Zeke_Sherman

Karma: 99

Zeke_Sherman 14 Dec 2018 15:24 UTC
1 point
0 ∶ 0
on: Critique of Superintelligence Part 4
A lot of baggage goes into the selection of a threshold for “highly accurate” or “ensured safe” or statements of that sort. The idea is that early safety work helps even though it won’t get you a guarantee. I don’t see any good reason to believe AI safety to be any more or less tractable than preemptive safety for any other technology, it just happens to have greater stakes. You’re right that the track record doesn’t look great; however I really haven’t seen any strong reason to believe that preemptive safety is generally ineffective—it seems like it just isn’t tried much.

Zeke_Sherman 14 Dec 2018 15:11 UTC
2 points
0 ∶ 0
on: Critique of Superintelligence Part 3
For low probability of other civilizations, see https://arxiv.org/abs/1806.02404.
Humans don’t have obviously formalized goals. But you can formalize human motivation, in which case our final goal is going to be abstract and multifaceted, and it is probably going to be include a very very broad sense of well-being. The model applies just fine.
Because it is tautologically true that agents are motivated against changing their final goals, this is just not possible to dispute. The proof is trivial, it comes from the very stipulation of what a goal is in the first place. It is just a framework for describing an agent. Now, with this framework, humans’ final goals happen to be complex and difficult to discern, and maybe AI goals will be like that too. But we tend to think that AI goals will not be like that. Omohundro argues some economic reasons in his paper on the “basic AI drives”, but also, it just seems clear that you can program an AI with a particular goal function and that will be all there is to it.
Yes, AI may end up with very different interpretations of its given goal but that seems to be one of the core issues in the value alignment problem that Bostrom is worried about, no?

Zeke_Sherman 14 Dec 2018 14:55 UTC
3 points
0 ∶ 0
on: Critique of Superintelligence Part 5
The Pascal’s Mugging thing has been discussed a lot around here. There isn’t an equivalence between all causes and muggings because the probabilities and outcomes are distinct and still matter. It’s not the case that every religion and every cause and every technology has the same tiny probability of the same large consequences, and you cannot satisfy every one of them because they have major opportunity costs. If you apply EV reasoning to cases like this then you just end up with a strong focus on one or a few of the highest impact issues (like AGI) at heavy short term cost. Unusual, but not a reductio ad absurdum.
There is no philosophical or formal system that properly describes human beliefs because human beliefs are messy, fuzzy neurophysiological phenomena. But we may choose to have a rational system for modeling our beliefs more consistently, and if we do then we may as well go with something that doesn’t give us obviously wrong implications in dutch book cases, because a belief system that has wrong implications does not fit our picture of ‘rational’ (whether we encounter those cases or not).

Zeke_Sherman 16 Nov 2018 10:38 UTC
4 points
0 ∶ 0
on: Tiny Probabilities of Vast Utilities: Concluding Arguments
I think the same sheltering happens if you talk about ignoring small probabilities, even if the probability of the x-risk is in fact extremely small.
The probability that $3000 to AMF saves a life is significant. But the probability that it saves the life of any one particular individual is extremely low. We can divide up the possibility space any number of ways. To me it seems like this is a pretty damning problem for the idea of ignoring small probabilities.
We can say that the outcome of the AMF donation has lower variance than the outcome of an x-risk donation, assuming equal EV. So we could talk about preferring low variance, or being averse to having no impact. But I don’t know if that will seem as intuitively reasonable when we circle our new framework back to more everyday, tangible thought experiments.

Zeke_Sherman 15 Nov 2018 10:24 UTC
1 point
0 ∶ 0
in reply to: martlau’s comment on: Crohn’s disease
Only if this project is assumed to be the best available use of funds. Other things may be better.

