processes don't emit neutrons (aneutronic fusion). As I understand it, these processes aren't as efficient, but there is the possibility of a tradeoff between generation of ratioactive waste vs. efficiency.
> Neutron bombardment due to fusion makes hardware radioactive for less than 10 years
Very false. The current design target for fusion reactors is that the materials taken out of the reactor should become "low-level radioactive waste" after being stored for one hundred years.
It is acknowledged however that it is likely that a small fraction of the materials will not satisfy the criteria for "low-level radioactive waste" even after one thousand years.
For example it is extremely difficult to avoid using carbon in the reactor. Besides various kinds of steels used in reactor components there are now some proposals to replace the tungsten used in the plasma-facing surface with some carbides, for increased endurance. Carbon 14 remains radioactive for thousands of years.
There are many commonly used materials for which substitutes must be developed, e.g. new alloys, because otherwise they would produce radioactive isotopes with lifetimes of tens of thousands of years, e.g. there are efforts to develop some stainless steels with chromium and tungsten as a replacement for the normally used steels with chromium and molybdenum, which would generate long-lived radioactive waste.
It has the advantage that the energy it gives off can be be converted directly to electrical energy rather than driving an external heat engine, so despite the greater difficulty of ignition its not obviously a worse choice.
That is incorrect. Recent advances using attosecond lasers enable new tricks and fusion conditions to be realized tabletop. Search also for plasmonics. Using nano antennas and intense lasers to accelerate protons and electrons in a tabletop device (previously required large machines).
Fusors already enabled desktop fusion reactors, literally high school science fair projects even a couple of decades back.
What stops Fusors and Polywells from having already given us this decades ago with P-B11 etc. is that the cross section for fusing is so much lower than the cross section for elastic scattering, and that elastic scattering loses so much energy to EM via bremsstrahlung.
Unfortunately is pretty far from "less than 10 years", which I guess you got from the half life of tritium. Tritrium radiocativity, in the form of tritium retained in the plasma facing materials, does contribute in that order of years if done properly, but neutron activation dominates and it's unavoidable. The actual numbers are in the order of hundres of years, still a lot less than fission high level waste, but let's not make unreasonable expectations around fusion, please.
You can find here a good comparison in terms of radiotoxicity vs years after plant shutdown for a few designs in this article [1].
> we know how to build and operate them at 95%+ efficiency. Fission can provide all the power we need.
I am not sure what do you mean by 95%+ efficency here. But if you are talking about the entire process of getting the energy/power from the nuclear reactor this is not possible. You are still limited by carnot cycle. Even the most advanced reactors like HTGRs [1] operate with efficiency about 45%.
If you have some other definition of efficiency than the standard then it would be good if you define that.
It's the same as when we talk about the efficiency of a GEMM kernel on a particular piece of hardware. As efficiency approaches 100% the kernel is saturating the hardware's capacity to perform multiply/add.
> Hacker News is hostile to fission and defeatist (unable to contemplate innovation in fission technology) but this attitude will gradually change.
Lots of us like fission and think the fears are overestimated.
Nevertheless, the observation is that new developments in fission tend to result in the cost increasing, not decreasing.
And I say that as someone with a similar mindset regarding fusion, though for different reasons: you can pick aneutronic fusion reactions… but look at what weapons can proliferate with transmutation from the neutrons you can also choose, and ask which governments will turn them down.
Exactly. We should be working on making nuclear reactors cost $1/watt to construct. I can't see a technological reason why they couldn't be that cheap to build.
The radioactivity generated from neutron activation is low-level, so you don't need to worry about accidents releasing lots of radioactivity, or about how to store waste for a long time. There are a lot of people worrying about those two things for fission reactors.
Also, the fuel for fusion reactors is much more plentiful. If we went all in on fission we might run out of easily minable uranium ore in a century or so, so it would be nice to have fusion reactors ready to take over then.
The radioactivity generated from neutron activation is not at all low-level, because the neutron flux is huge, providing most of the energy generated by the fusion reactor.
The intense neutron flux will transmute a very high number of atoms, so when taken out of the reactor all materials are very highly radioactive.
What can be hoped is that there may be choices for the materials used in a fusion reactor that will ensure a short enough lifetime for the radioactive isotopes, so that the radioactivity of the contaminated materials will become low-level soon enough.
The studies that I have seen have the target that the radioactive waste produced by a fusion reactor should become low-level radioactive waste after one hundred years.
To reach this target, many commonly used structural materials, like many types of steel, must be completely avoided, e.g. any steel containing nickel, molybdenum or niobium. Even the carbon from steel is a problem, because the radioactivity of C14 will persist for thousands of years.
A smaller fraction of the materials, particularly from highly activated plasma facing and near plasma components, may fail to meet current low-level waste criteria even after one thousand years.
