4 ms·

I had the interesting experience of working in a startup under a VP of Engineering who was a Nuclear Scientist and was part of the team the US sent to Chernobyl

I had the interesting experience of working in a startup under a VP of Engineering who was a Nuclear Scientist and was part of the team the US sent to Chernobyl's sister site - exact same plans, basically - to ensure its' safety. He told me that overwhelmingly what went wrong at Chernobyl is that the head of the plant consistently ignored all of the alarms and monitoring in place, because he was overconfident at what the results of his test would be. He apparently died in the control room. Interestingly, one thing the Russians did notably different was to have a big guy, I want to think at the sister site his name was Sergei, basically live in a hut next to the cooling lake, which was uphill from the plant. In the event of a disaster, he had a humongous mallet he would use to slam open the valve, draining the lake. ... or maybe this was all made up, I couldn't tell you tbh, but it sounded legit to me. ;)

Tomte 11y ago
He told me that overwhelmingly what went wrong at Chernobyl is that the head of the plant consistently ignored all of the alarms You may consider this an excuse of the kind "but that's not nuclear power's fault" and you'd be right-ish, but I simply don't care who exactly is at fault for whole regions to become uninhabitable in case of disaster. That is connected to nuclear power in my risk assessment, even if you can make a rational argument that it shouldn't be. And I'm very risk-averse when the magnitude of consequences is that big. Same with Fukushima: yes, everything was safe. Unconnected to the reactor per se they made huge planning errors, underestimating a tsunami. Again, I don't care. It may be a bias or even a fallacy that Kahnemann has already described, but it's a useful and correct fallacy, I would argue.
- readams 11y ago
  I'm fascinated by the idea of a "correct fallacy." How could such a thing ever make sense? Are you saying that you use this fallacy to reach a conclusion that you believe correct for other reasons? If so, why not state those other reasons and not use your fallacious reasoning? If not, how could your fallacy be correct?
  - Tomte 11y ago
    Me too. :-) I think that this fallacy is correct in the sense that averting extreme outcomes at the cost of a lower expected outcome is rational when it comes to survival. I suspect you might be able to explain the existence of those "irrational" behaviours this way, although I haven't searched the literature. Kind of the opposite of playing the lottery (which I also find rational, despite all the nerds decrying it): There you are giving up expected value, in order to gain variance.
    - Retra 11y ago
      Something is not rational simply because you think it is. A decision is rational if and only if the invested resources are proportional to the likelihood of success. A fallacy is a fallacy because it violates this proportionality. Fallacious reasoning can never be rational.
- wrsh07 11y ago
  > for whole regions to become uninhabitable Chernobyl has a 30km exclusion zone, and that's unnecessarily large. > Same with Fukushima Literally nobody died in Fukushima. It was just a memorable piece of destruction caused by a natural disaster that killed 16,000 people.
  - Tomte 11y ago
    And you can mathematically prove that no possible future disaster could possibly affect a larger area? No? Okay, my point stands. I simply don't care about probability-weighted outcomes there.
    - wrsh07 11y ago
      That's not how this works. We're not dealing in absolutes because this is real life. We're dealing with probabilities and expected values. If nuclear power provides a net good, we should consider using it, not blindly reject it. Saying you "don't care about probability-weighted outcomes" means you aren't willing to objectively value nuclear power. Why not? Why are you afraid of nuclear power? The point this article makes is that nuclear power is an incredibly safe technology, and the concerns about it are exaggerated.
      - mohawk 11y ago
        It would be good if these risks were objectively valued, for example by insuring the plant against meltdown costs. Total economic loss estimates for the Fukushima meltdown are 250-500 bn USD[1]. And if the wind had been slightly different those numbers would be a lot higher, as Tokyo real estate would be affected. And that insurance should also cover human error, terrorism, natural disasters. Won't be cheap is my guess. Which brings us to the biggest myth of all, that nuclear power is cheap when you don't externalise so many costs. [1] http://www.psr.org/environment-and-health/environmental-health-policy-institute/responses/costs-and-consequences-of-fukushima.html http://www.psr.org/environment-and-health/environmental-heal...
        11y ago

Tomte 11y ago

He told me that overwhelmingly what went wrong at Chernobyl is that the head of the plant consistently ignored all of the alarms You may consider this an excuse of the kind "but that's not nuclear power's fault" and you'd be right-ish, but I simply don't care who exactly is at fault for whole regions to become uninhabitable in case of disaster. That is connected to nuclear power in my risk assessment, even if you can make a rational argument that it shouldn't be. And I'm very risk-averse when the magnitude of consequences is that big. Same with Fukushima: yes, everything was safe. Unconnected to the reactor per se they made huge planning errors, underestimating a tsunami. Again, I don't care. It may be a bias or even a fallacy that Kahnemann has already described, but it's a useful and correct fallacy, I would argue.

readams 11y ago
I'm fascinated by the idea of a "correct fallacy." How could such a thing ever make sense? Are you saying that you use this fallacy to reach a conclusion that you believe correct for other reasons? If so, why not state those other reasons and not use your fallacious reasoning? If not, how could your fallacy be correct?
- Tomte 11y ago
  Me too. :-) I think that this fallacy is correct in the sense that averting extreme outcomes at the cost of a lower expected outcome is rational when it comes to survival. I suspect you might be able to explain the existence of those "irrational" behaviours this way, although I haven't searched the literature. Kind of the opposite of playing the lottery (which I also find rational, despite all the nerds decrying it): There you are giving up expected value, in order to gain variance.
  - Retra 11y ago
    Something is not rational simply because you think it is. A decision is rational if and only if the invested resources are proportional to the likelihood of success. A fallacy is a fallacy because it violates this proportionality. Fallacious reasoning can never be rational.
wrsh07 11y ago
> for whole regions to become uninhabitable Chernobyl has a 30km exclusion zone, and that's unnecessarily large. > Same with Fukushima Literally nobody died in Fukushima. It was just a memorable piece of destruction caused by a natural disaster that killed 16,000 people.
- Tomte 11y ago
  And you can mathematically prove that no possible future disaster could possibly affect a larger area? No? Okay, my point stands. I simply don't care about probability-weighted outcomes there.
  - wrsh07 11y ago
    That's not how this works. We're not dealing in absolutes because this is real life. We're dealing with probabilities and expected values. If nuclear power provides a net good, we should consider using it, not blindly reject it. Saying you "don't care about probability-weighted outcomes" means you aren't willing to objectively value nuclear power. Why not? Why are you afraid of nuclear power? The point this article makes is that nuclear power is an incredibly safe technology, and the concerns about it are exaggerated.
    - mohawk 11y ago
      It would be good if these risks were objectively valued, for example by insuring the plant against meltdown costs. Total economic loss estimates for the Fukushima meltdown are 250-500 bn USD[1]. And if the wind had been slightly different those numbers would be a lot higher, as Tokyo real estate would be affected. And that insurance should also cover human error, terrorism, natural disasters. Won't be cheap is my guess. Which brings us to the biggest myth of all, that nuclear power is cheap when you don't externalise so many costs. [1] http://www.psr.org/environment-and-health/environmental-health-policy-institute/responses/costs-and-consequences-of-fukushima.html http://www.psr.org/environment-and-health/environmental-heal...
      - 11y ago