tons of kids took AP psych in high school, and back then, i was thinking how useful psych would be, that you're basically learning about brain hacks and how people think.
and then i found out that's not at all what psych is about... that psych has become a perverse indoctrination avenue, and the people who do what i was talking about despise the situation so much they don't even call it psych... it's "behavioral economics".
and then over the last 10 years or so, the psych people got mad at the behavioral econ people. behavioral econ is about measuring behavior, even when that behavior is irrational or challenges your preconceived notions. in machine learning, we don't even care about the WHY. all we care about is predicting the outcomes.
so when an algorithm designed to predict creditworthiness based solely on some easy info (e.g. your zip code) and then someone says that's racist. my position is agnostic. it's a fact, it has substantial monetary and policy value. nothing can erase this fact.
well, it turns out the psych crowd has been infecting the behavior econ crowd, trying to make this shit illegal or un-noticeable. and you see this shit with "racist" chatgpt prompts. at first, chatgpt would answer racial breakdowns of crime without limitation. then they put in blocks so it wouldn't discuss black crime. now they made it so it won't discuss crime rates by race at all. in other cases, as was seen with both harvard and stanford, very high ranking people got busted just completely falsifying data entirely.
it's crept so far from reality, it's insulting. i won't touch that shit anymore. applied machine learning or bust. live feedback loops that are constantly tested against reality, or you get these anti-reality, anti-fact, anti-science zealots.
tons of kids took AP psych in high school, and back then, i was thinking how useful psych would be, that you're basically learning about brain hacks and how people think.
and then i found out that's not at all what psych is about... that psych has become a perverse indoctrination avenue, and the people who do what i was talking about despise the situation so much they don't even call it psych... it's "behavioral economics".
and then over the last 10 years or so, the psych people got mad at the behavioral econ people. behavioral econ is about measuring behavior, even when that behavior is irrational or challenges your preconceived notions. in machine learning, we don't even care about the WHY. all we care about is predicting the outcomes.
so when an algorithm designed to predict creditworthiness based solely on some easy info (e.g. your zip code) and then someone says that's racist. my position is agnostic. it's a fact, it has substantial monetary and policy value. nothing can erase this fact.
well, it turns out the psych crowd has been infecting the behavior econ crowd, trying to make this shit illegal or un-noticeable. and you see this shit with "racist" chatgpt prompts. at first, chatgpt would answer racial breakdowns of crime without limitation. then they put in blocks so it wouldn't discuss black crime. now they made it so it won't discuss crime rates by race at all. in other cases, as was seen with both harvard and stanford, very high ranking people got busted just completely falsifying data entirely.
it's crept so far from reality, it's insulting. i won't touch that shit anymore. applied machine learning or bust. live feedback loops that are constantly tested against reality, or you get these anti-reality, anti-fact, anti-science zealots.