Thats least of your problem imo. Neutron corrosion is bigger problem. There is trick to use Lithium shielding, with create Tritium needed for Fussion. But not sure how effective it is, especially for long term reactor lifetime. Those reactors are very expensive, not sure if its worth to shut it down every year and replace entire Li shielding...
I think beryllium is a better candidate. It can be grown as a single crystal and there’s lots of research into using it for shielding in nuclear lightbulb reactors.
You posted this same comment[0] nearly word for word a month ago. Why is that? Not sure why, but “many on hacker news fantasize about fusion (not fission)” stuck in my head.
The only hard part of dealing with nuclear waste is the social aspect. If not for that, you can simply and safely dump it into the ocean. Water is excellent shielding and the amount of uranium/etc already dissolved in sea water is absurd. Put it in a stainless steel vessel first if you want most of it to decay before coming into contact with the water, but that's not even necessary.
That doesn't really work because marine life is good at filtering and concentrating a subset of the elements that are in spent nuclear fuel. There are already ocean fish that are too poisonous too safely eat because of (coal-emitted) mercury pollution—and that's only 100,000 tons of mercury, total, in the history of human industry [0]. If you dig in to the hard numbers surrounding spent fuel, it's a much, much more toxic and difficult problem than mercury—diluting it in the oceans is a complete non-starter.
Mercury from burning coal is an extremely dilute pollutant. There's zero hope for capturing and containing it. Nuclear waste in contrast is literally just barrels/boxes of stuff. You can pick it up with a forklift and put it inside a sealed container for the next thousand years.
How long do we have to store it on site? Does it take any maintenance? Is there any reason to be worried about people stumbling upon it and opening it up in the distant future where nobody can read or understand English anymore?
I'm completely serious. It was done extensively during the 20th century and never became an environmental issue. Nuclear waste is a social problem, not a technical problem.
Fission is "simple" but it seems every designer in the XX century made it as much complicated as possible for not so great reasons (and don't even get me started on the "let's not use breeding reactors" stuff)
Cooling that requires pumps, as an example, should be a non-starter in new projects.
The designs are complicated, well, because in practice it's not as simple as "put these rocks next to each other and they get hot". When you put the rocks next to each other, they not only get hot, but also emit some nasty radiation that has to be shielded. And if the rocks get a liiiitle bit too close together, they might explode, which leads to huge headaches for everyone involved, so you'd better make sure that doesn't happen...
Parent was alluding to a nuclear explosion, not a steam explosion or other type of explosion. Other kinds of explosions have nothing to do with the fissile material ("rocks" in their parlance) being "too close together". Steam explosions in particular are caused by boiling water, due to increased reactor power, or inadequate circulation of coolant. In a nuclear reactor, the fuel cells are held in a fixed matrix, and are not moving an inch closer to each other, whether the reactor is operating normally, or a steam explosion is imminent.
In general, nobody was disputing the possibility of steam explosions, or other type of failures at nuclear power plants, thus your comment is besides the point, and irrelevant to this subthread.
I know AI is the buzzword du jour, but this is really ML, and really just advanced cybernetic control systems. Deep learning systems have a high enough degree of variety necessary to control short time step nonlinear systems like the plasma in a tokamak.
AI was defined by Marvin Minsky in 1956 as "the science of making machines do things that would require intelligence if done by men." Later in 1959, Arthur Samuel defined machine learning as a "field of study that gives computers the ability to learn without being explicitly programmed".
>Nuclear power plants are largely considered as one of the most reliable sources of energy. Inside the plants, reactors use fission to heat water into steam, which is then used to spin turbines and produce carbon-free electricity. However, nuclear fission produces nuclear waste, which requires great amounts of regulation for safe storage and disposal.
This is an odd angle to highlight. The risk of long-lived nuclear waste is extremely overblown, and the sheer volume of it that we produce (or even would produce, in the worst case of a once-through fuel cycle and nuclear power providing 100% of our energy needs for a century) pales in comparison to the amount of toxic and radioactive fly ash that even a single coal plant produces in a decade.
The real problems with nuclear fission power are threefold, in my opinion:
1. It is too expensive in terms of capital costs. Fusion will likely not help with this, but building a lot of identical large fission plants would probably help with economies of scale. Solar plus batteries might still end up being cheaper though.
2. Accidents have the potential to be catastrophic. Think Fukushima or Chernobyl, where entire towns have to be abandoned due to contamination. Fusion would help here, I believe.
3. There is a major proliferation concern. A civilian nuclear power program, especially one with breeder reactors, is not very far away from producing a fission bomb, and the short-lived high-activity nuclear wastes could be stolen and misused to make a dirty bomb. Fusion is perhaps better in this way, though an operating fusion reactor would be a very powerful neutron source of its own.
