Is the World Getting Safer? | Hacker News Reader

Is the World Getting Safer?

unintendedconsequenc.es

18 pointspaulorlando28 comments

So it's lies, damned lies and statistics again. When I was young I believed that rational debate could actually solve issues; now it feels like a sufficiently clever actor can basically argue for any point of view with solid 'looking' reasoning to back them up. All very postmodern (post-truth) although it was always thus.

Is the point of this article that we shouldn't trust anything (even when there are numbers to back up the conclusion) or that we can find the truth if we are smart enough to ask the right questions?

This plays out every day in the court system. Having a sufficiently clever lawyer does help. Having the truth on your side does too. Neither, not even both together, is a 100% guarantee.

It's not exactly a new problem: https://people.uwplatt.edu/~ciesield/platovsoph.htm

This article doesnt have much substance, I mean if you want to challenge the (almost) scientific consensus on an issue, you need to have a few good smoking guns. I like the idea of lower mortality due to better medical care, but the US is hardly representative of the world.

It would have been a much stronger article if the completely-hypothetical graphs had been replaced with ones that actually show the effect in real data, even if on a relatively small example.

`Problem might be copyright issues here, but generally I agree.

Yeah seriously, there's almost no substance here. Just stuff like guesses about earlier societies inflating death counts.

"Sure, these so called social scientists and historians have data on their side, but I have something even more powerful: supposition!"

Pointing out the limitations in data-driven arguments (that everyone should know & come to accept) doesn't actually move any particular position forward, since using data-driven arguments is agnostic to any particular position.

So at best, the argument this article posits is: Pinker isn't right, but I don't know who is

Outliers with high impact become news. This is one reason why people become more afraid for an airplane or terrorist than a car accident.

Recommended related material: the book Factfulness by Hans Rosling [1] and Terrorism and Counterterrorism: Comparing Theory and Practice by Universiteit Leiden [2].

[1] https://en.wikipedia.org/wiki/Factfulness%3A_Ten_Reasons_We%...

[2] https://www.coursera.org/learn/terrorism

The narrative now is to acknowledge that the world was getting safer, but that the progress has now stopped.

For example https://marginalrevolution.com/marginalrevolution/2020/10/th...

It's interesting that the science seems to trail people's intuitions. How long time does it take from the time a phenomenon happens until it's seen in the curves, articles are written, reviewed and trickle into news and textbooks?

No..it's not:

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

Not sure how the link supports your claim. The question was not "is the doomsday clock moving closer to midnight" and safety is non synonymous with time on the doomsday clock.

And it really just detracts from a well written and reasoned counter argument to what Steven Pinker said. I tend to agree with the article. And it is good counter to the old "don't mind the gangs, they only kill each other" claptrap really.

The core point this article raises is that it is not clear why homicide rate should be accepted as the sole and primary measure of safety. And here you come and suggest that instead we should take the doomsday clock ... which I would argue would be a step back.

The Doomsday clock idea has a lot of respectable people behind it but I've always been unclear how it isn't just a bias-o-meter.

Kahneman and others have done a lot of persuasive work examining how experts are often quite bad at judgement/prediction. How can we be sure the Doomsday Clock actually means anything substantive (as compared to data)?

>bias-o-meter

Oh i think it absolutly is...but a good one ;)

Meh. That's a very political thing at this point.

As is ignoring it (assuming it is factual).

>That's a very political thing at this point.

Yes and?

Political in the sense of "they decide the time disingenuously to score political points".

You mean once a year?

No I mean x minutes to midnight.

What has that todo with "disingenuously"?

Self-destruction is a kind of violence

And the doomsday clock is not a measure of violence, safety or self-destruction. It is a measure of the opinion of the Bulletin of the Atomic Scientists.

It represents something about the greatest tail risk that our society faces. This afternoon 300m people could die, all of our major ports, cultural centres, top 250 research institutes and universities, top 100 ports, most major hospitals, most power stations and on and on could be destroyed. This afternoon.

The rolling consequence of this event would lead to a large number of subsequent deaths - possibly billions - within 2 years. Progress and development would be stopped for years everywhere, many regions would fall backwards by 50 or 100 years. It's possible that human civilisation would never recover.

Some matter of opinion.

> It represents something about the greatest tail risk that our society faces.

through what mechanism? How is the time connected to this and what procedure can I undertake to devise the number other than poll the people who set it for what they think it should be? And why should I substitute this for the definition of safety?

There are some clear indicators that you can use :

- number of warheads active - number of anti-ballistic measures in play - active conflicts involving nuclear armed powers - active conflicts between nuclear armed powers - new nuclear technologies (ie. Bayesian fuses, hyper-sonic maneuverability, suppressed trajectories) - new nuclear powers/capabilities (ie. NK gets H-bomb, NK and India deploy boomers) - change in treaties (withdrawals) - threats made by national leaders

I don't know what you mean by "definition of safety" - while there are thousands of nuclear weapons in play no one is actually safe, ever.

Reply on news.ycombinator.com