In fact, why do we have to talk about intelligence? What about the Kardashians? What would they do if they showed up here? What would we do if we met the Kardashians? Would we try to eradicate them?

This article isn't about existential risks in general, it's specifically about AI.

There is one good reason to assume that: we are their creators, not a series of random processes smoothed out by natural selection. That has several consequences:

1. We can (attempt to) create strong AI in such a way that it doesn't want to kill us, or is unable. The want thing could have bugs, but we can work through bugs. The ability thing seems stronger at first brush -- consider a strong AI whose entire existence is inside a virtualized environment and which has no direct external sensors -- essentially an AI in the matrix, a matrix which may or may not even be recognizable to humans. It is not going to be able to conceive of a way of killing its extradimensional overlords.

The other problem is if some crazy person subverts this because they want apocalypse. Suicide bombing the species, so to speak.

2. We are their creators and they know it. You can twist that around and also say that's a reason they'll destroy us, but it is nonetheless unique to AI compared to every other example you have cited.

3. There is no reason that strong AI needs to have a survival instinct built in. It's fundamental to us because if it wasn't, we wouldn't live to produce as many children as we do (and even then suicide rate is alarming in humans, and documented in other mammals). I can think of no reason it's logically necessary to a strong AI to have a survival instinct. I know that's Asimov's third law of robotics.

In fact there's no reason to presume just about anything about the motivations of an AI. We'll probably try to give it some, which align with our motivations, so it's probably not entirely distinct from a human motivation, but that doesn't mean its actions overall would be human-like.

I will give one big exception: if strong AI is first developed by full-brain simulation of a human, that would imply human-like motivations.

Errr...did they change the rules that day? Or are you treating sunk costs oddly?

There are lots of studies that use equivalent consequences that show this to be the case. You can see citations just reading the summary paragraph of the wikipedia article: http://en.wikipedia.org/wiki/R...

They are referencing this well-known result, not using their work as evidence for it.

Canada (at least some provinces) gets by with a separate, lower minimum wage for high-schoolers. Pretty much annihilates the "but the YOUTHS" argument.

Your scenario is broken because you believe that doubling some people's earnings causes all prices to double, which is absurd.

The total money circulating in the marketplace did not come even close to doubling, so there's no reason to expect 100% inflation. Prices will probably rise marginally on low-end goods.

In fact, if you passed a law that said that literally everybody on earth got their wages exactly doubled instantly, you *still* wouldn't see a precise doubling of prices, because you haven't increased their existing asset base.

If you double literally all incomes on Earth, *and* the value of all financial instruments, overnight, then and only then do I expect everything's price to also double overnight (provided there's no legal impediment to doing so eg. rent control). Nobody is doing that.

It's like Lenz's Law. Yes, running a current causes electromagnetic interference and, ultimately, a countercurrent. No, that doesn't mean that electric current is impossible.

No evidence? That is itself a pretty strong claim.

Here's some really quick ones:

Effective: http://www.vaccines.gov/basics...
Safe: http://www.cnn.com/2014/07/01/...

These are just two of them. These results are so well-known (especially with people who lived when measles was rampant) that citations are generally regarded as unnecessary to provide -- anybody can look up the source on their own. You can maybe find a vaccine (particularly one not yet FDA-approved) that isn't safe and effective because it's a broad category -- it's like asking whether "liquids are safe and effective at quenching thirst", and the answer is yes, but don't drink mercury or poison or anything that isn't safe and effective.

There seems to be confusion as to what Quantum Mechanics is.

Quantum Mechanics is accurate enough that it is necessary to modern semiconductor design. There is no other model humankind has invented that works. Therefore it is a good answer to the GGGP's question, "So why is this nonsense still science?".

Your misgivings about entanglement are not actually relevant to whether Quantum Mechanics is a real thing. Consider: Classical mechanics makes predictions about how GPS satellites should work that are empirically incorrect, which is why General Relativity is involved in GPS systems. Thus, even though Classical Mechanics makes a prediction that we are uncomfortable with, it's still not nonsense.