Zeke_Sherman 14 Nov 2018 22:28 UTC
1 point
0 ∶ 0
in reply to: martlau’s comment on: Crohn’s disease
>Zeke estimates the direct financial upside of a successful replication to be about 33B$/year. This is a 66000:1 ratio (33B/500K = 66000).
This is not directly relevant, because the money is being saved by other people and governments, who are not normally using their money very well. EAs’ money is much more valuable as it is used much more efficiently than Western people and governments usually do. NB: this is also the reason why EA should generally be considered funders of last resort.
If the study has a 0.5% (??? I have no idea) chance of leading to global approval and effective treatment then it’s 35k QALY in expectation per my estimate which means a point estimate of $14/QALY. iirc, that’s comparable to global poverty interventions but at a much lower robustness of evidence, some other top EA efforts with a similar degree of robustness will presumably have a much higher EV. Of course the other diseases you can work on may be much worse causes.
Also that $33B comes from a study on the impact of the disease. Just because you replicate well doesn’t mean the treatment truly works, and is approved globally, etc. Hence the 0.5% number being very low.

Zeke_Sherman 14 Nov 2018 22:14 UTC
2 points
0 ∶ 0
on: Tiny Probabilities of Vast Utilities: Solutions
Last thread you said the problem with the funnel is that it makes the decision arbitrarily dependent upon how far you go. But to stop evaluating possibilities violates the regularity assumption. It seems like you are giving an argument against people who follow solution 1 and reject regularity; it’s those people whose decisions depend hugely and arbitrarily on where they define the threshold, especially when a hard limit for p is selected. Meanwhile, the standard view in the premises here has no cutoff.
> One needs a very extreme probability function in order to make this work; the probabilities have to diminish very fast to avoid being outpaced by the utilities.
I’m no sure what you mean by ‘very fast’. The implausibility of such a probability function is an intuition that I don’t share. I think the appendix 8.1 is really going to be the core argument at stake.
Solution #6 seems like an argument about the probability function, not an argument about the decision rule.

Zeke_Sherman 14 Nov 2018 10:02 UTC
3 points
0 ∶ 0
in reply to: RyanCarey’s comment on: Crohn’s disease
I know. It’s ten years of savings, because curing is accelerated by ten years.

Zeke_Sherman 14 Nov 2018 0:05 UTC
1 point
0 ∶ 0
in reply to: martlau’s comment on: Crohn’s disease
Going from moderate disease to remission seems to be an increase of about 0.25 QALY/year (https://academic.oup.com/ecco-jcc/article-pdf/9/12/1138/984265/jjv167.pdf). If this research accelerates treatment for sufferers by an average of 10 years then that’s an impact of 5 million QALY.
Crohn’s also costs $33B per year in the US + major European countries (https://www.ncbi.nlm.nih.gov/pubmed/25258034). If we convert that at a typical Western cost-per-statistical-life-saved of $7M, and the average life saved is +25 QALY, that’s another 1.2 million. Maybe 2 million worldwide because Crohn’s is mostly a Western phenomenon (https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(17)32448-0/fulltext?code=lancet-site). So that’s 7 million QALY overall. Which of course we discount by whatever the probability of failure is.
It’s very rough but it’s a step forward, don’t let the perfect be the enemy of the good.

Zeke_Sherman 13 Nov 2018 19:24 UTC
2 points
0 ∶ 0
on: Crohn’s disease
We need to factor in QALY or WALY benefits of health improvement in addition to the money saved by users, but we also need to discount for how many people won’t get the new treatment.

