>Intelligence too cheap to meter is well within grasp
And also:
>cost of intelligence should eventually converge to near the cost of electricity.
Which is a meter-worthy resource. So intelligence effect on people's lives is in the order of magnitude of one second of a toaster use each day, in present value. This begs the question: what could you do with a toaster-second say 5 years from today?
Unsurprisingly, this scared the crap out of the fossil fuel industry in the US and countries like Russia that are net exporters of fossil fuels, so they've spent decades lobbying to bind nuclear plant construction up in red tape to prevent them being built and funding anti-nuclear propaganda.
You can see a lot of the same attempts being made with AI, proposals to ban it or regulate it etc., or make sure only large organizations have it.
But the difference is that power plants are inherently local. If the US makes it uneconomical to build new nuclear plants, US utility customers can't easily get electricity from power plants in China. That isn't really how it works with AI.
The primary cost of a traditional plant is fuel. However a nuclear plant needs a tad more (qualified) oversight than a coal plant.
In the same way that the navy has specialists for running nuclear propulsion systems versus crew needed for diesel engines. Not surprisingly nuclear engines cost "more than fuel".
That cost may end up being insignificant in the long run, but cost is not zero. And shortages of staff would matter (like it does with ATC at the moment.)
Construction cost should be much lower than it is, but I don't think it'll be as cheap as say coal or diesel. The nature of the fuel would always require some extras, if only because the penalty-for-failure is so high. There's a difference between a coal-plant catastrophe and Chernobyl.
So there are costs gor running nuclear, I don't think it necessarily gets "too cheap to measure".
Solar is on its way to do something incredibly disruptive because it's the same "too cheap to meter" but only when the sun is shining, and then you still need the independent capacity to supply power from something else when it isn't. So now instead of "you pay ~$0.12/kWh all the time" you have a situation where power during sunshine hours is basically free but power at other times costs dramatically more than it used to because the infrastructure to supply that power has to recover its costs over significantly less usage.
[0] https://www.mdpi.com/1996-1073/10/12/2169 - Fig 3
This is conjecture. If you wanted to establish this, you would have to show that cost of (skilled) labor was unchanging or negligible.
It is also important to consider that nuclear power deaths/damages are much more localized and traceable than excess deaths from air pollution, and thus much less acceptable to the voting population-- you could argue that this should not make any difference (I disagree), but I don't want to digress here too much.
> we're all lucky that the Chinese chose a different tack to the West's policy of energy failure
What do you believe that is? Because from my point of view, China generates a negligible amount of electricity from nuclear power (<5%), this is not going to change within the next decades, and the main "purpose" from what I can tell is to in-house reactor/turbine know-how (instead of relying on Alstom/Siemesn).
> It is also important to consider that nuclear power deaths/damages...
Maybe you can answer this for me - what deaths and damages? So far I've never been able to pin down any actual death or damage to a nuclear meltdown. I'm sure there are some, but most of the actual attempts to quantify it require appealing to hypothetical deaths and damages that no-one can specifically point to, or tiny numbers that are irrelevant to industrial policy.
I know people who lived in a town next to a lead-zinc mine. That appears to be about as bad as a nuclear crisis from what I can gather and it doesn't seem to be causing anyone undue stress. We're still using lead and zinc. People still live in the town.
> What do you believe that is?
They're building reactors. https://en.wikipedia.org/wiki/List_of_commercial_nuclear_rea... is a happy tale of new and planned plants.
Some of them are really cool too, there is one by the Gobi desert, apparently to prove that they don't need to use water as a coolant.
> Those are constant dollars. Are you claiming that the real cost of labour went up 4- to 8-fold in the nuclear industry?
I'm not saying that is the only effect. But if you want to blame "onerous regulations" for nuclear power being so expensive, then you have to at least put bounds on other factors driving cost (and you also have to show those relaxed regulations would not have led to a significant amount of additional incidents; the current rate of 2 meltdowns for ~500ish reactor sites is already problematic enough).
