The News Thread

[Einherjar86]

That quote was from the same article you linked.

If you could give a specific example of his "leaps of character," that would make it easier to address. As far as I can tell, he's simply going off of what Musk, Zuckerberg, and the like have already said about AI. His main point seems to be that tech moguls project an unconscious skepticism toward their own behavior onto the prospect of superintelligent AI. In and of itself, that's a very speculative suggestion, and not one that has a lot of argument backing it; it's a kind of cultural critique performed very succinctly. It's a thought piece, not much more; but people like Chiang are often asked for such pieces, given their extensive body of work.

I do think there's been some push-back however from people who have somewhat misinterpreted the piece. I don't think Chiang is giving us his stance on AI one way or another. I think he's just comparing contemporary visions of superintelligent AI to corporate models and systems. This isn't a totally ridiculous analogy if one thinks in terms of emergence and complexity theory. Emergence has been used in the sciences to explain phenomena ranging from termite colonies to the internet to processes of urbanization. Chiang is just picking up on a theory of complexity that observes comparable patterns between corporate entities and hypothetical superintelligent systems.

 

[rms]

I know it's from the same article, it's what made me go from "awesome headline" to "god damn it wasn't really about the headline but about capitalism"

Billionaires like Bill Gates and Elon Musk assume that a superintelligent AI will stop at nothing to achieve its goals because that’s the attitude they adopted. (Of course, they saw nothing wrong with this strategy when they were the ones engaging in it; it’s only the possibility that someone else might be better at it than they were that gives them cause for concern.)

you wouldn't say this is a leap of character for both Gates and Musk?

that's a very speculative suggestion, and not one that has a lot of argument backing it;

agreed, which is why I say it's a leap of character. Musk made his money on PayPal, not enslaving a foreign people and forcing them to do cheap labor.

it's a kind of cultural critique performed very succinctly.

I would say it goes past a cultural critique once names are thrown into the mix

Chiang is just picking up on a theory of complexity that observes comparable patterns between corporate entities and hypothetical superintelligent systems.

I thought the piece was going to be about machine learning and how we, as humans, don't know how to properly utilize it. I was mistaken

 

[Einherjar86]

I know it's from the same article, it's what made me go from "awesome headline" to "god damn it wasn't really about the headline but about capitalism"

Sorry, the way you said "unless he's done character pieces before on Gates and Musk in the quote you cited" made me think you assumed the excerpt was from something else he wrote.

you wouldn't say this is a leap of character for both Gates and Musk?

I wouldn't call it a leap of character. I think it's a legitimate interpretation based on their observable behavior and past comments about AI. It's only a leap of character to the extent that it makes assumptions about someone's character based on their actions.

agreed, which is why I say it's a leap of character. Musk made his money on PayPal, not enslaving a foreign people and forcing them to do cheap labor.

I don't think that's quite what Chiang has in mind.

He's comparing corporate entities to superintelligent AI, but Chiang never blames an AI for its behavior. Basically, he's saying that superintelligent AIs would evolve and behave in accordance with their prime directive, and would optimize their functionality in order to achieve their goals efficiently. Since machines lack insight, they wouldn't acknowledge when their behavior poses a legitimate ethical concern.

Likewise, Chiang notes that corporations also exhibit no insight; we don't expect corporate entities to reflect upon their decisions, which is why some level of government oversight is (or should be) mandatory.

Chiang is pointing out that the de facto leaders of these tech giants don't seem to think it's necessary for them to promote any culture of insight or self-reflection within the corporate community, or as a part of the corporation's prime directive. They justify their corporation's potentially unethical behavior by appealing to a corporate body's prime directive--i.e. to deliver the goods to its stakeholders by making as much money as it can. Yet people like Musk and Gates have no trouble suggesting that superintelligent AIs might be dangerous because they lack insight.

I don't think he's criticizing Musk's or Gates's personal character. I don't even think he's saying corporations shouldn't necessarily do what they do, or that superintelligent AIs shouldn't be created. I think he's just trying to say that there's a little bit of a double-standard going on here.

