Which was also rumored to be the work of Christians

I figure it was the same as the reason Elon bought Twitter, kinda. They got it as part of a larger purchase so that's different, but the goal seems to be the same - destroy what's good about it so that the people who were accomplishing things with it stopped. We used to have discussions here that moved conversations forwards... Occasionally anyway.

Tolerance of abuse is enablement, but enablement of abuse is all you want because you want to be enabled to abuse, so that checks out. You have that in common with all rapeublicans. That's why you all voted for the child rapist, you admire his rapes.

Not to mention that Trump is also a white supremacist, and there's lots and lots of evidence of that available.

Those are all good reasons why you would do the updates, and very bad reasons not to, so those are puzzling statements to make in this context even though they are accurate.

The example of minstrel shows was given upthread.

They perpetuated harmful racist stereotypes, which contributed to harm.

You want to bring those back or what?

I think Slashdot has pretty successfully picked apart this study, though I see no need to depend on anything to discredit this study but the weird avatar choices. They make the entire study invalid because of the confounding factors of the avatars' appearances.

Yeah, that question is not just bad, it's going to cause you legal problems if any applicant who was turned down finds out. Interview questions and processes are supposed to be designed to eliminate bias, but that question is literally designed to allow bias to influence the hiring decision, and will be a major exhibit in a discrimination case. Your employer's HR department is run by a spectacular fuckup.

Yeah, there's a reason for that.

No they fucking are not.

Yes, they approve of people who protect child rapists. See? Not irrelevant.

I'm not complaining about the company doing advertising, I'm complaining about Slashdot presenting it without irony.

>not [seeking] an argument

Oh! Oh! I'm sorry! This is slashdot, in Room 12. This is Abuse!
For arguments, you want 12A, next door!

>people don't hire on merit
>Which candidate would you feel most comfortable working with for 40+ hours a week?

Why do you think "merit" and "works well with the employees I already have" are unrelated?

>Of course the candidate with a similar socioeconomic profile as your own would be preferred:

I suspect that's often true, but certainly not universal.

For example, allegedly, some people want slaves and want to lord themselves over their slaves. So they would never hire people like themselves, SES wise; they would only hire people where it was clear (in their own minds) that there was SES or other status daylight between themselves and their "underlings" So I'm not sure there's a universal law of hiring. I agree that "hiring someone like me" is a common -- but not universal -- pattern.

(Oddly enough, in the USA in 2026, I think you'll find that white males are the _least likely_ to use the "someone like me" hiring strategy - if for no other reason than decades of targeted enforcement against that that group for appearing to employ that strategy. Eg the Indian preference for hiring Indians, or the female preference for hiring females, is stronger than the white-male preference for hiring white-males. Maybe not in terms of hidden preferences, but in terms of actual hiring outcomes, yes, that's what the data will show)

Back to the non-universality of hiring strategies: A friend of mine with a tech business intentionally tries to hire mid-field performers. He doesn't want a rapid climber, because he cannot afford to keep them on staff once they understand what they are worth. He wants mid-field performers who are going to do solid, reliable work, and be content to work for him for a long time. So he doesn't hire on "merit" - he hires for "good enough, low turnover"

Other businesses have other strategies.

> but "not understanding cost" is a dangerous assumption to make.

One of the best, most succinct phrases in this discussion. Thank you.

We don't have to wait for AGIs to develop their own agency and build terminator robots. Industrial robots with no agency and with auditable and deterministic programming already kill humans. The industrial robot doesn't understand there's a human in the safety cage with it. It doesn't understand that moving through space occupied by a human has a cost. It doesn't know there's a human there and it doesn't know what a human is.

Designing automation that can never kill a person is already an unsolved problem with standard, non-intelligent technology.

I don't know if AI makes it harder or easier. On one hand, AI behaves, in effect, non-deterministically (even though at some level it is deterministic, it's not observably so).

OTOH, in organic intelligences, adding sensors, adding alignment, giving feedback - tend to produce humans who accidentally kill less frequently. So maybe something like the 3 laws will conceptually be part of the System Prompt of future AIs, and they'll figure out how to avoid HAL 9000 problems.

Anything that challenges your assumptions must be a joke. Therefore you never need to change your mind on anything!

Thanks for the interesting response.

The strongest man today; the hardest workers today... are not 10x or even 2x as strong or as physically productive as the best in human history. Depending on what you think about ancient man, we are possibly weaker and dumber than some of our ancestors.

The transmission of new ideas across memory, time, and place is the extra term in the equation of human achievement, civilization, and prosperity. It's not that labor doesn't matter. It's that ideas are like compound interest. Labor adds principal to our civilizational account balance. But Ideas compound it.

In the long term, ideas are all that matter.

I have the same reservations you do - powerful people will continue to want to corner and concentrate their power.

It's unclear how well they will be able to do that, however. The capability lag time between a paid frontier model and a self-hostable open-weights model is months.

The ability to train a good-enough model will also be commoditized. In my youth, we had distributed crypto challenges.... to get super-computing work done on mass democratized hardware of like-minded individuals. So in the future, we can fall back on analagous distributed model training efforts if some better alternative doesn't exist.

I am skeptical about a post-scarcity society and I am skeptical about man's "labor" ever going to zero. Of the infamous twelve possible AI outcomes, I am somewhere in middle. For biblical reasons, I believe the extinction of man is not going to happen, and I believe the utopia of man is also not going to happen.