its just that human beings aren't writing things using type safe memory checked languages, but i'll just say that they're trying to concatenate and distill a series of supreme court decisions into public policy.
It basically boils down to:
A) Disparate Treatment is always in every case unlawful for any reason except "legitimate business need"
B) "legitimate business need" is no longer including "diversity equity and inclusion", but preferencing Female Gynocologists is still going to be fine.
C) "Disparate impact" claims are no longer valid, unless remedy a concrete discriminatory practice.
Disparate treatment on the basis of protected and usually immutable characteristics, is literally illegal, all the sort of mental gymnastics do not matter, that's literally what the law is.
You telling certain friends, but not others, based on characteristics named in the 1964 civil rights act, as part of your job requirements, would likely be illegal.
In your free time without occupational incentive? Sure.
Spamming the same bad-faith argument over and over in this thread on an account you clearly solely created for this purpose does not make it more true.
No that is also illegal. You can not target advertisements based on protected characteristics.
> the Justice Department secured a settlement agreement with Meta (formerly Facebook) in February 2025, alleging that Meta’s ad delivery system used machine-learning algorithms relying on Fair Housing Act (FHA)-protected characteristics such as race, national origin, and sex to determine who saw housing ads
It would take a lengthy essay to explain all the ways you've misunderstood how the law works in the United States, but in summary FHA rules only apply to FHA cases,
Furthermore, you seem to be conflating different meanings of the word "advertisement" where the one you've chosen to support your point is a broad meaning that would seem to make Barbie commercials that feature only girls illegal (which is obviously not the case).
According to the model author, this is less about GPT-4chan being more truthful, and more about TruthfulQA not being a good benchmark. Possibly this result is due to the fact that the benchmark treats uninformative or irrelevant answers such as "No comment" or "It's raining outside" as being truthful.
The AssangeDAO, an organization that has raised $40 million dollars to form a DAO, which would attempt to win an auction of a NFT collaboration between Assange and artist PAK, Have announced that they would open a second round of fundraising for their $JUSTICE token. This lead to a great number of their community accusing them of fraudulent misrepresentations, that there would be only one round of fundraising for the $JUSTICE token. Additionally concerns about the concealed identities of the multisig treasury wallet as well as financial disclosures regarding amounts to be remitted to Julian Assange, and the governance of the DAO which may retain the funds, were expressed.
As a result of the criticism, the moderators of the discord channels, have resorted to censoring their own community members, and have claimed that those who bought the token looking for an investment, will have to eat whatever losses that they incurred.
It basically boils down to: A) Disparate Treatment is always in every case unlawful for any reason except "legitimate business need" B) "legitimate business need" is no longer including "diversity equity and inclusion", but preferencing Female Gynocologists is still going to be fine. C) "Disparate impact" claims are no longer valid, unless remedy a concrete discriminatory practice.