Hacker News Reader

Comment by amarshall | Hacker News Reader

Very unsafe? In the past 14 years there have been 72 fatalities involving US Air Carriers, out of around 250 million flight hours flown[1]. That’s fewer fatalities in 14 years than there are US motor vehicle fatalities in a single day (on average).

[1]: https://www.ntsb.gov/safety/StatisticalReviews/Pages/CivilAv...

View on HN

Humans are bad at statistics. For example the incident at three mile island in 1979 didn't kill anyone, but the accident crystallized anti-nuclear safety concerns among activists and the general public, and accelerated the decline of efforts to build new reactors.

Don’t underestimate the direct effects.

Three mile island directly cost well over 1 billion in 1979 dollars (2+B today) just in terms of destroyed assets and initial cleanup costs. The wider impact was even more expensive.

Such a visible failure changed the risk/reward calculations which then hurt the nuclear industry quite a bit. We did keep building US nuclear reactors afterwards, but they were never that profitable in the first place making the industry very sensitive to disruption.

Timeline of US reactor construction: https://en.wikipedia.org/wiki/Three_Mile_Island_accident#/me...

The problem as I see it is that the costs of Three Mile are acute and fairly direct. Whereas fossil fuel power may have even an even bigger cost (from the environmental impact), but as they are slow, chronic, and indirect, they’re easily overlooked.

Humans just seem to be disproportionately responsive to rare, acute events.

To "slow, chronic, and indirect", I would also add "spanning most of a human lifespan." For these kinds of things we often don't realize something is brand new as of our lifetime. Said in another way, if something changes in our life when we are 2, we will tend to think that thing was always that way, even though it is very recent.

An example, forest fires in the West (coast US). They have always been around. So, many say, nothing new here. Yet, we don't quite grok that they are 10x worse than 50 years ago [1]. Thus, if you look at it across multiple human lifetimes we can see there is a radical difference. Across one lifetime and it might not seem like it is so different.

[1] https://www.nytimes.com/interactive/2020/09/24/climate/fires... (has several graphs that indicate magnitude more fire & magnitude more burn area)

Those differences need to be prices in via Coal taxes or the market doesn’t care.

However, the market does care if something has proven to be a risky bet. Meltdowns involve actual money on the line.

Looks like Chernobyl really killed nuclear.

Compare active reactors + reactors in construction during three mile island vs where nuclear production stabilized.

Things stabilizing like that is very suspect. It probably has to do with capacity factors and the long lead time, but I wouldn’t assume there’s no correlation.

Yeah, I'm not saying there's no correlation, just that you see the full-on halt kick in after Chernobyl. Which makes intuitive sense to me. Even though our reactors were very different from Chernobyl, the incident was so much more extreme and the distinctions were probably largely lost to the general public.

You have it backwards. Industries that are profitable are subject to disruption.

Wrong kind of disruption. If your profit margins are 0.5% then a tiny dip in demand, 2% spike in interest rates, or spike in steel costs etc can be devastating.

If you’re a software company with profit margins over 40% then you don’t care much about rent etc.

This question sort of answers itself, but why skirt the real concern, just add a "Chance of crashing" filter to the flight search? Bonus jobs for data scientists and takes the consumer out of the odds calculation.

the impact of violence vs other causes of death is similar, 9/11 killed 3K people and we went to war for 20 years and spent trillions as a result. Meanwhile 100K Americans died from opioids last year and the government does nothing

requiring events to kill anyone as evidence of danger of death is a foolish standard, obviously.

People are "bad at statistics" in the sense that the very real evidence coming from, e.g. 3 mile, is hard to bring into statistical models appropriately, not in the sense that there was no evidence there.

For Three Mile, indeed there is some disagreement on the long-term non-fatal health effects. That any causal effect is unclear does not mean it does not exist, but does suggest that, even if it does, it is likely small. (Admittedly, I have not done a thorough review of the data on this). Nevertheless, your point carries to the true comparison of nuclear power: fossil fuel power and its long-term economic and health impact.

For the airline risk example: US Airline Carriers in past 14 years: 268 serious injuries (same source as above). Depending on the year, that’s about the same or less than the number of road traffic related injuries in the US in just one hour (on average)[1].

[1]: https://www.statista.com/statistics/191900/road-traffic-rela...

So what are you basing it on if not deaths/flight hours

Scary news reports per minute

Reply on news.ycombinator.com