Hacker News Reader

Comment by inigyou | Hacker News Reader

I believe this is why the USA recently declared that it's no longer going to follow the "linear no-threshold model" and "as low as reasonable achievable" principle of radiation exposure. In other words they're declaring that a small amount of radiation is good for you. Building a reactor without shielding and with a single open-cycle water loop should reduce costs, right?

(Do you know what else reduces costs? Apparently, putting a foot of graphite at the bottom of your control rods and not paying close attention to a dangerously unstable reactor during an unusual experiment. These aren't official American policy yet.)

Also, the USA's recent ban on solar electricity equipment should help make other forms more competitive.

View on HN

Linear no-threshold and ALARA are pretty objectively overly conservative, to the point of potentially doing more harm than good (Fukushima is mentioned down below. More people died as a result of the evacuation than would have resulted from the radiation) as well as making many things obscenely expensive. It's a little absurd that nuclear plants obsess over radiation doses smaller than a short-haul flight.

Linear no-threshold being false doesn't necessarily mean a little bit of radiation is good for you, though there is some shaky data to suggest that might be true, just that small amounts of radiation are not as proportionally bad as large amounts.

The harms of radiation are massively overhyped.

Nobody from the Fukushima accident died from radiation sickness or radiation syndrome, and so far there's been no evidence of long term elevated cancer rates either.

Meanwhile the nuclear industry has been drowning for 50+ years from egregious layers of overregulation.

It costed NuScale $500+ Million just to write the documents required for regulatory approval for a new reactor design.

Microsoft is spending $60 Million a year to build an LLM whose only purpose is to automate writing nuclear regulatory documents.

I don't support the Trump admin one iota, but I'm also saddened that democrats tend to be too safetyist to do an honest cost benefit evaluation of some of these insane regulations.

Country of lawers vs country of engineers

>The harms of radiation are massively overhyped.

>Nobody from the Fukushima accident died from radiation sickness or radiation syndrome, and so far there's been no evidence of long term elevated cancer rates either.

You say these two things as if the second statement supports the first, but it doesn't.

The dangers of radiation are very well known and don't change at all based on what happened at Fukushima. If you fire a gun and miss, the gun is still dangerous.

>Meanwhile the nuclear industry has been drowning for 50+ years from egregious layers of overregulation.

Here's an account of two gruesome deaths caused by poor regulation of nuclear materials in Japan. It happened 12 years before Fukushima.

https://science.howstuffworks.com/hisashi-ouchi.htm

Interesting but people die (or are severely injured) in all kinds of industrial accidents all the time. Unfortunately, with or without regulation it comes with the territory of a modern industrial society.

>Interesting but people die (or are severely injured) in all kinds of industrial accidents all the time.

Very few other industrial accidents cause effects that have to be dealt with for decades over 1000 miles away, even across oceans.

Nuclear power really is in a league of its own when it comes to risk. I am not aware of anything that comes close. Dam accidents have killed many more people and devastated towns, but the day after that happens, the area can be reused and the risk is entirely gone. The risks of nuclear power remain for decades or even centuries.

Finding ways to deal with that risk is fine, arguing that it can be managed is fine, but pretending that the risk is exaggerated based on the accidents that have happened so far is a mistake.

The risk is things that could happen, not what has actually happened.

You probably mean the Chernobyl accident.

There are estimates that pollution from European coal power plants cause more deaths EACH year in Europe, than the total estimated deaths from Chernobyl accidents (most pessimistic estimate 4000-8000 deaths).

For example, just the Great Smog event that affected London, England, in December 1952 caused by modern estimate 10,000–12,000 deaths.

https://en.wikipedia.org/wiki/Great_Smog_of_London

>There are estimates that pollution from European coal power plants cause more deaths EACH year in Europe,

First and foremost - coal sucks. Secondly - the choice isn't between nuclear or coal.

That said, the estimates you mention are not believable. Pollution as a whole causes lots of premature deaths, but coal burned in electricity plants are a very small part of total pollution.

Especially in Europe, where most countries no longer operate coal power plants. Not even the UK.

https://ourworldindata.org/grapher/death-rates-from-energy-p...

People die from the operation & externalities of every energy source, it's just that Nuclear power has orders of magnitude fewer deaths per kilowatt-hour than the dirtier energy sources these regulations have pushed energy production towards.

I'm sure some lives would be lost if Nuclear Power faced the same regulatory costs that other energy sources do, rather than the egregious and absurd costs they face today.