Zeke_Sherman 13 Nov 2018 14:31 UTC
1 point
0 ∶ 0
in reply to: kokotajlod’s comment on: Tiny Probabilities of Vast Utilities: Defusing the Initial Worry and Steelmanning the Problem
>The shape of your action profiles depends on your probability function
Are you saying that there is no expected utility just because people have different expectations?
>and your utility function
Well, of course. That doesn’t mean there is no expected utility! It’s just different for different agents.
>I’m arguing that even if you ignore infinitely valuable outcomes, there’s still a big problem having to do with infinitely many possible finite outcomes,
That in itself is not a problem, imagine a uniform distribution from 0 to 1.
>if the profiles are funnel-shaped then what you end up doing will be highly arbitrary, determined mostly by whatever is happening at the place where you happened to draw the cutoff.
If you do something arbitrary like drawing a cutoff, then of course how you do it will have arbitrary results. I think the lesson here is not to draw cutoffs in the first place.
>That’s what I’d like to think, and that’s what I do think. But this argument challenges that; this argument says that the low-hanging fruit metaphor is inappropriate here: there is no lowest-hanging fruit or anything close; there is an infinite series of fruit hanging lower and lower, such that for any fruit you pick, if only you had thought about it a little longer you would have found an even lower-hanging fruit that would have been so much easier to pick that it would easily justify the cost in extra thinking time needed to identify it… moreover, you never really “pick” these fruit, in that the fruit are gambles, not outcomes; they aren’t actually what you want, they are just tickets that have some chance of getting what you want. And the lower the fruit, the lower the chance...
There must be a lowest hanging fruit out of any finite set of possible actions, as long as “better intervention than” follows basic decision theoretic properties which come automatically if they have expected utility values.
Also, remember the conservation of expected evidence. When we think about the long run effects of a given intervention, we are updating our prior to go either up or down, not predictably making it seem more attractive.

Zeke_Sherman 11 Nov 2018 16:17 UTC
2 points
0 ∶ 0
in reply to: kokotajlod’s comment on: Tiny Probabilities of Vast Utilities: Defusing the Initial Worry and Steelmanning the Problem
>Imagine you are adding new possible outcomes to consideration, one by one. Most of the outcomes you add won’t change the EV much. But occasionally you’ll hit one that makes everything that came before look like a rounding error, and it might flip the sign of the EV.
But the probability of those rare things will be super low. It’s not obvious that they’ll change the EV as much as nearer term impacts.
This would benefit from an exercise in modeling the utilities and probabilities of a certain intervention to see what the distribution actually looks like. So far no one has bothered (or needed, perhaps) to actually enumerate the 2nd, 3rd, etc… order effects and estimate their probabilities. All this theorizing might be unnecessary if our actual expectations follow a different pattern.
>So this is a problem in theory—it means we are approximating an ideal which is both stupid and incoherent.
Are we? Expected utility is still a thing. Some actions have greater expected utility than others even if the probability distribution has huge mass across both positive and negative possibilities. If infinite utility is a problem then it’s already a problem regardless of any funnel or oscillating type distribution of outcomes.
>Arguably this behavior is the predictable result of considering more and more possibilities in your EV calculations, and it doesn’t represent progress in any meaningful sense—it just means that EAs have gone farther down the funnel-shaped rabbithole than everybody else.
Another way of describing this phenomenon is that we are simply seizing the low hanging fruit, and hard intellectual progress isn’t even needed.

Zeke_Sherman 10 Nov 2018 16:05 UTC
4 points
0 ∶ 0
on: Tiny Probabilities of Vast Utilities: Defusing the Initial Worry and Steelmanning the Problem
Good post but we shouldn’t assume the “funnel” distribution to be symmetric about the line of 0 utility. We can expect that unlikely outcomes are good in expectation just as we expect that likely outcomes are good in expectation. Your last two images show actions which have an immediate expected utility of 0. But if we are talking about an action with generally good effects, we can expect the funnel (or bullet) to start at a positive number. We also might expect it to follow an upward-sloping line, rather than equally diverging to positive and negative outcomes. In other words, bed nets are more likely to please interdimensional travelers than they are to displease them, and so on.
Also, the distribution of outcomes at any level of probability should follow a roughly Gaussian distribution. Most bizarre, contorted possibilities lead to outcomes that are neither unusually good nor unusually bad. This means it’s not clear that the utility is undefined; as you keep you looking to sets of unlikelier outcomes you are getting a series of tightly finite expectations rather than big broad ones that might easily turn out to be hugely positive or negative based on minor factors. Your images of the funnel and bullet should show much more density along the middle, with less density at the top and bottom. We still get an infinite series, so there is that philosophical problem for people who want a rigorous idea of utilitarianism, but it’s not a big problem for practical decision making because it’s easy to talk about some interventions being better than others.

