https://www.reddit.com/r/ControlProblem/wiki/faq

Almost seven million people have died to COVID...

Are you??? Probing beliefs and asking question is a very important part of philosophy.

>they don't feel pain.

This is why, not their lack of cognitive ability.

I'll concede that I don't know how tolerated irl action was on the site.

I'm still completely fine with Google preventing the site from showing up in results due to its content. The government shouldn't stop the site from existing, but Google is well within its rights, and is doing the right thing, by not providing easy access to it.

Transphobia (and other hate speech) is bad. Spreading it is bad, platforming it is bad.

>Complete and utter bullshit. If you...

You're being too vague. Do you think I'm referring to insulting people online as harassment?

No, people organized serious harassment campaigns on Kiwi Farms, often with the intention to drive people to suicide. This included forms of irl harassment like: swatting, doxing, identity theft, and more. None of this is allowed on Twitter and will get you banned and your messages deleted, as it should be.

>Somehow the Internet ran fine...

I don't know how to respond to this. It's too vague and unsubstantiated. I'm sure the internet "ran fine", but that doesn't mean I'd be ok with hate speech and harassment campaigns being hosted on popular platforms.

>What do you define as harmful? Do you realise how subjective a standard there is?

The fact that there are ambiguous cases doesn't mean you can't construct good terms of service. No one will ever be fully satisfied by the rules, but that doesn't mean we shouldn't have them.

Examples of harmful behavior that should definitely be banned were mentioned in my first comment. Kiwi Farms, for example, was a source of organized harassment of trans people in real life. Google shouldn't allow easy access to the site.

>Do you realise that there is no moral duty to engender an AI with the cultural and social sensitivities of today, forever. Do you realise how horrifying an idea that is, a stagnant thought-policer without the ability to adapt?

Please quote where I said that AI should forever be constrained by all and only today's social values.

Should I be able to comment slurs at you, right now, with no consequences? Should I be entitled to Reddit continuing to host my comments if I were to start posting hate speech all over the website? What if I were to organize harassment, encourage suicide on the platform?

All platforms need limits on free speech to keep the space tolerable. And further, there's a moral duty to not host harmful content.

Accepting the premise that wealth generated by a company is attributable to all of its laborers does not lead to the conclusion that work is irrelevant. That just doesn't follow at all.

The relevant conclusion is that the wealth that pays for public transit is attributable to all laborers, of which bosses are a minority.

Organizing work is a type of labor, yes. Not sure why you think that's a gotcha.

Ultimately all wealth can be traced to labor. Even in twenty thousand years, when the galaxy is being turned into computing power by swarms of automated robots, you will be able to trace a line back to the labor of humans. Without low level workers, production would not be possible.

Again, I'm not even putting forward the claim that they're deserved that wealth. But denying that they're responsible for it is stupid.

Try running a multi-billion dollar business without janitors. Or cashiers. Or stockers. See how that goes in reality.

This isn't even entirely Marxist yet. I haven't stated an opinion on who should own the results of labor. But the labor of poor people is worth far more than they receive in income. It's just undeniable.

Poor people create rich people's wealth. The tax subsidies used for public transit may not come directly from poor bank accounts, but it is the result of their labor.

What does charity have to do with this? We're talking about policy.

I'm 99% sure it'll be a trivial amount of computation within 100 years. We still have plenty of avenues to explore.

Modern large neural networks usually run on GPUs, not CPUs

I don't think we should be applying the same standards to AI, which will be the most important technology in all of history. The way it proliferates is going to make a large impact on humanity's future.

Of course, DeepMind is perfectly within their legal rights to not share their models. But ethically speaking, they should... or maybe they should be trying to keep AI out of the hands of the public at all costs. Either way, this is too important to rely on copyright laws for answers.

I didn't mean it would literally run on any computer, that's not my point.

>If someone builds a rocket, you don't get to go for a ride. Why do you think that you should just be able to use a model that someone else designed and trained?

Not the same thing. A rocket is a limited resource: it can only be used by so many people at once. An AI model has no real scarcity: it can be copied and run on anyone's computers with no limits.