But many many more lives would be saved by deregulating nuclear energy and making it more cost competitive with coal and natural gas.

Many children wouldn't contract asthma & the other respiratory illnesses they do today if nuclear energy hadn't been killed via overregulation and replaced by coal and natural gas.

Where would unregulated nuclear power go on that chart?

Most of the nuclear power that makes up that chart was built when there was significantly less regulation on reactors than there is today.

Even if deregulation increased nuclear deaths per watt by two orders of magnitude (which I strongly doubt!), nuclear would still be orders of magnitude safer than the fossil fuels that have replaced it across the western world.

>Even if deregulation increased nuclear deaths per watt by two orders of magnitude (which I strongly doubt!)

Why doubt it?

Chernobyl alone was very close to causing millions of deaths. The actual accident released only about 5% of the core, but had the graphite fire not been extinguished, places as far away as Scotland and Scandinavia would not be able to grow enough edible food within their borders for decades. Europe would starve. Leukemia cases would be in the millions.

The exclusion zone around Chernobyl would be hundreds of kilometers instead of 30.

If the ground water barrier had been broken - which was likely - the ground water for tens of millions of people would be poisoned for centuries.

The only reason to doubt it is that two orders of magnitude is probably not enough.

I hadn't heard any of these claims before.

Without a source, I decided to ask chatgpt. Its assessment is that essentially every claim in this comment is false. If you think chatgpt is wrong can you explain why?

https://chatgpt.com/s/t_6a72ee6f965c81919cc6e4bf8da64262

By all means use chatGPT, but your linked chat session hasn't understood the conditions. It answers mostly based on what actually happened, not taking into account the scenario I described.

Try asking chatGPT something like:

"just a simple yes or no - Could the Chernobyl accident in theory under worst case conditions have developed to a disaster causing local food shortages in western europe and affect the water supply for millions of people in the Ukraine and neighboring countries?"

For me, I get "Yes" from ChatGPT, Gemini and Claude Opus.

You may argue that Chernobyl was a hunk of junk plant and that it can't happen anywhere else, so ask ChatGPT this as well:

"just a simple yes or no - is there any scenario where fuel stored in ponds at US nuclear plants could ever release pollution like Chernobyl?"

Yes.

It really is in a class of its own when it comes to risk.

>Nuclear power has orders of magnitude fewer deaths per kilowatt-hour

By this logic the minuteman missile is harmless. It hasn’t killed anyone, right?

Nuclear power has the potential to cause orders of magnitude more damage than any other power plant.

Again, what has happened is irrelevant, the risk is all about what could happen.

I think you're seriously misunderstanding the statistic I posted.

I didn't post aggregate deaths from different power sources, I posted deaths per terawatt-hour of energy produced!

This figure includes deaths from every nuclear accident, and yet it shows getting energy from nuclear power is orders of magnitude safer than getting the same energy from the fossil fuels that have replaced nuclear power in countries that have phased them out. (like germany)

> I think you're seriously misunderstanding the statistic I posted.

Not at all, i am very familiar with this argument, but it’s flawed.

It doesn’t matter at all how many people have died so far. It is entirely irrelevant for determining the risk.

How is it irrelevant? It establishes the base rate under specific regulatory regimes.

Most of the nuclear plants that make up that base rate were built when nuclear regulations were much much more lax than they are now, and yet Nuclear ends up with the safest track record out of all energy sources.

>How is it irrelevant?

Because it only looks at what has happened, not what could happen. And what could happen is the interesting thing.

Before Chernobyl happened, the statistics didn't tell us it could happen. Before Fukushima happened, the statistics didn't tell is it could happen.

They happened anyway. The next thing that happens will similarly be completely independent from anything that happened before.

So - what happened before is irrelevant. What COULD happen is relevant.

> Nobody from the Fukushima accident died from radiation sickness or radiation syndrome, and so far there's been no evidence of long term elevated cancer rates either.

There has actually been 1 radiation death

How can writing documents possibly cost $500,000,000?

How could writing code possibly cost that much? And yet tech companies spend significantly more than that per year on it. You need a lot of highly paid, experienced people to do it (negotiating regulatory frameworks is pretty much its own profession).

https://www.nuscalepower.com/press-releases/2020/nuscale-pow...

"NuScale spent over $500 million, with the backing of Fluor, and over 2 million labor hours to develop the information needed to prepare its DCA application. The company also submitted 14 separate Topical Reports in addition to the over 12,000 pages for its DCA application and provided more than 2 million pages of supporting information for NRC audits."

Reply on news.ycombinator.com