Zeke_Sherman 28 Nov 2017 6:29 UTC
4 points
0 ∶ 0
in reply to: Liam_Donovan’s comment on: An Exploration of Sexual Violence Reduction for Effective Altruism Potential

Considering that most people would be unhappy to be told that they’re more likely to be a rapist because of their race, we should have a strong prior that many Effective Altruists would feel the same way.

Well I saw statistics that suggest that I’m more likely to be a rapist since I’m a man, the post explicitly said that I have a 6% chance of being a rapist as a man in EA, and that didn’t make me unhappy. And I haven’t seen anyone who has actually expressed any personal discomfort at the OP nor any of my posts, leaving aside the secondhand outrage expressed by characters such as yourself. So my prior is that this is false.

Apart from Lila’s argument, this “non-white people are more likely to be rapists” is a terrible line of thinking because (IMO) it’s likely to build racist modes of thought: assigning negative characteristics to minorities based on dubious evidence

Well you can actually ask rape victims what race their attacker was and then see what the statistics are, as RAINN did in the link I provided. That’s not dubious evidence.

If the evidence were incontrovertible, this might be acceptable, but it’s nowhere near the required standard of proof to overcome the strong prior that humans are equally likely to commit crimes regardless of race

Huh? Why on earth would you have that prior, given the long long history of different ethnic groups behaving differently and being treated differently throughout Western history? And we have damningly strong evidence that people of difference races commit crimes at different rates, as a pure statistical fact backed up by mountains of data gathered by the Bureau of Justice Statistics and many other institutions. What you want to attribute that to is up to you, but refusing to acknowledge it is the height of denialism which even race and progressive activists don’t do. The actually well-informed progressive/leftist activists and philosophers don’t grasp at concepts of rationality to try to throw together some skepticism about whether different races commit crimes at different rates, they just say that the cause of these differential rates is social and a result of structurally racist society, but if you gave a whiff of charity to my posts then you would know that nothing that I have said in any way assumes that the increased propensity of blacks to commit rapes relative to whites in the Western world is not a direct result of structurally racist society.

among other reasons, because race is largely a social construct).

This is a silly cop-out. Only uninformed right wing pundits who strawman poststructuralism think that for something to be a social construct implies that it doesn’t matter and isn’t real. We can define race as minimally as “the color that people check on surveys when they are asked about their race” and it is still true that people commit crimes at different rates in correspondence with how they identify. Dodging these issues by disputing how much biological reality there is or isn’t associated with social race constructs blindly sweeps over enormous realities about how race, ethnicity and skin color are perceived and operate in contemporary society.

Additionally, the long history of using false statistics and “science” to bolster white supremacy should make one more skeptical of numbers like this

Those dastardly white supremacists at the Rape, Abuse & Incest National Network! Or is it the rape victims—you think we shouldn’t believe the rape victims when they tell us what race their aggressor was, is that right? I’m almost offended by that. And yet you try to lecture me about things that will “strengthen bad cognitive patterns and weaken good judgement”...

Zeke_Sherman 22 Nov 2017 1:20 UTC
−1 points
0 ∶ 0
in reply to: Lila’s comment on: An Exploration of Sexual Violence Reduction for Effective Altruism Potential
It’s nice to imagine things. But I’ll wait for actual EAs to tell me about what does or doesn’t upset them before drawing conclusions about what they think.

Zeke_Sherman 21 Nov 2017 22:13 UTC
−1 points
0 ∶ 0
in reply to: Lila’s comment on: An Exploration of Sexual Violence Reduction for Effective Altruism Potential
I think it’s pretty odd of you to try to tell me about what upsets EAs or how we feel, given that you have already left the movement. To speak as if you have some kind of personal stake or connection to this matter is rather dishonest.

