Hacker News Reader

"We" now and "we" ten thousand years from now are very different people. Those in the future will laugh at your concerns for them. It's as if ancient Egyptians would worry about whether we, now, have enough flood plain space to harvest crops.

Nuclear fuel is a non-trivial resource. There is no basis for thinking we will ever find anything better.

Scientific progress is not a law of nature. It needs very specific environment to thrive. Before scientific revolution kicked in the pharaohs would have easily adjusted to europe and elsewhere.

No one can predict whether there is an infinite amount of scientific progress ahead of us or will we hit a wall.

There are plenty of areas to look at for new energy generation just within particle physics alone. Not to mention that we're at the finish line with a net positive fusion reactor (I don't know a single fusion scientist that doesn't think iter will work).

Scientific progress may not be law, but there isn't a thousand year period where humans haven't advanced.

I don't mean to be negative, but we've been hearing fusion is 5 years out for how long?

Point is: you can't really put a deadline on R&D. We won't really know when it works, until, well it does. And who knows how long it will to take commercialize once we do get it working.

I mean, Musk said fully autonomous vehicles in 2015 were easy and would be solved in 2 years. 4 years later, Tesla'd are killing people by driving into highway barriers or broadsiding a semi trailer. Ambitious? Yes. Are we there yet? No. I think autonomous driving is a worthy goal and we should keep working at it. Same with fusion. But do I believe the projected timelines? No. Doubt that I'll see fusion as a power source in my timeline. Maybe never fully autonomous vehicles (I mean trully fully). And I'm still under 40. Maybe I'm overly pessimistic, but I think a lot of people underestimate the true difficulty of some of these projects.

> Musk said fully autonomous vehicles in 2015

Quoting Musk is probably the worst example you can give. His track record for prediction isn't great.

You're right that it is hard to predict how long R&D takes, but governments are different than companies. Companies have a lifespan in decades, possibly a few centuries (if you're REALLY really lucky and pivot a lot). Profits are focused at the yearly level and long shots are decades out. Governments on the other hand generally have much longer lifespans. Decades are short periods of time for them. Investing in things that will pay off in a century is great business here (lots of science and math falls into this category). This is the problem. People are comparing how companies operate to how governments operate. They have completely different agendas, priorities, and operate at completely different time scales.

> I don't mean to be negative, but we've been hearing fusion is 5 years out for how long?

Fusion is an amount of money away more than an amount of time.

Unless you believe markets are completely inefficient I just don’t buy this argument. Think about the reward awaiting the person or company that finally cracks fusion power: trillions of dollars in wealth; the gratitude of an entire planet; instant, everlasting fame, and so on. So, ample incentives already exist. No, the reason we haven’t solved fusion is that it’s a hard problem, and the right genius(es) have yet to come along.

Companies are just too risk averse for expensive moonshots like this.

Construction costs alone for ITER are estimated to be $22 to $65 billion dollars. This doesn't even factor in the cost of actually running it after that or the cost of feeding all the scientists involved in it (which receive government funding through their respective countries).

You can probably count the amount of companies with enough cash lying around to finance this on one hand. And then they'd still need the people for it.

Even once ITER is completed you still don't have a realistic product to sell. ITER is of no use besides getting us one step closer to fusion energy.

And at the end of it we may just realize that fusion is never going to be cost efficient compared to other energy sources.

That's no risk any sane company will take.

As always the free market will wait until all the groundwork has been laid by taxpayer funded scientists, and even then only make a move if there are massive government subsidies - as has happened and is still happening with fission energy.

The free market is much better at low-risk and incremental innovation. Stuff like this? Not so much.

This is exactly right. If we look at history most of the major moonshots have at minimum had a substantial help from the government. Even silicon valley had a lot of help. Places like Bell Labs did as well, and partially forced. We see things now like X, but they're moonshots aren't decades out from being useful. That would be dumb for a company to do, which generally don't last on the other of centuries. But countries do! Even when they don't, the use lives on in the birth of whatever new country. Countries are in the business of succeeding on the time line of centuries, so funding things that takes decades of research and billions of dollars (mind you this is an international effort so that money is spread out over decades AND by many countries) makes a lot of sense. It's a simple fact that countries and companies operate on completely different time frames.