I personally think that collective standards regarding risk and pollution have risen significantly since the early days of nuclear power, and a lot of things that were done back then would be considered unacceptably reckless/negligient nowadays (=> Hanford became a superfund site for a reason), so claiming that those historical costs are repeatable seems very dubious in the first place to me.
Ok, so say they don't evacuate the city. What is the actual risk here that we're talking about?
> I'm not saying that is the only effect.
What other effect are you considering? The materials aren't getting that more expensive and the labour will cost similar too. There aren't a lot of options left apart from regulatory changes.
> I personally think that collective standards regarding risk and pollution have risen significantly...
Then do you potentially think that the reason the price is rising is because of regulation? Because this whole paragraph reads like a justification for the costs incurred due to regulation. I'm not seeing what your complaint is here pinning the inverted learning curve on the regulators - you seem to be saying that is what happened and it is reasonable in this paragraph.
The risk is that hundreds of people avoidably die of cancer over the following decades, and that you lose an immense amount of trust and credibility in government. This is not easy to de-risk because there are a lot of potential mechanisms for bio-accumulation/concentration, and by the time you find out what those exact mechanisms are empirically, the damage might already be done (and by then its too late, and people are no longer gonna trust anything you tell them).
I'm not arguing that the Fukushima evacuation was 100% the right call to make even in hindsight, but it was most certainly a defensible position (even in retrospect).
> Then do you potentially think that the reason the price is rising is because of regulation?
Strongly disagree with the framing. My opinion is that most early reactors did not meet modern safety/pollution expectations and quite frequently resulted in a toxic mess that still requires cleanup today (Sellafield would be a non-US example).
Despite all regulations, the safety track record (2 meltdowns at 500ish reactor sites) is arguably really poor, especially when reactor meltdowns had been portrayed as "virtually impossible" by early proponents.
> What other effect are you considering?
The very same effects that make bespoke heavy machinery comparatively expensive nowadays in general:
- Low benefit from industrial automation
- Mechanical engineering sectors have shrunk in the west (=> parts/subassemblies and qualified labor more expensive)
- Generally higher wages
Those apply much less to renewables because those are less labor intensive and benefit from economies of scale more (by design). Its also much easier to benefit from cheap foreign labor with those (especially solar).
As a sanity check:
I would always expect coal power plants to stay significantly cheaper than nuclear power (when disregarding CO2/fuel cost). But even in that setting, modern coal power (just like nuclear power) already struggles to compete with solar/wind on levelized cost basis today (and trends are very clear!)
So even if we dropped all nuclear regulation today completely (with disastrous likely effects) then the absolute best we could hope for would be to get competitive (on levelized cost) with the most expensive forms of renewables (offshore wind and rooftop solar). Given that it would likely take decades to act on such a plan, all the risks (and current downward trends on renewables and storage costs) chasing that approach seems utterly pointless to me.
Thus I believe the current nuclear approach (build a few reactors here and there so the know-how is not lost completely), which is exactly what both China and the US are doing, is perfectly sensible.
For long-term displacement, many people (mostly sick and elderly) died at an increased rate while in temporary housing and shelters. Degraded living conditions and separation from support networks are likely contributing factors. As of 27 February 2017, the Fukushima prefecture government counted 2,129 "disaster-related deaths" in the prefecture. This value exceeds the number that have died in Fukushima prefecture directly from the earthquake and tsunami. .... As of the year 2016, among those deaths, 1,368 have been listed as "related to the nuclear power plant" according to media analysis
~ https://en.wikipedia.org/wiki/Fukushima_nuclear_accident_cas...
Do you actually mean 100s over decades? Given that the evacuations from Fukushima caused thousands (well, thousand - but we're talking the highest hundreds before we get to thousands) of deaths in years, it seems like the risk of nuclear meltdowns would be completely acceptable. The government forced people to accept higher risk because bureaucrats panicked and it seems that was acceptable. What is the problem with a meltdown supposed to be if more damage is done just for theatre?