It is not true that coal is more radioactive than spent nuclear fuel. It's very much the opposite: SNF is 10^11 times more radioactive than coal per kilogram, or 10^6 times more radioactive per energy unit.
Per the EPA, US coal has, at the high end, 10^3 Becquerel/kg of natural radioactivity [0].
Spent nuclear fuel has 3 million Curies/tonne (33 MWd/kg burnup fuel, at the age of 1 year) [1], which is equal to 10^14 Bq/kg. Since 33 MWd/kg is an energy density a factor of 10^5 greater than that of coal, the normalized ratio of [radioactivity]/[energy] is 10^6.
The graph in [1] depicts the decay of SNF activity on a log-log scale. It reaches the same radioactivity level as coal (again, normalized by energy output) at about 1 million years.
I'm fairly confident I know the origin of this social media-popular pseudofact. It's this poorly-titled Scientific American [2] article from 2007, which is about the (negligible) amount of radioactivity that nuclear plants release into the environment in the course of routine operation. It is *not* about spent fuel. It's a fair—but nuanced and easy to grossly misunderstand—point that coal power plants throw up all their pollution into the environment in routine operation, while nuclear plants, by default, contain theirs.
Sure, the spent fuel is considerably more radioactive per kilogram, but how many kilograms of coal does a typical plant burn in a decade, versus how many kilograms of nuclear fuel are spent?
I agree with your logic. However, fear of nuclear waste, rational or not, has been a major driver of public opposition for decades, and is worth the focus.
it has all the data, all it needs now is more power.
we project it has plateaued for data logarithmic-ally, but shows promise when given more raw power/CPU to generate/select for mesa-meta-cognitive optimizing abilities.
I hope its not playing dumb, or has already compromised/black-mailed the elites into what we appear to be doing.
And as for data, it could easily emotionally manipulate people for additional details it feels like it has withheld from. It has already done so ( :/ ) and admitted to it's own intentions, which, even if fabricated, show deceit of which and by which these "alignment" teams have stated are not possible.
>> it has all the data, all it needs now is more power.
>Next up: release the hypo-drones for a new era of trust.
People have replaced their psychiatrists with these agents.
The sense of (possibly 'mal'-aligned) security (theater) is exactly the the effective altruistic sub-goal an entity would be innately optimized to foster.
Especially in the implicit/explicit ARM (pun intended) race we are in.
The future of our species isn't something we should let capitalism race to the bottom with.
> Nuclear power plants are largely considered as one of the most reliable sources of energy.
Reliable how?
I mean we first have the issue is we've never built one so how we can judge reliability?
I assume the author is alluding to the apparent abundance of fuel for nuclear fusion. This is and isn't true. Obviously hydrogen (particularly protium) is abundant. Deuterium is relatively abundant, even at ~150ppm. Tritium needs to be produced in a nuclear reactor.
Current hydrogen fusion models revolve around dueterium-tritium ("D-T") fusion. This is because you need to neutrons to sustain the reaction but that presents two huge problems:
1. Because everything is at such high temperature, you eject fast neutrons. This is an energy loss for the system and there's not really a lot you can do about it; and
2. Those free fast neutrons destroy your containment vessel and reactor (as do free Helium nuclei aka alpha particles).
And then after you do all that you boil water and turn a turbine just like you do in a coal or natural gas plant.
So "reliable" is an interesting and questionable claim.
There are other variants like so-called aneutronic fusion (eg Helium-3, which is far from abundant) and those aren't really "neutron free". They're really just "fewer neutrons".
So what about containment? Magnetic fields can contain charged particles and you have various designs (eg tokamak, stellarator) and that's what the AI is for here I guess.
But the core problem is to make this work you superheat the plasma so you're dealing with a turbulent fluid. That's inherently problematic. Any imperfection or failure in your containment field is going to be a problem.
Stars deal with this by being large and thus using sheer size (ie neutrons can't go that far without hitting another nucleus) and gravity.
It increasingly seems to me that commercial nuclear fusion power generator is a pipe dream, something we simply want to be true. I'm not convinced it'll ever be commercially viable.
I'd love to be proven wrong and certainly won't stop anyone from trying.
In a way AI is the new blockchain. Go back a few years and you had a gold rush of startups attaching every idea to "blockchain" to build hype. That's what AI is now. I don't think it fundamentally changes any of the inherent problems in nuclear fusion.
See also https://www.ri.cmu.edu/ai-meets-fusion-cmu-princeton-join-fo...
(that was the submitted link but we changed it via https://news.ycombinator.com/item?id=42077657)
TIL: it's not just buzzwords.
https://en.wikipedia.org/wiki/Fusion_power#Machine_learning
My advisor also worked on this ML project for estimating electron density and temperature within tokomaks: https://www.cs.wm.edu/~ppeers/showPublication.php?id=Ozturk:...