That being said, I have friends in fusion. I know plenty of others that work in various areas of it. No one thinks iter isn't going to work. All the major problems have been solved. The infighting you see is about money, because there isn't much going around for this because the gp is right that this is a difficult problem and we thought it was easier. The argument now is that some think we need iter to prove that it works and that will cause more funding to come in and then we can build the nice fancy power reactors. The other camp is that we don't need iter and that we should put that money into <insert popular method> (most will probably work). This is a very different conversation then what it was 20 years ago. 20 years ago we had neutron problems.

> Companies are just too risk averse

The risk calculation is: expected cost minus expected reward. The expected reward is astronomical (no pun intended!), so you're simply restating my point: fusion is not likely to work regardless of how much money we throw at it. Amazon's R&D annual budget exceeds $20b. That's one company spending half an ITER a year. The private sector has no problem coming up with those sums for ideas it deems promising. The problem is that industry has decided, quite rationally in my opinion, that fusion power is not one of those ideas.

> That's one company spending half an ITER a year.

No it's not even close to the total cost of ITER. It may just pay for the cost of actually physically constructing it.

The R&D that got taxpayer funded scientists to the point where we are able to construct it at all is going to be a multiple of that.

> The problem is that industry has decided, quite rationally in my opinion, that fusion power is not one of those ideas.

So you're saying that things like fission reactors, spaceflight and GPS, treatments for some rare and not-so-rare diseases, the Human Genome Project, ARPANET, everything that came out of CERN, etc. were bad ideas just because no private company could be assed to develop and build those first?

No. It's a fact that all of these were vitally important to get our species where it is today, even though they would have looked like a terrible investment of time and money from the perspective of a private company. Today private companies have happily entered some of those spaces that have been pioneered by government funded scientists.

It's true that many moonshots fail. That's why they're called (sometimes literal) moonshots. But it's important to try anyways so we can advance as a species.

And most of the time we learn a lot even if we fail.

It's always 5 years out until it isn't! :)

Fusion is powered off of hydrogen, the single most abundant element in the universe. Seems to me like the smart bet is that sometime within 3000-300k years humanity will have developed fusion.

Cold, net-positive fusion. Current fusion reactors either require more power than they output or would run far too hot to be sustainable.

While I too am hopeful that humanity will be able to crack cold fusion in the next few hundred millennia, it should be noted that cold fusion is something which has simply not occurred in our universe before. Yes, fusion is the most commonly-occuring process in the universe, but all of it runs hot. It is possible that net-positive cold fusion might never come to pass and we'll have to go back to solar panels so we can exploit the output of already-existing hot fusion from the Sun.

Why would you need cold fusion? Simply net-positive energy output from a (hot) fusion reactor would do the trick.

Of course, there is a risk there that we will hit some fundamental efficiency limitation that simply makes it not worth it. But st least it's not known to be a physical impossibility, like cold fusion is believed to be.

In fact, more than cold fusion, I think there is a reason to hope for warm superconductors at some point in the future.

That depends on how hot we're talking. I'm not up-to-date on the latest hot fusion work (so I might very well be completely wrong here) but I'd have naively assumed that containment would be a significant limitation of having stable hot fusion reactors. Most research fusion reactors I've seen online only run for minutes at most -- and that's with net-negative fusion.

https://www.thedrive.com/the-war-zone/29074/skunk-works-exot...

Basically there's lot of research into how to do containment that's far more efficient than Tokamak reactors that suffer from the problem you mention. A 100 years is a really long time to bet against given how the pace of our innovation is constantly increasing, let alone 3000 years. Keep in mind that thorium has 3x the capacity & we don't even have reactors to process that so even "regular" has a very long shelf-life. Uranium scarcity isn't a thing when compared with global warming & meets our energy needs drastically better than solar/wind & even hydro while having a much more compact ecological foot-print.

