i/o Profile picture
Nov 1, 2023 1 tweets 1 min read Read on X
WaPo: Artificial intelligence generates "stereotypes" about crime and success "that don’t reflect the real world."



Uhm, stereotype accuracy is one of the most robust findings in social science for a reason: People notice patterns that really exist.

If AI wants to "reflect the real world" then generating stereotypes is inevitable.

But the WaPo would rather AI spit out one soothing lie after another, like that African-Americans aren't any more violent than Asian-Americans (even though they commit murder at over 20 times the rate that Asians do), and that they're just as cognitively capable (an American of East Asian ancestry is over 100 times more likely than a black person in America to have an IQ of over 140).washingtonpost.com/technology/int…

• • •

Missing some Tweet in this thread? You can try to force a refresh
 

Keep Current with i/o

i/o Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @eyeslasho

Nov 14
No person has ever brought ignominy and disrepute upon a once-respected publication as quickly and decisively as this ridiculous woman. Image
No, women aren't as fast as men. No, racism doesn't explain the near-complete absence of blacks in scientific innovation. No, EO Wilson wasn't a white supremacist. No, the lab-leak theory for COVID origins isn't a "rightwing hoax." And on and on.
Science. Image
Read 5 tweets
Oct 7
[THREAD]

1/ At a South African university, an argument is made to eliminate science from study (because it's a "product of Western modernity) and to "restart [it] from an African perspective."

What should be studied? The speaker suggests "black magic."
2/ When a "science person" objects, he is scolded by the organizer for "disrespecting the sacredness of this space," and asked to apologize, which he does. But that doesn't stop the scolding. Opinions can only be expressed under rules that appear to guide outcomes.
3/ The "black magic" advocate then adds that, despite the fact that she took some science in high school, she decided to not be on the science faculty because science stands in the way of "decolonization."

(Note: Africa was decolonized over 50 years ago.)
Read 6 tweets
Sep 5
Sex offending rate of women: 3 per one million

Sex offending rate of men: 395 per million

Sex offending rates of transwomen: 1,916 per millionImage
Conviction/incarceration claim:



Population claim:

archive.is/5bIVF
ons.gov.uk/peoplepopulati…
"In 2018, the BBC Reality Check team found that 48 per cent of transgender offenders were serving time for a sexual offence."



2023 data on trans incarcerated population in UK:

crimeandjustice.org.uk/resources/pris…
crimeandjustice.org.uk/resources/pris…
Read 4 tweets
Mar 30
"Intelligence has been the primary evolutionary vector for humans. What is underappreciated is how much happened in the last 10k years. There was an intelligence explosion ~6k years ago with sudden development of writing, agriculture, elementary math, etc."

[Italics mine]
@EPoe187 @Russwarne @Steve_Sailer @AporiaMagazine
In other words: Evolution didn't magically stop when humans exited Africa, and the most important evolutionary adaptation to the novel environments encountered by humans occurred not below the neck, but above it.
Read 4 tweets
Mar 7
[THREAD]

1/ This is fascinating.

Women of which race/ethnicity are most likely overall to receive a response from a man (across all ethnicities) to a message they send on a dating app?

Answer: Middle Eastern women.

Almost 1/2 of men will respond to a Middle Eastern woman. Image
2/ By contrast, only about 1/3 of all men overall will respond to a black woman— the only female demographic with less than a 40% overall response rate.
3/ The results below show how different ethnicities rate each other on attractiveness on OK Cupid.

Overall, women find white men to be the most attractive and Asian men the least attractive. Men find Asian women to be the most attractive overall and black women the least. Image
Read 6 tweets
Jan 13
How do we know that blacks were admitted to Harvard at about 20 times the rate they would have been admitted had admissions been based solely on academic performance?

Because the consultant to the plaintiffs in the Harvard lawsuit (a Duke professor) found that if Harvard had admitted only those students in the top academic performance decile of applicants, the percentage of those admitted who were black would have been 0.9%, not the 15.8% which were actually admitted — in other words, with a 10% admissions rate, and using academic merit as the only admissions criterion, black applicants would have had their chance of admission reduced by about 16 times.

But Harvard's admission rate isn't 10%. It's about 3%. So what would have happened if Harvard had only admitted the top 3% of the applicants solely on academic merit?

I don't have the raw data set, so I can't use it to estimate whether the 16 times preference accorded blacks would have increased or decreased, but I don't need it because we know that EVERY large-sample data set of group cognitive performance in the US has shown the same thing: As one approaches the far right-tails of the group cognitive distribution curves, black representation (compared to whites and especially Asians) almost vanishes. We see this with both standardized college admissions testing and with IQ testing.

This means that an admissions policy looking only at high school GPA and standardized test performance, when combined with a 3% admissions rate, would have produced a percentage of admitted blacks that was even less than 0.9%. Extrapolating from what see in standard group cognitive testing distributions, I estimate that the black percentage of admits would have shrunk from 0.9% to about 0.4-0.6%.

That 20 times advantage for blacks that I referenced earlier is a very conservative estimate. It's likely that blacks applying to Harvard under the previous institutionally-racist admissions policies enjoyed an advantage in excess of that.
OK, I can see that some repliers are confused, and perhaps I contributed to this confusion.

I use the term "admission rate" above, not "enrollment rate." In other words, I'm referring to those students who were offered admission to Harvard, not those who actually enrolled.

It occurs to me that a less confusing term might have been "acceptance rate."
The black enrollment rate at undergrad Harvard (which has typically been in the upper single-digits) has always less than the black admissions rate because many blacks accepted at Harvard decide to enroll at another school.
Read 4 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us!

:(