 

[rms]

Likewise, Chiang notes that corporations also exhibit no insight; we don't expect corporate entities to reflect upon their decisions, which is why some level of government oversight is (or should be) mandatory.

I disagreed with this claim as well. What Chiang probably meant here is benevolent insight, but I still wouldn't agree.

I don't think he's criticizing Musk's or Gates's personal character

the thesis of the piece is : shitty business leaders fear AI because AI will act like them (shitty and unethical). How is this not an attack of character?

If Musk's words on AI have been so damning of his own character you'd think any quote of his would make it into the piece and later be analyzed. Hell, his own admission is a story about how the machine would have no capacity for human empathy while the average business would.

Thus, in its pursuit of a seemingly innocuous goal, an AI could bring about the extinction of humanity purely as an unintended side effect.

Musk can say this because no corporation has done this. Because they have insight!

 

[Einherjar86]

I disagreed with this claim as well. What Chiang probably meant here is benevolent insight, but I still wouldn't agree.

I think he's just talking about ethical reasoning. A corporation can't ethically reason any more than an ant colony can. Individuals can ethically reason, and can institute ethical parameters; but it's not incumbent on corporations to do this. That's part of the point.

the thesis of the piece is : shitty business leaders fear AI because AI will act like them (shitty and unethical). How is this not an attack of character?

Fair point. I don't think it's a stretch to say that Chiang has an opinion of the character of these business leaders; but I also think this requires more extensive commentary than Chiang has room for in this piece.

Ethics in business is a really tricky subject because ethical responsibility can be said to apply in multiple contradictory areas. A company is ethically responsible to its workforce; but privileging ethical responsibility to workers means a company is less responsible to its stockholders. If a business's primary goal is to make money, then it's difficult to see how ethical concerns should even fall within the scope of a company's operations. This is why an outside system is needed to regulate the company in a manner that is palatable to the consumers who use said company's products/services.

This is part of the point Chiang makes: that AIs, like multinationals and tech conglomerates, have no ethical scope. It's not built into their functionality. Chiang certainly blames "capitalism" for this (albeit a vague, abstract notion of capitalism), but I still don't think we can read a full-fledged condemnation of capitalism in his piece. It's more like an astute observation of the contradictory applications of ethical regulation--i.e. that tech moguls have no problem issuing ethical demands to others but not their own companies.

If Musk's words on AI have been so damning of his own character you'd think any quote of his would make it into the piece and later be analyzed. Hell, his own admission is a story about how the machine would have no capacity for human empathy while the average business would.

I'm pretty sure Chiang is assuming that Musk's stance on AI is well-known by this point. It's virtually common knowledge.

Thus, in its pursuit of a seemingly innocuous goal, an AI could bring about the extinction of humanity purely as an unintended side effect.

Musk can say this because no corporation has done this. Because they have insight!

It's not because corporations have insight; it's because they're subject to oversight (and even then, probably not enough oversight--but that's another conversation). Chiang is pointing out that corporations don't see the need for their own regulations but do see the need for AI regulation.

 

[Dak]

I scanned most of the headlines at Buzzfeed today. Still trying to decide if Buzzfeed or Huffpost are more vacuous. Even FOXNews looks like a serious media site by comparison.

For such a supposed visionary, Chiang makes a series of moves common to these sorts of soft-hit pieces:

1. Identify trendy topic
2. Pick some big names with controversial or maybe even popular stances
3. Lump these disparate big names together because reasons
4. Complain about their businesses and/or policies
5. Blame Capitalism because nominated thing I don't like is in the world and has a place in the global marketplace.

To address one concern from the article:
Corporations operate for the interest of humans, even if not for every single one. There is no reason to expect that AI will eventually operate in the interest of even one human.