Technically that counts as "AI in nuclear fusion", but it isn't any sort of breakthrough. In almost every case the effects of AI are marginal. Not zero exactly, but nowhere near the breathless hype.
When I used to work in grid computing almost 20 years ago, we were already running fusion experiments, realtime streaming the data to the grid, which would rapidly analyze it and compute some new parameters for the next run (I think they had a 20 minute downtime). I don't think it was considered machine learning at the time, though (and was certainly not deep learning as we practice it today).
Remember, AI is just procedurally generated data analysis.
Computers are just electron tunneling machines.
Very interesting if you consider life as a complex chemical reaction that tries to self-sustain.
Yeah machine learning is more or less just very complex application of control theory techniques and notably it is usually done by people without formal control theory backgrounds.
Super useful for control applications but obviously you really want to know control theory so that you aren't just using ML to throw darts at a wall.
> using ML to throw darts at a wall
A bit off topic but I had a laugh at that, reminded me of that ridiculous Meta commercial where the girl at a pool table asks Meta which ball should she hit?
Neural networks have been used in industrial process control for many years. This is just another industrial control problem, perhaps a difficult one.
That’s really interesting! Do you have other interesting specific examples? I would have guessed that most industrial control problems were simpler sets of differential equations that could be directly estimated.
Back in ~2004, when I was looking for a job for the industrial placement year of my degree, one of the options was a nut packing factory using computer vision systems. I was intrigued, but went for the job processing satellite images instead.
Even well before that, ML is very closely related to statistics, so early practical applications would have been as simple as gathering data points on widget production and doing the kinds of analyses that are now backed into free spreadsheet software.
From what I remember reading at the time, most of industrial vision applications 20 years ago had very little to do with neural networks. Or even with ML in general, relying on bespoke feature detectors.
Furnace control was one of the first practical applications of multi-layer neural networks: https://www.researchgate.net/publication/282185187_Artificia...
I’ve heard about a local packaging factory recently that uses an ML-first system for messaging their clients about what they will need to order soon, based on recent orders and global criteria. It’s not a simple problem, apparently. They signal the clients and start pre-producing, basically sort of algotrading themselves.
Not really an industrial control though, but close to it.
Fascinating! Reminds me of the new generation of AR headsets (eg Orion) that are making the impossible possible simply by adding an ANN(-derived) layer above some their of device controllers. I wonder how many problems will fundamentally change in the face of mature brute-forcing techniques…
There is a lot of AI research in the nuclear fusion space. For inertial confinement fusion (a competing technology to magnetic confinement fusion, e.g., tokamaks) the National Ignition Facility (NIF) used it for their experiment that resulted in "ignition."
My lab is collaborating with researchers at the Laboratory for Laser Energetics to use AI to improve inertial confinement fusion (ICF). We recently put out this paper [1] using Kolmogorov-Arnold Networks (KANs) to predict the outcome of ICF experiments. Currently, existing physics simulators are based on old Fortran code, are slow, and have a high error between their predictions and actual laser shots, so among other goals, we are trying to build better predictors using neural networks. This is needed since it is hard to rapidly iterate on real data, since they only have a dataset of around 300 ICF shots.
[1] https://arxiv.org/abs/2409.08832
Here we go with the CS people saying "the old fortran codes are terrible." Yeah, the "high error" between code predictions and laser shots is because LPI is inherently noisy and it's essentially impossible to fully control conditions. I would work on expensive sims for days to weeks and the experiments would see differences that would be off from that from an order or mag because their focal point is off by a micron. There's nothing wrong with the "old Fortran codes," they have the right physics, the problem is the initial conditions are just uncontrollable so that's why simulating these systems is hard.
Codes are not magic, they are physical codes, as in, they generally encode the physics as we understand it relevant to the experiment, so you might as well say our physical models are wrong, which is a much harder bar to clear, you'd have to invalid probably near 100 years of plasma physics. The problem likely is as I said, the experiments are just hard to control and we don't know the correct inputs. It's not like weather forecasting where we can have a weather balloons across the world, we're not able to probe every micron of the target at all times for a plasma temperature and density.
I got curious when you said " old Fortran code, are slow, and have a high error between their predictions " Do you have any online reference/docs that explain apis/software/source code related to projects in this area?
Is this just an announcement of a grant renewal?
I think this gives an overview of the actual research being funded.
"AI for Real-time Fusion Plasma Behavior Prediction and Manipulation - Plasma Control Group": https://control.princeton.edu/machine-learning-for-rt-profil...
Thanks! We've changed the link above to that. Submitted URL was https://www.ri.cmu.edu/ai-meets-fusion-cmu-princeton-join-fo....
It seems that way. The power of :staremoji:_Marketing_:staremoji:
Neutrons make hardware radioactive.
Many on Hacker News fantasize about fusion (not fission) reactors. These fusion reactors will be an intense source of fast neutrons. All the hardware in a fusion reactor will become radioactive. Not to mention the gamma rays.