You're proposing people respond to a threat that isn't really there. Hundreds of people routinely die over decades from all sorts of industrial processes. A once in a generation risk that somewhere on the globe will suffer, maybe, 100s of deaths over the course of a few decades is nothing. Coal is still worse, is everywhere and people accept that. Is this the limit of the damage you're worried about?
Here's an MIT study that dug into the reasons behind high nuclear construction cost. They found that regulatory burdens were only responsible for 30% of cost increases. Most of the cost overruns were because of needing to adapt the design of the nuclear plant to the site.
https://news.mit.edu/2020/reasons-nuclear-overruns-1118
Now, you can criticize the methodology of that study, but then you have to bring your own study that shows precisely which regulatory burdens are causing these cost overruns, and which of those are in excess. Is it in excess that we have strict workplace safety regulation now? Is it in excess that we demand reactor containment vessels to prevent meltdowns from contaminating ground water supplies? In order to make a good red tape argument I expect detail in what is excess regulation, and I've never seen that.
Besides, if "red tape" and fossil industry marketing was really the cause of unpopularity, and the business case for nuclear was real when installing plants at scale, you would see Russia and China have the majority of their electricity production from nuclear power.
- Russia is the most pro-nuclear country in the world, and even they didn't get past 20% of electricity share. They claim levelized cost for nuclear on the order of that of solar and wind, but I am very skeptical of that number, and anyone who knows anything about the Russian government's relation to the truth will understand why. When they build nuclear plants in other countries (e.g. the bangladesh project) they are not that cheap.
- China sits at low single digit percentages of nuclear share, with a levelized cost that is significantly higher than Russia's and well above that of solar and wind. While they're planning to grow the nuclear share they assume it will be based on new nuclear technology that changes the business case.
Both Russia and China can rely on cheap skilled labor to bring down costs, a luxury western countries do not have.
And this is ultimately the issue: the nuclear industry has been promising designs that bring down costs for over half a century, and they have never delivered on that promise. It's all smoke and mirrors distracting from the fact that building nuclear plants is inherently freaking expensive. Maybe AI can help us design better and cheaper nuclear power plants, but as of today there is no proven nuclear plant design that is economical to build, and that is ultimately why you see so little new nuclear plant construction in the west.
Not France?
> France derives about 70% of its electricity from nuclear energy, due to a long-standing policy based on energy security.
— https://world-nuclear.org/information-library/country-profil...
Wouldn't there still be the OpEx of maintaining the power plants + power infrastructure (long-distance HVDC lines, transformer substations, etc)? That isn't negligible.
...I say that, but I do note that I live in a coastal city with a temperate climate near a mountain watershed, and I literally do have unmetered water. I suppose I pay for the OpEx of the infrastructure with my taxes.
These are fixed costs. They don't go down if people use less power, so the sensible way to pay for them is with some kind of fixed monthly connection charge (e.g. proportional to the size of your service) or via taxes. You're still not measuring how much power people use at any given time.
There are quite a few businesses that can scale limited only by power consumption — but due to that, they today need connections massively overbuilt compared to their current usage (as they project their usage growth to be extremely fast, potentially as much as doubling each year) in order to not need to be constantly re-trenching connections or browning out. Which in turn means that, under a leased-line electrical system, they'd be massively overpaying for unused power capacity at all times — possibly to the point of being unprofitable.
To achieve profitability, they'd need to negotiate billing based on only the fraction of the capacity of the available grid capacity that they're actually demanding on any given day... or, in other words, metered billing.
Many things are cheap when you ignore externalities.
This is frankly nonsense, and my hope is that this nonsense is coming from a person too young to remember the real, valid fears from disasters like 3 Mile Island, Chernobyl, and Fukushima.
