@BrickedKeyboard @gnomicutterance
I think Timnit Gebru nailed it when she pointed out that we can't define Intelligence, which means we can't scope it, which means we can't build it.
The cult of IQ tests which rests on a foundation of science trying to prove that:
A) races are real and have real, heritable differences in intelligence
And
B) that a general intelligence, g, exists
Have done quite solid work proving that neither of those things are true, unintentionally, but still.
Timnit Gebru demonstrates that AI is the intellectual heir of IQ tests which spring from the eugenics movement. If only I was a fraction as smart.
This was debunked over 40 years ago.
What is the reason for the pernicious resilience of the IQ myth?
Spearman was never able to find a g. It's always shown to be a statistical artefact.
*smh* and then the fallacious comparison to climate change.
IQ tests, in common with most "race science", doesn't produce good results.
Climate change constantly does.
You're more like Room Temperature superconductors. (Which may even be becoming real! As opposed to IQ.)
Yeah, ignorance of history (things that happened more than 20 years ago) is strong in these people.
The basis of IQ tests is Spearman's, g, a general intelligence. Inventing a new branch of statistical analysis, Spearman exhaustively shows that scientists can ignore errors if they believe hard enough. (Mismeasure of Man, SJ Gould).
As T. Gebru points out, how can you design or verify a system you can't spec? There's no definition and no evidence g exists.
@saucerwizard @dgerard
For those interested, what David Nutt did _after_ he got fired for saying "actually the evidence is that cannabis is less dangerous than alcohol".
https://www.drugscience.org.uk/drug-information/lsd/