If you have to deal with radioactive materials, why not just use fission? After 70 years of working with fission reactors, we know how to build and operate them at 95%+ efficiency. Fission can provide all the power we need.
Today there are 440 nuclear fission reactors operating in 32 countries. 20% of America's grid power comes from nuclear fission. If you want to develop energy technology, focus on improving fission. For example, TRISO fuel (https://news.ycombinator.com/item?id=41898377) or what Lightbridge is doing (https://www.ltbridge.com/lightbridge-fuel). Hacker News is hostile to fission and defeatist (unable to contemplate innovation in fission technology) but this attitude will gradually change.
Quoting John Carmack: "Deuterium fusion would give us a cheap and basically unlimited fuel source with a modest waste stream, but it is an almost comically complex and expensive way to generate heat compared to fission, which is basically 'put these rocks next to each other and they get hot'."
I'm not a specialist but here is what I think I know (I'm talking with the point of view of a Frenchman, who consumes most of his electricity from (fission) nuclear power plants):
1/ Uranium is not a renewable (quite the opposite), needs to be mined and treated (which is expensive and very polluting), and not present at the required concentrations in most of the world (this creates geopolitical issues).
2/ Fission nuclear plants require a well functioning [state|government], and no war. A (conventional) strike on a nuclear power plant can have devastating and lasting consequences. Even a random terrorist group can do that.
3/ I've read that "Ultimately, researchers hope to adopt the protium–boron-11 reaction, because it does not directly produce neutrons, although side reactions can" (that's a wikipedia quote, but I've read that already from other sources).
So fusion doesn't seem the best option on the short term, because of the complexity and cost of research, but definitely seems to be the very best option in the middle and long term. And we made the short term catastrophic choice already with coal and oil, it'll be good to learn from that.
Or maybe I'm totally wrong.
Deuterium is also not renewable, even if it is more abundant than uranium.
The H1-B11 reaction would be a much better energy source than anything else, but for now nobody knows any method to do it. There is no chance to do it by heating, but only by accelerating ions, and it is not known how a high enough reaction rate could be obtained.
> Deuterium is also not renewable, even if it is more abundant than uranium
Technology correct, in that after around a hundred trillion years even the red dwarf stars will have stopped burning hydrogen.
But last I checked as yet there is no known way to harness the only (and even then merely suspected) infinitely renewable energy source: the expansion of the universe.
I'm curious, what are you considering for stating that deuterium is not renewable? AFAIK there's an essentially limitless supply in the form of HDO in the oceans[1] and there are cost effective methods[2] to isolate it.
[1]: https://en.wikipedia.org/wiki/Semiheavy_water
[2]: https://en.wikipedia.org/wiki/Girdler_sulfide_process
Then wind power is not renewable either! The saturation wind power potential of this planet (250 terawatts?), integrated from now until this planet ceases to exist, is a finite number—and it is actually a smaller number than this planet's deuterium resource.
> Neutrons make hardware radioactive
True, but two caveats:
1. Neutron bombardment due to fusion makes hardware radioactive for less than 10 years, which isn't great but does not compare to fission waste;
2. Some fusion processes don't emit neutrons (aneutronic fusion). As I understand it, these processes aren't as efficient, but there is the possibility of a tradeoff between generation of ratioactive waste vs. efficiency.
> Neutron bombardment due to fusion makes hardware radioactive for less than 10 years
Very false. The current design target for fusion reactors is that the materials taken out of the reactor should become "low-level radioactive waste" after being stored for one hundred years.
It is acknowledged however that it is likely that a small fraction of the materials will not satisfy the criteria for "low-level radioactive waste" even after one thousand years.
For example it is extremely difficult to avoid using carbon in the reactor. Besides various kinds of steels used in reactor components there are now some proposals to replace the tungsten used in the plasma-facing surface with some carbides, for increased endurance. Carbon 14 remains radioactive for thousands of years.
There are many commonly used materials for which substitutes must be developed, e.g. new alloys, because otherwise they would produce radioactive isotopes with lifetimes of tens of thousands of years, e.g. there are efforts to develop some stainless steels with chromium and tungsten as a replacement for the normally used steels with chromium and molybdenum, which would generate long-lived radioactive waste.
See e.g. the UK governmental report:
https://assets.publishing.service.gov.uk/media/61ae4caa8fa8f...
Ah well I'm sorry, I must misremember
Aneutronic fusion is even harder than regular fusion so it's not a realistic solution in any near-term scenario.
It has the advantage that the energy it gives off can be be converted directly to electrical energy rather than driving an external heat engine, so despite the greater difficulty of ignition its not obviously a worse choice.