 

[Einherjar86]

I scanned most of the headlines at Buzzfeed today. Still trying to decide if Buzzfeed or Huffpost are more vacuous. Even FOXNews looks like a serious media site by comparison.

 

[Dak]

At this moment, 2 of the top 3 headlines at Buzzfeed.com are "13 Impossible Decisions for Netflix Addicts" and "18 Times Anna Kendrick was the greatest person on the internet"

If you click over to "News" (The fact that you need to click to another area of the site is noted), it's just "omg sexual harrassment" all the way down, plus PS "omg look at the gop tax plan".

Edit: Seinfeld looks like Andy Sandberg for a second in that gif. Dem Jew genes.

 

[tagradh]

Does anyone even think of BuzzFeed as a real news site, though?



I like that the US ambassador to the UN reminded everyone that Trump would take personally voting against America, as if there's anyone out there unaware of the fact that he's a total child. And that no one gave a fuck about this threat, anyway.

Taking climate change outa the National Security Strategy seems pretty fucking daft but I guess there are lobbyists and morons to win over. Who cares if climate change, regardless of its cause, is going to seriously impact large parts of America, destabilize large parts of the world in which the US has interests, see the development of technologies and industries that, for whatever reason, America would rather countries like China dominated...

 

[Einherjar86]

If you click over to "News" (The fact that you need to click to another area of the site is noted), it's just "omg sexual harrassment" all the way down, plus PS "omg look at the gop tax plan".

I'm sorry, I wasn't aware those topics aren't news.

 

[HamburgerBoy]

Taking climate change outa the National Security Strategy seems pretty fucking daft but I guess there are lobbyists and morons to win over. Who cares if climate change, regardless of its cause, is going to seriously impact large parts of America, destabilize large parts of the world in which the US has interests, see the development of technologies and industries that, for whatever reason, America would rather countries like China dominated...

Not familiar with the specifics of what you're talking about; is there some proposed law/treaty/trade agreement/whatever that curtails China's pollution/productivity relative to that of America and greener Europe?

 

[Dak]

I'm sorry, I wasn't aware those topics aren't news.

Well if they labeled it "Titillating Sexual Harrassment News" it would have been understandable. Based on the sum total of articles on Buzzfeed, and the ridiculous tweets from BF editors, it's pretty obvious the writers and editors live in a very tiny "bougie bubble", one of enough privilege where serious matters given careful attention are "like, so totally boring <vocal fry>". Or maybe just a dog whistle or something.

 

[TechnicalBarbarity]

this happened a few days ago here...


excessive force much?

 

[viewerfromnihil]

Maybe if some of these pigs got fired from time to time it wouldn't happen so often. I really don't understand the end game of the police unions defending these actions tooth and nail, nor how de-escalation training hasn't been implemented comprehensively in the country since this problem hit the mainstream in the 90s.

 

[HamburgerBoy]

I don't see the excessive force in that one. Dude was resisting from the start, and then went full "I'd like to speak to the manager" a minute in. Maybe that very last baton whack on the leg was unnecessary and that may have been the one that broke it, but whatevs.

EDIT: Wait, he stood up and grabbed the baton from the officer before that. What a fucking moron. Lucky he wasn't shot.

 

[Dak]

Maybe if some of these pigs got fired from time to time it wouldn't happen so often. I really don't understand the end game of the police unions defending these actions tooth and nail, nor how de-escalation training hasn't been implemented comprehensively in the country since this problem hit the mainstream in the 90s.

There are plenty of cases where the force isn't excessive to the actual threat (not the one the bleeding hearts try to paint - EG Michael Brown) but the problem is that the police unions defend every case the same, which weakens their claims.

 

[Satanstoenail]

A much bigger and more powerful button.

 

[Serjeant Grumbles]

The bigger the button, the bigger the nukes it can launch. That's just science.

 

[Dak]

The funny part is Trump actually can launch nukes. You guys can't even launch memes.

 

[Satanstoenail]

The funny part is Trump actually can launch nukes.

Failing to see the humour there.