Can “effective altruism” maximise the bang for each charitable buck? • Hacker News Board

Re: Can “effective altruism” maximise the bang for each charitable buck?

#201

Post by buvanshak » Sat, Jun 02, 2018, 10:49 AM UTC

gowld wrote: post #102

buvanshak wrote: post #60

Earlier quoted context omitted.

I cannot make much sense of your comment. But let me expand on my original comment and tell this. The only charitable work that I consider genuine is the one that you does not know about. When you read about the work of a person 'x' of a foundation in papers/media, their work cease to be charity. It is just that the person is gaining something, in return, other than "feeling good" you mention. Superficially, or in ot…

If the only genuine charity is invisible, how can we maintain a society with a culture of genuine charity? If charity is secret, it's too easy to simply not do charity, since there are no consequences, neither positive nor negative reinforcement. I have my suspicions about the PR motives of https://givingpledge.org/ , but ultimately, doing good in exchange for positive PR (a form of "buying advertising") is a deal I'…

>If charity is secret, it's too easy to simply not do charity,..

That is like saying if every one have sex in secret, then it is simply too easy not to have sex, so no one will do it.

The act of charity itself is the reward, and not its attribution.

Re: Can “effective altruism” maximise the bang for each charitable buck?

#202

Post by emodendroket » Sat, Jun 02, 2018, 12:22 PM UTC

perl4ever wrote: post #197

emodendroket wrote: post #194

Earlier quoted context omitted.

Here's a classic problem for utilitarianism: You are a doctor. You have four patients. Three are young men in otherwise good health who need an organ or they will die very soon -- two of them need kidneys and one of them needs a heart. Your fourth patient is a homeless person who's broken his leg but is otherwise well. You know for a fact he has no friends or family who will notice he is gone, and that nobody will kn…

I don't think "utilitarianism" incorporates any principle that homeless people are worthless. Really, depending on how you define utility, it can be molded into anything. Note that harvesting organs means he will definitely die, but the organ transplants may fail, or may only work for a limited time. It could be rather than assuming "two lives saved minus one life lost" you should use "probability/quality adjusted ye…

Who said it did? The point is that saving multiple lives produces greater utility than allowing one person to live, so if utility is your only consideration it is fine to murder people to save lives. The point of making the donor homeless and a loner is just to remove objections about how eventually someone is going to notice nobody is paying the rent, or you'll make other people sad by killing him, or you'll reduce their trust in the medical system, or various other answers that sidestep the problem.

Re: Can “effective altruism” maximise the bang for each charitable buck?

#203

Post by emodendroket » Sat, Jun 02, 2018, 12:25 PM UTC

Vinnl wrote: post #198

emodendroket wrote: post #194

Earlier quoted context omitted.

Here's a classic problem for utilitarianism: You are a doctor. You have four patients. Three are young men in otherwise good health who need an organ or they will die very soon -- two of them need kidneys and one of them needs a heart. Your fourth patient is a homeless person who's broken his leg but is otherwise well. You know for a fact he has no friends or family who will notice he is gone, and that nobody will kn…

> I'm suspicious of any "formula" to solve all ethical problems. Right, I think that's what I'm mostly interested in - what if you treat utilitarianism as a guideline? If you only apply it in cases where you think most people would agree on what is "most useful", would that not lead to better decisions on average, without influencing potential hypothetical edge cases? (I also happen to think that the most useful thin…

If utility is simply one factor of many then we're not doing utilitarianism.

I don't think the law thing works at all. Laws can be and often are unjust. Imagine I live in a twisted society where the doctor killing patients to harvest organs is legal. Is it now ethical? This is not purely theoretical -- the people operating concentration camps in Germany weren't violating German law.

Re: Can “effective altruism” maximise the bang for each charitable buck?

#204

Post by Veedrac » Sat, Jun 02, 2018, 1:51 PM UTC

the_af wrote: post #193

Veedrac wrote: post #190

Earlier quoted context omitted.