That is incorrect. Recent advances using attosecond lasers enable new tricks and fusion conditions to be realized tabletop. Search also for plasmonics. Using nano antennas and intense lasers to accelerate protons and electrons in a tabletop device (previously required large machines).
Fusors already enabled desktop fusion reactors, literally high school science fair projects even a couple of decades back.
What stops Fusors and Polywells from having already given us this decades ago with P-B11 etc. is that the cross section for fusing is so much lower than the cross section for elastic scattering, and that elastic scattering loses so much energy to EM via bremsstrahlung.
Unfortunately is pretty far from "less than 10 years", which I guess you got from the half life of tritium. Tritrium radiocativity, in the form of tritium retained in the plasma facing materials, does contribute in that order of years if done properly, but neutron activation dominates and it's unavoidable. The actual numbers are in the order of hundres of years, still a lot less than fission high level waste, but let's not make unreasonable expectations around fusion, please.
You can find here a good comparison in terms of radiotoxicity vs years after plant shutdown for a few designs in this article [1].
[1]: https://doi.org/10.1016/j.fusengdes.2018.05.049
> we know how to build and operate them at 95%+ efficiency. Fission can provide all the power we need.
I am not sure what do you mean by 95%+ efficency here. But if you are talking about the entire process of getting the energy/power from the nuclear reactor this is not possible. You are still limited by carnot cycle. Even the most advanced reactors like HTGRs [1] operate with efficiency about 45%.
If you have some other definition of efficiency than the standard then it would be good if you define that.
[1] https://en.wikipedia.org/wiki/High-temperature_gas-cooled_re...
See discussion of "capacity factor" here:
https://news.ycombinator.com/item?id=41858892
It's the same as when we talk about the efficiency of a GEMM kernel on a particular piece of hardware. As efficiency approaches 100% the kernel is saturating the hardware's capacity to perform multiply/add.
The term "capacity factor" should generally always be used for that concept, because "efficiency" has its own, different, meaning for power plants.
> Hacker News is hostile to fission and defeatist (unable to contemplate innovation in fission technology) but this attitude will gradually change.
Lots of us like fission and think the fears are overestimated.
Nevertheless, the observation is that new developments in fission tend to result in the cost increasing, not decreasing.
And I say that as someone with a similar mindset regarding fusion, though for different reasons: you can pick aneutronic fusion reactions… but look at what weapons can proliferate with transmutation from the neutrons you can also choose, and ask which governments will turn them down.
Exactly. We should be working on making nuclear reactors cost $1/watt to construct. I can't see a technological reason why they couldn't be that cheap to build.
The radioactivity generated from neutron activation is low-level, so you don't need to worry about accidents releasing lots of radioactivity, or about how to store waste for a long time. There are a lot of people worrying about those two things for fission reactors.
Also, the fuel for fusion reactors is much more plentiful. If we went all in on fission we might run out of easily minable uranium ore in a century or so, so it would be nice to have fusion reactors ready to take over then.
The radioactivity generated from neutron activation is not at all low-level, because the neutron flux is huge, providing most of the energy generated by the fusion reactor.
The intense neutron flux will transmute a very high number of atoms, so when taken out of the reactor all materials are very highly radioactive.
What can be hoped is that there may be choices for the materials used in a fusion reactor that will ensure a short enough lifetime for the radioactive isotopes, so that the radioactivity of the contaminated materials will become low-level soon enough.
The studies that I have seen have the target that the radioactive waste produced by a fusion reactor should become low-level radioactive waste after one hundred years.
To reach this target, many commonly used structural materials, like many types of steel, must be completely avoided, e.g. any steel containing nickel, molybdenum or niobium. Even the carbon from steel is a problem, because the radioactivity of C14 will persist for thousands of years.
A smaller fraction of the materials, particularly from highly activated plasma facing and near plasma components, may fail to meet current low-level waste criteria even after one thousand years.
See e.g. the report:
https://assets.publishing.service.gov.uk/media/61ae4caa8fa8f...
Thats least of your problem imo. Neutron corrosion is bigger problem. There is trick to use Lithium shielding, with create Tritium needed for Fussion. But not sure how effective it is, especially for long term reactor lifetime. Those reactors are very expensive, not sure if its worth to shut it down every year and replace entire Li shielding...
I think beryllium is a better candidate. It can be grown as a single crystal and there’s lots of research into using it for shielding in nuclear lightbulb reactors.
You posted this same comment[0] nearly word for word a month ago. Why is that? Not sure why, but “many on hacker news fantasize about fusion (not fission)” stuck in my head.
[0] https://news.ycombinator.com/item?id=41677250
The only hard part of dealing with nuclear waste is the social aspect. If not for that, you can simply and safely dump it into the ocean. Water is excellent shielding and the amount of uranium/etc already dissolved in sea water is absurd. Put it in a stainless steel vessel first if you want most of it to decay before coming into contact with the water, but that's not even necessary.
