I know just one person over 100 with an actual birth certificate.
Across U.S. states, the total and per capita numbers of supercentenarians dramatically decline right after the introduction of birth certificates (blue line).
Also, have you ever noticed that supercentenarians are more common in areas with more crime, more poverty, and lower average life expectancies?
Here's data for England:
The same pattern of supercentenarian numbers being correlated with poverty holds in (A, D) England, (B, E) France, and (C, F) Japan.
Across countries, you just see the same things over and over, from age heaping to weird correlations, so the conclusion is clear:
Supercentenarian numbers are driven less by regionally exceptional longevity and more by people defrauding pension systems and making up their ages.
Oh, and if you wanted to learn how to live a long life from the "blue zones" in Sardinia, Okinawa, and Icaria, good luck. Those places have low life expectancies and literacy levels, high crime, and lots of poverty.
Their long-lived people are not able to validate their ages.
This also applies to Loma Linda (not all that exceptional of a place).
In fact, across the whole U.S., at least 17% of centenarians were found to be non-centenarians in 2019 when someone just read through two plain-text files and found dates didn't match.
And this also applies to Nicoya, which is riddled with fraudulent ages:
If someone says they know someone super old, ask them: Where were they born? If it's in some place that was poor in the not-too-distant past, then they probably have the wrong age.
I simulated 100,000 people to show how often people are "thrice-exceptional": Smart, stable, and exceptionally hard-working.
I've highlighted these people in red in this chart:
If you reorient the chart to a bird's eye view, it looks like this:
In short, there are not many people who are thrice-exceptional, in the sense of being at least +2 standard deviations in conscientiousness, emotional stability (i.e., inverse neuroticism), and intelligence.
To replicate this, use 42 as the seed and assume linearity and normality
The decline of trust is something worth caring about, and reversing it is something worth doing.
We should not have to live constantly wondering if we're being lied to or scammed. Trust should be possible again.
I don't know how we go about regaining trust and promoting trustworthiness in society.
It feels like there's an immense level of toleration of untrustworthy behavior from everyone: scams are openly funded; academics congratulate their fraudster peers; all content is now slop.
What China's doing—corruption crackdowns and arresting fraudsters—seems laudable, and I think the U.S. and other Western nations should follow suit.
Fraud leads to so many lives being lost and so much progress being halted or delayed.
British fertility abruptly fell after one important court case: the Bradlaugh-Besant trial🧵
You can see its impact very visibly on this chart:
The trial involved Annie Besant (left) and Charles Bradlaugh (right).
These two were atheists—a scandalous position at the time!—and they wanted to promote free-thinking about practically everything that upset the puritanical society of their time.
They were on trial because they tried to sell a book entitled Fruits of Philosophy.
This was an American guide to tons of different aspects of family planning, and included birth control methods, some of which worked, others which did not.
One of the really interesting studies on the psychiatric effects of maltreatment is Danese and Widom's from Nat. Hum. Behavior a few years ago.
They found that only subjective (S), rather than objective (O) maltreatment predicted actually having a mental disorder.
Phrased differently, if people subjectively believed they were abused, that predicted poor mental health, but objectively recorded maltreatment only predicted it if there was also a subjective report.
Some people might 'simply' be more resilient than others.
I think this finding makes sense.
Consider the level of agreement between prospective (P-R) and retrospective (R-P) reports of childhood maltreatment.
A slim majority of people recorded being mistreated later report that they were mistreated when asked to recall.
The Reich Lab article on genetic selection in Europe over the last 10,000 years is finally online, and it includes such interesting results as:
- Intelligence has increased
- People got lighter
- Mental disorders became less common
And more!
They've added some interesting simulation results that show that these changes are unlikely to have happened without directional selection, under a variety of different model assumptions.
They also showed that, despite pigmentation being oligogenic, selection on it was polygenic.
"[S]election for pigmentation had an equal impact on all variants in proportion to effect size."
I still think this is one of the most important recent papers on AI in the job market🧵
The website Freelancer added an option to generate cover letters with AI, and suddenly the quality associated with cover letters stopped predicting the odds of people getting hired!
LLMs do a few things to cover letters.
Firstly, they increase the quality, as measured by how well tailored they are to a given job listing.
Second, they make job applications in expensive, so people start spending less time shooting off applications.
More, rapidly-produced job applications becomes the norm.