I am aware of one reproduction of the experiment, the goals seem pretty darn explicit, and its results are public. He has stated the rules of engagement, and has said that he did it "the hard way". If nothing else, one should at least be confident that Yudkowsky is honest. His claim that "a transhuman mind will likely be able to convince a human mind" is what his experiment demonstrates , not what is assumes , and fr…

- What reproduction? What would you consider a successful reproduction, for that matter? If I told you I reenacted the experiment at home with a friend, would you consider this a reproduction? Someone saying they reproduced it on the internet would convince you? What are your standards of quality? - What is the goal of the experiment? Is the goal "show that a transhuman AI can convince a human gatekeeper to set it fr…

The only experiment I know of and would consider a serious attempt at reproduction would be Tuxedage's series,

https://www.lesswrong.com/posts/FmxhoWxvBqSxhFeJn/i-attempte...

https://www.lesswrong.com/posts/dop3rLwFhW5gtpEgz/i-attempte...

https://www.lesswrong.com/posts/oexwJBd3zAjw9Cru8/i-played-t...

His total is 3 for 3. I do not know how to explain these results without either taking them to be honest attempts at a fair experiment or by assuming those involved colluded. I find the latter absurd, given my priors about the honesty of members of LessWrong (Yudkoswky in particular, though he wasn't involved in the reproduction).

> If I told you I reenacted the experiment at home with a friend, would you consider this a reproduction? Someone saying they reproduced it on the internet would convince you?

It is not so simple. I would want evidence that you and your friend were smart and had a decent understanding of the domain, and that your friend was in a similar state of unbelief about the plausibility of being convinced. I would want a statement that it was a serious attempt at doing things "the hard way" and true to the experiment, on both sides, lest you get [1]. Of course I would want the standard rules, or a reasonable modification publicly stated, in addition.

[1] https://pastebin.com/Jee2P6BD

> What is the goal of the experiment?

To show that "I can't imagine anything that even a transhuman could say to me which would change [my mind]" is not evidence, and should not be treated as such. To provide evidence that "humans are not secure systems".

You say "very low stakes", but Yudkowsky convinced someone who had offered a $5000 handicap. That hardly seems like a trivial quantity.

> [maybe it's all a lie]

You have to be very cynical to take this worldview.

> we don't even know what a "transhuman" mind would be like

The experiment is under assumption of a true singularity, ergo. nigh-unlimited intelligence. I can discuss what outcomes I think are likely for AI development, or which are merely plausible, but the experiment is about one particular hypothetical, so that would be a different conversation.

> the participants seem to be selected from a pool of people fascinated by this kind of questions and who would be open to suggestion that a "transhuman" mind can convince them of stuff

I am unconvinced that this experiment would work if the gatekeepers did not have an understanding of the topic; they are meant to play a gatekeeper, after all. A person who considers the singularity plausible but thinks an AI box is effective seems like the perfect control should the singularity happen and people want to figure out whether to AI box it.

Re: Can “effective altruism” maximise the bang for each charitable buck?

#205

Post by probably_wrong » Sat, Jun 02, 2018, 4:47 PM UTC

justinpombrio wrote: post #170

probably_wrong wrote: post #103

Earlier quoted context omitted.

I assume that investing in AI risk is nonsense for one simple reason: I have yet to see a single study showing that a single undesired behavior has been stopped or even slowed down. DDOS attacks are as stupid as you can get from an intelligence point of view, and yet not a single proponent for AI risk has come up with a way to stop even a single one AFAIK. I haven't heard either of a single military program that has…

I'm confused. The research in question is "how can we make sure that a superintelligence would be morally aligned with humans". Why would dealing with DDOS attacks would be the right first step? It seems completely unrelated.

Because (IMHO) if they want to morally align a superintelligence, the first step would be to either morally align a dumb intelligence or to show that you can convince the people building this intelligence to steer it in this direction.

If they do neither, they risk either coming up with a plan that doesn't work (because it was not tested) or that no one cares about.

Re: Can “effective altruism” maximise the bang for each charitable buck?

#206

Post by the_af » Sat, Jun 02, 2018, 5:23 PM UTC

Veedrac wrote: post #204

the_af wrote: post #193

Earlier quoted context omitted.

- What reproduction? What would you consider a successful reproduction, for that matter? If I told you I reenacted the experiment at home with a friend, would you consider this a reproduction? Someone saying they reproduced it on the internet would convince you? What are your standards of quality? - What is the goal of the experiment? Is the goal "show that a transhuman AI can convince a human gatekeeper to set it fr…

The only experiment I know of and would consider a serious attempt at reproduction would be Tuxedage's series, https://www.lesswrong.com/posts/FmxhoWxvBqSxhFeJn/i-attempte... https://www.lesswrong.com/posts/dop3rLwFhW5gtpEgz/i-attempte... https://www.lesswrong.com/posts/oexwJBd3zAjw9Cru8/i-played-t... His total is 3 for 3. I do not know how to explain these results without either taking them to be honest attempts at…

But that's just it: I simply don't think Yudkowsky or any of the sort of people who would be enthusiastic about the sci-fi theories on SL4, or host extropy.org, or believe in Roko's Basilisk, or read Harry Potter fanfic and find it philosophically insightful, have a decent understanding of the AI domain. Everything about him and his followers smacks of fringe cultists completely outside mainstream research.