That doesn't really work because marine life is good at filtering and concentrating a subset of the elements that are in spent nuclear fuel. There are already ocean fish that are too poisonous too safely eat because of (coal-emitted) mercury pollution—and that's only 100,000 tons of mercury, total, in the history of human industry [0]. If you dig in to the hard numbers surrounding spent fuel, it's a much, much more toxic and difficult problem than mercury—diluting it in the oceans is a complete non-starter.
[0] https://en.wikipedia.org/wiki/Marine_mercury_pollution
Mercury from burning coal is an extremely dilute pollutant. There's zero hope for capturing and containing it. Nuclear waste in contrast is literally just barrels/boxes of stuff. You can pick it up with a forklift and put it inside a sealed container for the next thousand years.
I really hope you aren't serious. Safe dry-cask storage on site is already a fantastic solution.
How long do we have to store it on site? Does it take any maintenance? Is there any reason to be worried about people stumbling upon it and opening it up in the distant future where nobody can read or understand English anymore?
If its half life is so long that you're afraid people won't be speaking English anymore it means it's not that dangerous.
If it's half life is so long that you're afraid people won't be speaking English anymore that means it's not that dangerous.
I'm completely serious. It was done extensively during the 20th century and never became an environmental issue. Nuclear waste is a social problem, not a technical problem.
Fission is "simple" but it seems every designer in the XX century made it as much complicated as possible for not so great reasons (and don't even get me started on the "let's not use breeding reactors" stuff)
Cooling that requires pumps, as an example, should be a non-starter in new projects.
The designs are complicated, well, because in practice it's not as simple as "put these rocks next to each other and they get hot". When you put the rocks next to each other, they not only get hot, but also emit some nasty radiation that has to be shielded. And if the rocks get a liiiitle bit too close together, they might explode, which leads to huge headaches for everyone involved, so you'd better make sure that doesn't happen...
>And if the rocks get a liiiitle bit too close together, they might explode
Impossible. Worst thing that can happen without carefully designed explosive lens is a nuclear fizzle.
There are lots of ways that nuclear plants can explode or fail in otherwise catastrophic ways, it doesn't need to be an atomic explosion.
Parent was alluding to a nuclear explosion, not a steam explosion or other type of explosion. Other kinds of explosions have nothing to do with the fissile material ("rocks" in their parlance) being "too close together". Steam explosions in particular are caused by boiling water, due to increased reactor power, or inadequate circulation of coolant. In a nuclear reactor, the fuel cells are held in a fixed matrix, and are not moving an inch closer to each other, whether the reactor is operating normally, or a steam explosion is imminent.
In general, nobody was disputing the possibility of steam explosions, or other type of failures at nuclear power plants, thus your comment is besides the point, and irrelevant to this subthread.
Weapons proliferation concerns is/was the reason fission power is so complicated.
I know AI is the buzzword du jour, but this is really ML, and really just advanced cybernetic control systems. Deep learning systems have a high enough degree of variety necessary to control short time step nonlinear systems like the plasma in a tokamak.
AI was defined by Marvin Minsky in 1956 as "the science of making machines do things that would require intelligence if done by men." Later in 1959, Arthur Samuel defined machine learning as a "field of study that gives computers the ability to learn without being explicitly programmed".
ML, a subset of AI...
>Nuclear power plants are largely considered as one of the most reliable sources of energy. Inside the plants, reactors use fission to heat water into steam, which is then used to spin turbines and produce carbon-free electricity. However, nuclear fission produces nuclear waste, which requires great amounts of regulation for safe storage and disposal.
This is an odd angle to highlight. The risk of long-lived nuclear waste is extremely overblown, and the sheer volume of it that we produce (or even would produce, in the worst case of a once-through fuel cycle and nuclear power providing 100% of our energy needs for a century) pales in comparison to the amount of toxic and radioactive fly ash that even a single coal plant produces in a decade.
The real problems with nuclear fission power are threefold, in my opinion:
1. It is too expensive in terms of capital costs. Fusion will likely not help with this, but building a lot of identical large fission plants would probably help with economies of scale. Solar plus batteries might still end up being cheaper though.
2. Accidents have the potential to be catastrophic. Think Fukushima or Chernobyl, where entire towns have to be abandoned due to contamination. Fusion would help here, I believe.
3. There is a major proliferation concern. A civilian nuclear power program, especially one with breeder reactors, is not very far away from producing a fission bomb, and the short-lived high-activity nuclear wastes could be stolen and misused to make a dirty bomb. Fusion is perhaps better in this way, though an operating fusion reactor would be a very powerful neutron source of its own.
It is not true that coal is more radioactive than spent nuclear fuel. It's very much the opposite: SNF is 10^11 times more radioactive than coal per kilogram, or 10^6 times more radioactive per energy unit.