From what I’ve been told, the runtime limit is mainly because they have a stringent and tiny limit on the lifetime radiation production of their reactors. These limits are not technological, they are because the reactors are research projects.

I believe the main losses are plasma instabilities, rather than the energy cost of the containment. This is somewhat secondhand knowledge though, I merely used to hang out in a pub with someone who worked in this field.

Those in the near- and medium- future are already lamenting our lack of concern for them :(

This assumes we don't screw ourselves into another dark age via societal breakdown climate disasters. Both events seem increasingly likely within 100 years, let alone 10,000 years.

Humanity in Y12019 might be exactly where we are now.

In that case there's no problem with fuel.

Exactly where we are, but without all of the easily available petrochemical and coal deposits.

This is utterly laughable!

If we can't make any improvements or find an alternative energy source (like fusion, who's fuel is essentially limitless) then we probably don't deserve to live anymore.

The assumption that we'll make absolutely no technological progress in a (at a ridiculous lower bound!!) 3000yr time frame (and that we shouldn't use an energy source because of that) is laughable! Let alone calling something the least renewable because of it. Specifically with nuclear we've made tremendous progress in the last decade alone. There is no evidence to support this kind of thinking. Let alone that humanity won't make any progress.

If we took this kind of thinking to any other kind of problem we'd never make any progress. There's reasons to not use nuclear (I'm pro), but this isn't one of them.

> Put aside the science fiction of asteroid mining nuclear fuel for now

> If we have dreams of ever being able to do mass interstellar travel

We probably should put aside the science fiction of mass interstellar travel for now too then?

Especially given that asteroid mining (or lunar mining) is relatively achievable in the near term compared to interstellar travel...

> If we magically went to 100% nuclear fission for energy (which is not reasonable for personal transportation, at least)

Not quite so fast. A nuclear power plant under the hood of a car may be unfeasible for a long time to come, but what is becoming feasible is electric cars. Right now, a Tesla is still powered by coal in most places for now, but as more renewable energy sources come online, they magically get switched over to being solar powered, and then they become "nuclear powered", if/when that time comes.

There is effectively an unlimited amount of uranium dissolved in ocean water which can be mined at a much higher cost, but which is an insignificant cost in the operation of a nuclear plant. There's no reasonable reason to delay consuming uranium, and absolutely ridiculous to consider saving thorium (not that we can actually consume thorium yet).

We have 300k years of uranium at current usage rates. If we switched over to Thorium the supply is orders of magnitude greater.

What is the foot dragging on the momentum of Thorium generators?

They don't produce plutonium. Department of War (later the Dept of Energy) had a LFTR (liquid fluoride thorium reactor) running at Oak Ridge but the primary goal there was to produce plutonium for the military's nuclear weapons arsenal, not electricity.

In addition to having a far more plentiful fuel source, a LFTR runs at ambient atmospheric pressure so it's not necessary to have a massive containment dome surrounding the reactor in case of a core loop breach. Also, the chemistry of a LFTR tolerates the addition of partially spent uranium fuel (popularly mislabeled as "nuclear waste") which is then consumed in the normal fuel lifecycle of a LFTR. Why this wonder of chemistry isn't more commonly known is beyond me: we could simply solve the Yucca Mountain problem by recycling the partially spent fuel stored there into liquid metal reactors like the LFTRs. The downside is that if you have to scram a LFTR the core chemistry solidifies (it's liquid metal, after all) and restarting is time and energy intensive.

Ultimately, though, because the LFTR didn't produce plutonium for bombs the Department of Defense didn't have interest in running it.

I disagree with this. 233U is part of the Th cycle. You definitely get transuranic elements out, though admittedly less. There is an advantage though, you can produce power and pull out enriched uranium while doing that; an enticement plant that generates power. (Note: we already use reactors that have low plutonium yields)

The real reason that we don't have thorium reactors is that it's difficult. All the kinks aren't worked out. Compound that with that it's hard to implement new reactor designs and the process becomes really slow to work out those kinks.