I don't think the chosen participants have a particularly deep understanding of the domain, they just think they do (because that's what defines Singularity believers, LessWrong readers, and people who believe they are hyper "rational" and that this is some kind of superpower). I think they understand AI no more than a Star Wars fan understands space travel.

Re: Can “effective altruism” maximise the bang for each charitable buck?

#207

Post by Veedrac » Sat, Jun 02, 2018, 5:29 PM UTC

the_af wrote: post #206

Veedrac wrote: post #204

Earlier quoted context omitted.

The only experiment I know of and would consider a serious attempt at reproduction would be Tuxedage's series, https://www.lesswrong.com/posts/FmxhoWxvBqSxhFeJn/i-attempte... https://www.lesswrong.com/posts/dop3rLwFhW5gtpEgz/i-attempte... https://www.lesswrong.com/posts/oexwJBd3zAjw9Cru8/i-played-t... His total is 3 for 3. I do not know how to explain these results without either taking them to be honest attempts at…

But that's just it: I simply don't think Yudkowsky or any of the sort of people who would be enthusiastic about the sci-fi theories on SL4, or host extropy.org, or believe in Roko's Basilisk, or read Harry Potter fanfic and find it philosophically insightful, have a decent understanding of the AI domain. Everything about him and his followers smacks of fringe cultists completely outside mainstream research. I don't t…

Sure, I don't particularly care if you or anyone else wants to disengage from LessWrong-esque ideas because they sound weird. I only entered this discussion because it sounded like you might have had an actual argument.

Re: Can “effective altruism” maximise the bang for each charitable buck?

#208

Post by davidivadavid » Sat, Jun 02, 2018, 6:10 PM UTC

emodendroket wrote: post #202

perl4ever wrote: post #197

Earlier quoted context omitted.

I don't think "utilitarianism" incorporates any principle that homeless people are worthless. Really, depending on how you define utility, it can be molded into anything. Note that harvesting organs means he will definitely die, but the organ transplants may fail, or may only work for a limited time. It could be rather than assuming "two lives saved minus one life lost" you should use "probability/quality adjusted ye…

Who said it did? The point is that saving multiple lives produces greater utility than allowing one person to live, so if utility is your only consideration it is fine to murder people to save lives. The point of making the donor homeless and a loner is just to remove objections about how eventually someone is going to notice nobody is paying the rent, or you'll make other people sad by killing him, or you'll reduce…

"The point is that saving multiple lives produces greater utility than allowing one person to live, so if utility is your only consideration it is fine to murder people to save lives."

That depends on what utility function you're using.

Re: Can “effective altruism” maximise the bang for each charitable buck?

#209

Post by emodendroket » Sat, Jun 02, 2018, 6:24 PM UTC

davidivadavid wrote: post #208

emodendroket wrote: post #202

Earlier quoted context omitted.

Who said it did? The point is that saving multiple lives produces greater utility than allowing one person to live, so if utility is your only consideration it is fine to murder people to save lives. The point of making the donor homeless and a loner is just to remove objections about how eventually someone is going to notice nobody is paying the rent, or you'll make other people sad by killing him, or you'll reduce…

"The point is that saving multiple lives produces greater utility than allowing one person to live, so if utility is your only consideration it is fine to murder people to save lives." That depends on what utility function you're using.

I can't think of one that avoids this problem. Can you?

Re: Can “effective altruism” maximise the bang for each charitable buck?

#210

Post by perl4ever » Sat, Jun 02, 2018, 7:04 PM UTC

emodendroket wrote: post #209

davidivadavid wrote: post #208

Earlier quoted context omitted.

"The point is that saving multiple lives produces greater utility than allowing one person to live, so if utility is your only consideration it is fine to murder people to save lives." That depends on what utility function you're using.

I can't think of one that avoids this problem. Can you?

I did throw out "probability/quality adjusted years of life gained" to address the obviously grossly inadequate metric of counting "lives gained/lost".