Yes, I fully understand that over the long term coal plants cause many more deaths, and of course climate change is an order of magnitude worse, eventually. The issue is that human fears aren't based so much on "the long term" or "eventualities". When nuclear power plants failed, they had the unfortunate side effect of making hundreds or thousands of square miles uninhabitable for generations. The fact that societies demand heavy regulation for something that can fail so spectacularly isn't some underhanded conspiracy.
Just look at France, the country with probably the most successful wide-scale deployment of nuclear. They are rightfully proud of their nuclear industry there, but they are not a huge country (significantly smaller than Texas), and thus understand the importance of regulation to prevent any disasters. Electricity there is relatively cheap compared to the rest of Western Europe but still considerably higher than the US average.
And the fears from Chernobyl was MOSTLY irrational.
The reason for the extreme fears that are generated from even very moderate spills from nuclear plants comes in part from the association with nuclear bombs and in part from fear of the unknown.
A lot (if not most) people shut their rational thinking off when the word "nuclear" is used, even those who SHOULD understand that a lot more people die from coal and gas plants EVERY YEAR than have died from nuclear energy throughout history.
Indeed, the safety level at Chernobyl may have been atrocious. But so was the coal industry in the USSR. Indeed, even if just considering the USSR, the death toll from coal alone caused a similar number of deaths (or a bit more) than the deaths caused by Chernobyl EVERY YEAR [1].
[1] https://www.science.org/doi/pdf/10.1126/science.238.4823.11....
I'm so tired of hearing about how regulation is this magic salve that saves everything. Regulation is what caused Chernobyl. Soviet regulations mandated that the flawed RBMK reactor design be used. They knew it was flawed and they forced people to use it anyway. Because that's what government does. There's a similar story in western countries, where it hasn't been feasible to use better designs due to antiquated government regulations, it's just no one here has screwed up as badly as the Soviets did.
It’ll never be too cheap to meter, but electricity will get much cheaper over the coming decades, and so will synthetic hydrocarbons on the back of it.
(that excludes a brief period when I camped with a solar panel)
There are better options, and at scale they're literally capable of producing electricity that literally is too cheap to meter.
The reasons they haven't been built at scale are purely political.
Today's AI is computing's equivalent of nuclear energy - clumsy, centralised, crude, industrial, extractive, and massively overhyped and overpriced.
Real AI would be the step after that - distributed, decentralised, reliable, collaborative, free in all senses of the word.
Another way for intelligence to get too cheap to meter is for the cost to fall so low it becomes hyperabundant. If you were to, for instance, take AI2027 as a benchmark and think ultimately we'll achieve something like the equivalent of John von Neumann in a box with a 2T dense equivalent parameter model and it will match such a Nobel prize winner's productivity when running inference at say 15 tokens a second (as fast as people can read) then you only need in principle 60 teraflops of AI infernce compute, which is roughly 2x the current Apple Neural Engine. So plausibly by the time you get to the 2030s, every laptop, smartphone, etc will be easily able to run models as powerful as the smartest people.
Somewhat longer term, I'm sure Altman expects the entire process to be automated and for the computational efficiency to rise significantly. If you take recent estimates from various players in the reversible computing space, you'd guesstimate that you ought to be able to do 60tflops by the late 2030s using under 0.1W or ~1kWh/yr which Helion could produce for ~1¢. I do feel like 1 year of cognitive labor from the smartest person a penny or two renders intelligence too cheap to meter out on a per-hour basis.
also this is weird i thought electricity prices only get cheaper? https://fred.stlouisfed.org/series/APU000072610
> The deal would help enable a revival of Unit 1 of the five-decades-old facility in Pennsylvania that was retired in 2019 due to economic reasons.
i wonder why they stopped producing energy in 2019 even though energy prices have gone up over the five decades?
but here is some data bro! https://fred.stlouisfed.org/series/APU000072610