The answer, unsurprisingly is extremely complicated but certainly isn't just because it doesn't produce Pu, or rather creates low yields. This myth needs to die.

I think it's fair to say our expertise in Uranium cycle fuel came out of the fact it generates substantially more plutonium for nuclear weapons. This was a huge reason it was selected over Thorium initially. Now, it may just be momentum. I argue it's because back in the day, it didn't product Pu.

Also keep in mind that the Uranium cycle is SIGNIFICANTLY easier. I'd argue that that played a major role. But I'm also not denying that weapons didn't play a role.

The fact that as a total percentage of the cost of generating the power Uranium is cheap as heck. All of these complex alternative fuel plans are solving a problem that doesn't currently exist.

If we went all in with fission we still have thousands of years before it's an issue, by which point we'll probably have figured out some even better alternatives.

the fact that they don't really exist yet?

They do! CANDU reactors, in operation all over the world, can use Thorium fuel today.

[1] https://en.wikipedia.org/wiki/CANDU_reactor

Oh dear this ancient repeated never ending debate. Nuclear lost - in cost, in failure modes, in lack of private funding, in plant build time, in risk of enrichment systems... dead horse is dead. Edit. Add the fact the reactors take decades to develop, cool story we have about ten of those decades before really shit has hit irreversible climate destabilization fan.

Tell that to the US Navy which operates dozens of reactors in wartime vessels travelling millions of miles and enduring rough conditions and even combat with one of the best safety records of any industry.

The US has 2 sunk Nuclear submarines (129 and 99 lives lost). With at least USS Thresher being related to a loss of propulsion and thus at least indirectly the reactor. Considering how few where in service and none being sunk in wartime that’s not a great safety record. https://en.m.wikipedia.org/wiki/List_of_sunken_nuclear_subma...

The cause of USS Scorpion’s sinking is unknown so it may be related to the reactor or it may not.

We're discussing safety record for nuclear power specifically, not lost ships for any reason. The USN has about half of all the reactors running by the USA. How is that a "few"?

If you read the page about USS Thresher then it clearly explains it had nothing to do with the reactor and was a "high-pressure water spraying from a broken pipe joint may have shorted out one of the many electrical panels, causing a shutdown ("scram") of the reactor, which in turn caused loss of propulsion" followed by other personnel and procedure issues that led to a sinking.

That quote directly says the loss of reactor power was part of the chain of events leading to a sinking.

Sure it did not blow up, but that does not mean the reactor was safe to use as a critical piece of equipment. Loss of power is a huge deal and the reactor being unable to recover in that situation is a major design flaw.

That's like saying a torpedo hit the sub and the reactor stopped working so it's a reactor problem. You're being disingenuous. The reactor was not the part that malfunctioned.

Again, if you read that page then you'll find several other causes like faulty ballast tanks that didn't eject, isolating the steam system too quickly and eliminating potential energy from the turbines, and inexperienced personnel not restarting the reactor after shutdown. If just the ballast tanks had worked then the sub would be safe for rescue, if the steam was used then it could continue to drive to surface.

According to an official report [1]: "U.S. Nuclear Powered Warships have safely operated for more than 50 years without experiencing any reactor accident or any release of radioactivity that hurt human health or had an adverse effect on marine life. Naval reactors have an outstanding record of over 134 million miles safely steamed on nuclear power, and they have amassed over 5700 reactor-years of safe operation." and this is from several years ago.

1. https://www.mofa.go.jp/region/n-america/us/security/fact0604...

Steam turbines are part of a nuclear reactor. If it had used batteries and Diesel engines there would have been no steam system to fuck up.

I absolutely agree it was not the single cause of failure, it was even recoverable in theory. But, if the reactor had continued to supply power, like large result battery systems tend to, then it would have likely made it. That directly means using a reactor was less safe to put on the sub.