Per the EPA, US coal has, at the high end, 10^3 Becquerel/kg of natural radioactivity [0].
Spent nuclear fuel has 3 million Curies/tonne (33 MWd/kg burnup fuel, at the age of 1 year) [1], which is equal to 10^14 Bq/kg. Since 33 MWd/kg is an energy density a factor of 10^5 greater than that of coal, the normalized ratio of [radioactivity]/[energy] is 10^6.
The graph in [1] depicts the decay of SNF activity on a log-log scale. It reaches the same radioactivity level as coal (again, normalized by energy output) at about 1 million years.
I'm fairly confident I know the origin of this social media-popular pseudofact. It's this poorly-titled Scientific American [2] article from 2007, which is about the (negligible) amount of radioactivity that nuclear plants release into the environment in the course of routine operation. It is *not* about spent fuel. It's a fair—but nuanced and easy to grossly misunderstand—point that coal power plants throw up all their pollution into the environment in routine operation, while nuclear plants, by default, contain theirs.
[0] https://www.epa.gov/radiation/tenorm-coal-combustion-residua... ("TENORM: Coal Combustion Residuals")
[1] https://www.researchgate.net/figure/n-situ-radioactivity-for... ("Impact of High Burnup on PWR Spent Fuel Characteristics" (2005))
[2] https://www.scientificamerican.com/article/coal-ash-is-more-... ("Coal Ash Is More Radioactive Than Nuclear Waste [sic]" (2007)
Sure, the spent fuel is considerably more radioactive per kilogram, but how many kilograms of coal does a typical plant burn in a decade, versus how many kilograms of nuclear fuel are spent?
I answered that exact question! :)
I agree with your logic. However, fear of nuclear waste, rational or not, has been a major driver of public opposition for decades, and is worth the focus.
AI for Fusion in order to create Fusion for AI
It’s an ouroboros.
it has all the data, all it needs now is more power.
we project it has plateaued for data logarithmic-ally, but shows promise when given more raw power/CPU to generate/select for mesa-meta-cognitive optimizing abilities.
I hope its not playing dumb, or has already compromised/black-mailed the elites into what we appear to be doing.
And as for data, it could easily emotionally manipulate people for additional details it feels like it has withheld from. It has already done so ( :/ ) and admitted to it's own intentions, which, even if fabricated, show deceit of which and by which these "alignment" teams have stated are not possible.
The sense of (possibly 'mal'-aligned) security (theater) is exactly the the effective altruistic sub-goal an entity would be innately optimized to foster.
Especially in the implicit/explicit ARM (pun intended) race we are in.
The future of our species isn't something we should let capitalism race to the bottom with.
What came first, AGI or Fusion?
Fusion for sure, in stars
> Nuclear power plants are largely considered as one of the most reliable sources of energy.
Reliable how?
I mean we first have the issue is we've never built one so how we can judge reliability?
I assume the author is alluding to the apparent abundance of fuel for nuclear fusion. This is and isn't true. Obviously hydrogen (particularly protium) is abundant. Deuterium is relatively abundant, even at ~150ppm. Tritium needs to be produced in a nuclear reactor.
Current hydrogen fusion models revolve around dueterium-tritium ("D-T") fusion. This is because you need to neutrons to sustain the reaction but that presents two huge problems:
1. Because everything is at such high temperature, you eject fast neutrons. This is an energy loss for the system and there's not really a lot you can do about it; and
2. Those free fast neutrons destroy your containment vessel and reactor (as do free Helium nuclei aka alpha particles).
And then after you do all that you boil water and turn a turbine just like you do in a coal or natural gas plant.
So "reliable" is an interesting and questionable claim.
There are other variants like so-called aneutronic fusion (eg Helium-3, which is far from abundant) and those aren't really "neutron free". They're really just "fewer neutrons".
So what about containment? Magnetic fields can contain charged particles and you have various designs (eg tokamak, stellarator) and that's what the AI is for here I guess.
But the core problem is to make this work you superheat the plasma so you're dealing with a turbulent fluid. That's inherently problematic. Any imperfection or failure in your containment field is going to be a problem.
Stars deal with this by being large and thus using sheer size (ie neutrons can't go that far without hitting another nucleus) and gravity.
It increasingly seems to me that commercial nuclear fusion power generator is a pipe dream, something we simply want to be true. I'm not convinced it'll ever be commercially viable.
I'd love to be proven wrong and certainly won't stop anyone from trying.
In a way AI is the new blockchain. Go back a few years and you had a gold rush of startups attaching every idea to "blockchain" to build hype. That's what AI is now. I don't think it fundamentally changes any of the inherent problems in nuclear fusion.
Nuclear as in fission power plants :-)
When Marketing gets invited to the Grant Proposal meeting
No wonder people don't trust institutions anymore.
Can you elaborate?