Egregoros

@nanook @fluffy All these reactors running 50 years and

I'm in favor of thorium salt reactors. I like them. I'm just not super worried about something that's only happened to one reactor after it got hit with an earthquake two orders of magnitude higher than it was supposed to.

Friend of mine worked in an oil refinery a while and I think *anything* is safer than California's oil refineries.

I support your efforts but I remain unconvinced that conventional reactors are so terrible and thorium-salt reactors are still in the design phase.

@cjd @bajax @tard @nanook @p
>you're saying safety of nuclear power is intractable
caleb i am not making such an assertion. in my post, in the brief part where i did mention safety issues, that was mentioned to illustrate the straw man nanook (unknowningly?) constructed.

i wrote that the safety of nuclear power is contested, this is a very weak assertion, we need only find some nontrivial contest. but even if you ill like this, i wrote that part ("which, as far... contested?") as dramatic flair anyway, the post as written loses nothing if you omit that sentence.


ultimately the issue with nuclear power is that it several order of magnitude more expensive to construct, unless you are sitting on huge piles of cash you will be financing it, this means you will raise power costs to pay off the loan, the prices per watt hour then end up flat or higher, it is just a wealth transfer from the state to bankers.

if you have a lot of cash, you could build them and it would be a good idea, at least compared to just letting the cash sit around, but nation states are not apple or mihoyo, come to think of it didn't mihoyo build a fusion reactor? that was quite entertaining, i wonder about the details there.

@cjd @bajax @tard @nanook @p
those indian numbers are certainly surprising, i will make a note to check what is going on. there certainly exists a possibility that india of all places has innovated and thrown out the paradigm. it deserves a closer look.

>Building nuclear plants is not that expensive, $2000 per KW is competitive with coal plants.
when you write 2000 per kw, this is neither the cost[1] , nor is this competitive with coal [2].

>engineering problems
indeed, i hope that things do change, it would be great for everyone to have cheap power.

1. on your own chart, ending in 1990, the most recent number is ~8000.
2. coal is ~800 per kw

@cjd @bajax @nanook @p @tard
i will also remark, since i made a coffee and thought about it, that the 800 is in modern dollars, that chart is in 2010 dollars, this 800 for coal should be about 500 in 2010 dollars, meanwhile the chart goes from about 1000 to 8000 in 25 years, it has been 35 years since 1990, so plotting a linear trend in the chart, we would (naively) expect the cost to be about 18000. this is about 35x coal,

of course this is just the most naive way to compare it, but i felt that it was necessary to at least make a naive comparison

Yesterday I went to ChatGPT to try to get down to the first-principles cost of a reactor, and as a result I've moderated my opinion somewhat.

My original position was that nuclear is *cheap* because getting rocks to get hot is easy, if you're permitted to buy said rocks. And the only reason it's not ubiquitous is regulation and bullshit.

GPT's claim is that there are two true first principles problems:
1. Neutron economy as cores are scaled down.
2. Moving heat away from the core as cores are scaled up.

This makes enough sense to me, and I will reform my position the the following: "Nuclear should be relatively cheap, but there are engineering problems to solve, and they are not solved because of regulation and bullshit".

---

> 800 for coal

Okay I'll take your number, 2000 was GPT's estimate and it sounds like you know this for a fact.

> meanwhile the chart goes from about 1000 to 8000 in 25 years

The point is that nothing's supposed to get MORE expensive with time, so it means the US got lost in the weeds of over-regulation and loss of will.

There's another chart I can't find with reactors from different countries, and you see that China today is pushing the bottom of those 1970s numbers. Picrel is the same story.

Now if $1700 is the cost of a PWR, and China has a molten salt thorium breeder (they have an experimental one and plans to build a 100MW small production unit), then we should expect that getting rid of the whole pressure vessel and all of the crap that comes with it will at least halve the cost, if not quarter it. So there should come a time when coal becomes uncompetitive - and then at that point, economies of scale will drive the cost down probably another 50% again and coal will be very dead.

@cjd @bajax @tard @nanook @p
>The point is that nothing's supposed to get MORE expensive with time, so it means the US got lost in the weeds of over-regulation and loss of will.
i am not sure this is a true statement. plenty of things get more expensive in time. furthermore, putting that aside, this alone would not conclude over-regulation. in fact it is neither necessary nor sufficient to conclude it.

>"Nuclear should be relatively cheap, but there are engineering problems to solve, and they are not solved
certainly nobody could disagree. that was the original promise of nuclear power. cheap, ubiquitous, reliable. the basic thesis is solid, but the details are muddled. actually, long term, my hypothesis is that this sort of atomic power will dominate: we just need to innovate.
>they are not solved because of regulation and bullshit".
do you have a "smoking gun" for this?

>China today is pushing the bottom of those 1970s numbers
can you pro-rate that to us construction costs? that is, take the cost of an identical building in china and usa, and look at the ratio. whatever is easily available, airport, high rise, whatever, just the same building. i think that chinese construction in general is cheaper, but tis way we can estimate its equivalent, i am interested in this pro-rated cost.

> my hypothesis is that this sort of atomic power will dominate

My original point was that power prices will come down by an order of magnitude (i.e. 10-20 cents -> 1-2 cents), and then after this, petroleum will be more uncompetitive vs. electric transportation.

If nuclear power dominates in the future as you say, and power demand (datacenters) continues on it's current trajectory, then I don't think this is unrealistic at all.

Nuclear is really interesting because its ongoing operating cost is so low that over time, it can eventually push prices into a whole other dimension. But demand growth is key because without it, you can't build supply or else you get a glut and then you're unable to pay for any of your reactors.

> i think that chinese construction in general is cheaper

It's an interesting question. I suspect some things are reasonably competitive (e.g. a car factory in Alabama) but some things are essentially impossible to do in the US - the California high speed rail project comes to mind. But that's kind of my point: If the US can't build nuclear reactors, this doesn't mean nuclear reactors are actually that expensive.