If an aircraft lost all engine power over the ocean because of a defective engine design and thus crashed. Well you can bet the navy would blame the engines, but with nuclear power they care about perception. The loss of a crew is acceptable risk, the loss of nuclear reactors from a public backlash is a loss of capability.

PS: They can and do mitigate this risk by having large battery systems for redundancy, it’s only a few hours of power. But, it can run a redundant electric engine to move the sub, and that’s an important lesson learned.

If it was a diesel-electric then the broken water-pipe would've shorted out the batteries and electrics, and the flooded engine room would've stopped the engines, resulting in the exact same loss of power.

The fact is that the reactor didn't fail, it worked normally. The steam system didn't "fuck up" but was used incorrectly. The part that did fail was the ballast tank safety system. Diesel-electrics aren't a magical answer and have major disadvantages (like the need for air intake in a submarine).

It seems you're arguing that nuclear-powered designs aren't good, which is an entirely different topic than the safety record of the reactors themselves. Bad design of a system is not an argument that that core principles are inherently wrong, which is exactly the same thing any nuclear expert will tell you about Chernobyl.

It’s very possible that a broken pipe would have shorted out all batteries on an equivalent non nuclear sub. But, that would also have been a significant design flaw.

Anyway, I think this is more a case of disagreements about what constitutes a system than about what happened. I would say if the design of a critical system lets people mess up then that’s a failure of the design and the system. Chernobyl was very much a case of operators doing a long series of dumb things as people do. Similarly, the operators caused an issue with the steam system, but that just means the system and it’s design was also faulty. (IMO, the steam system should be considered part of the nuclear power plant as civilian nuclear power plants also use a steam loop for similar reasons.)

I would rate the US Navy highly on nuclear safety operations especially over the last 50 years.

It's the safest form of energy in terms of deaths per terawatt-hour generated by an order of magnitude [1].

[1] https://www.nextbigfuture.com/2011/03/deaths-per-twh-by-ener...

Perhaps the dead horse is the claim that it’s dead? The only thing it’s held back by is ignorance.

3000 years of not worrying about energy by todays standards looks like a fantastic deal to me.

That's also 3,000 more years to figure out how to make a useful fusion reactor, or to figure out how to smoosh neutrons and protons and electrons together to turn uranium into a renewable resource, or to figure out how to blow up stars and harvest fissile material that way.

3,000 years is a long time at our current rate of scientific discovery.

Even if nuclear only lasted 3000 years it would be more than what fossil fuels will last.

And by then we will have many better alternatives

Why would we put aside the science fiction of astroid mining? In 3000 years, it probably won’t be a problem.

And I’m not sure how you think it’s the most non-renewable. How much oil do you think we have left?

The earth can make more oil given enough time. There will never be more fissile material in this solar system.

Enough time is millions of years for the earth to make oil.

If it makes you happy, i’m willing to ration the uranium so the earth’s supply lasts 10,000 years. If our exponential knowledge curve can’t solve the problem by then we might be in trouble.

The singularity should be within the next 10,000 years.

Once you get to the point where you mined all of the fissile material in the solar system you probably have fusion figured out or how to use solar power at large scales.

I'd say we have a reasonable expectation that fusion will be viable this century in the sense of working. Whether it will be cost effective remains to be seen.

it's the most non-renewable resource, but it also goes away whether we use it or not.

It’s going away on the scale of billions of years. Personally, I think it’s much better to use it for space exploration than unusually expensive electrical power.

Now if was even vaguely cost competitive with battery backed up wind and solar then that’s a reasonable argument, but until then I just don’t see the point.

With seawater extraction and breeders there's enough uranium to power the entire world many times over for hundreds of millions of years. Thus, just like the fusion fuel in the sun, uranium fission fuel is effectively as renewable as it gets.

http://ansnuclearcafe.org/2016/10/03/nuclear-power-becomes-c...

For all intents and purposes it is a renewable resource. Breeder reactors allow us to utilize our fissible materials for an absurd amount of time.

Yup. Just like the finite but immense fusion fuel in our nearest star.

Does your figure account for the uranium contained in seawater ?