The retort "Make QM based computing work. Ever." is also irrelevant. Doesn't matter whether QM based computing is practical. The existence of entanglement doesn't mean that quantum computers are a good idea, any more than the explanation for how sunlight reaches the Earth means that classical computers are real. They happen to be real, and they both operate on electromagnetism, but there is a huge disconnect between those two things.

As for entanglement itself, it's a confusing subject but the GGGP didn't do a great job refuting it.

Because "two forensic softwares" is ungrammatical.

It's like asking why somebody said they had two works of art when "works of art" could be replaced with just "art" without confusion. You can't just say you have "two arts". You could say you have "some art" but now you've lost information from the sentence.

Program is countable, but software is linguistically uncountable.

Can a good driver dodge accidents from the passenger's seat? Because 2 of those accidents were in self-driving cars with the self-driving turned off.

Without details on the accidents, we aren't in a good position to judge whether a good driver could have dodged this accident. The one detail we have is that the car was moving 10mph at the time. Not sure what to make of it -- was it read-ended, was it a parking lot accident, was it a bad lane change? Because being read ended is pretty cut-and-dried somebody else's fault (unless you *just* completed a lane change or something like that), but a parking lot accident could be very concerning.

Not that I disagree that the technology should be proven!

"It's disgusting that everybody doesn't immediately hail a poorly formatted near-unreadable mess of a post as among greatest. Also everybody who doesn't agree with it is literally evil and has rabies."

What the fuck. Seriously, the guy you're responding to had more weasel words than genus Mustelidae. How can you possibly read that and then act like he said all women are great forever and all men suck?

This is beyond a strawman. It's a whole straw village. And frankly, such an obviously warped of interpretation makes me struggle to give you the benefit of the doubt for the rest of your story, but I'm going to try not to dismiss it out of hand because I know that attitude is a problem for both men and women -- especially men. So: I agree that you shouldn't have been raped or falsely accused of being a rapist, that's shitty.

This said...

That term means something different from how you're using it. You're using it to mean "feminist" really, but not using the word feminist because you want to exclude radical feminists and academic feminists. I have little patience for semantic political correctness arguments like this. You can't use "humanist" is a politically correct way of saying "feminist". It's taken.

The assumption here is that 100% of the cause for every notable gender imbalance today is due to inherent differences. I think that's pretty naive. Can you name the year when equality happened? Because what I observe is consistent change from 200 years ago, which basically agrees was non-equal. On what basis do we say that today, or 10 years ago, or whatever was the time when we hit equality?

Frequently on slashdot even the suggestion that we formally study the reasons instantly triggers huge comment threads about inherent differences, bringing up red herrings like conscription or circumcision -- things that are generally agreed to be unfairly anti-male as if it disproved that there was anything unfairly anti-female. If you can flip it around and have somebody say they won't support ending conscription for men only until there is pay equity for women, it makes exactly as much sense.

When you see a persistent difference, there's two possibilities. There's an inherent difference, or there's a cultural difference (or some combination). The instantaneous assumption that all observed differences are inherent is extremely convenient. Even if true, it's worthy of testing.

So yes, a true meritocracy will end up with differences due to various natural factors*. But we haven't seen a true meritocracy so we don't know what it would actually look like.

*Actually, there's reason to fear that a true meritocracy is an unstable system, almost inevitably devolving into a less globally efficient one due to the local efficiencies of stereotyping. After all, it's not like you can rationally say that our 6 million year ago ape ancestors had an unnatural society. But there's clearly been weird cultural biases in many times and places. Any unreasonable cultural biases came about sometime.

No, it literally does mean that nothing else can affect us. If we are affected, we can observe these effects.

I'm pretty sure it's from the reference to the Alcubierre Drive. The summary on Wikipedia (which is what 99% of slashdotters know of the drive) states that it requires negative mass.