I hope you’re just using this as a demonstration and not seriously suggesting that we start racially profiling people in EA.

Racial profiling is something that is conducted by law enforcement and criminal investigation, and EA does neither of those things. I would be much more bothered if EA started trying to hunt for criminals within its ranks than I would be from the mere fact that the manner in which we did this involved racial profiling.

It should be clear why people find the following statements upsetting:

Neither of those statements are upsetting to me.

Zeke_Sherman 13 Nov 2017 12:49 UTC
3 points
0 ∶ 0
in reply to: Kathy_Forth’s comment on: An Exploration of Sexual Violence Reduction for Effective Altruism Potential
There are an estimated 276,000 annual cases of female suicide in the entire world (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3367275/). If, say, half of them are associated with sexual violence (guess), and you throw males in as well, then the eventual lifesaving potential is maybe 150,000 people per year.

Most of these suicides are in SE Asia and the Western Pacific where I believe healthcare and medication provision are not as comprehensive as they are here in the west.

Zeke_Sherman 13 Nov 2017 2:33 UTC
3 points
0 ∶ 0
in reply to: Kathy_Forth’s comment on: An Exploration of Sexual Violence Reduction for Effective Altruism Potential

How long do you think it would take you to upgrade every single estimate to the maximum quality level?

Um I don’t know, I just said I would estimate this one number. I think I was clear that I was talking about “this particular question”.

Assuming 2,300 people in EA per the survey, for every 100 rape victims:

Out of the 25 rape victims who are spouses or partners of the perpetrator (https://www.rainn.org/statistics/perpetrators-sexual-violence), 20 will be outside of EA, when the offender is in EA.

Out of the 45 rape victims who are acquaintances of the perpetrator, 30 will be outside of EA, when the offender is in EA.

Out of the 28 rape victims who are strangers to the perpetrator, 20 will be outside of EA, when the offender is in EA.

Out of the 6 victims who can’t remember or are victimized by multiple people, 4 will be outside of EA, when the offender is in EA.

For the 1 victim who is a non-spouse relative, the victim will be outside of EA.

This makes a total of 30% of rape victims of EAs being in EA.

Assuming 13,000 people in EA per the FB group, for every 100 rape victims:

Out of the 25 rape victims who are spouses or partners of the perpetrator (https://www.rainn.org/statistics/perpetrators-sexual-violence), 23 will be outside of EA, when the offender is in EA.

Out of the 45 rape victims who are acquaintances of the perpetrator, 40 will be outside of EA, when the offender is in EA.

Out of the 28 rape victims who are strangers to the perpetrator, 24 will be outside of EA, when the offender is in EA.

Out of the 6 victims who can’t remember or are victimized by multiple people, 5 will be outside of EA, when the offender is in EA.

For the 1 victim who is a non-spouse relative, the victim will be outside of EA.

This makes a total of 12% of rape victims of EAs being in EA.

Zeke_Sherman 13 Nov 2017 2:18 UTC
0 points
0 ∶ 0
in reply to: Kathy_Forth’s comment on: An Exploration of Sexual Violence Reduction for Effective Altruism Potential
I think you’d get better results if you spent your time simply including things that can easily be included, rather than sparking meta-level arguments about which things are or aren’t worth including. You could have accepted the race correlations and then found one or two countervailing considerations to counter the alleged bias for a more comprehensive overall view. That still would have been more productive than this.

Zeke_Sherman 12 Nov 2017 13:41 UTC
2 points
0 ∶ 0
in reply to: Kathy_Forth’s comment on: An Exploration of Sexual Violence Reduction for Effective Altruism Potential

The way in which gender is relevant while race is not is that sexual attractions are limited by gender preferences in most humans.

Sexual violence tendencies are correlated with racial status in most humans. Why treat it differently?

Given that most sexually violent people attack one gender but not the other, and given that our gender ratio is very seriously skewed, gender is a critical component of this sexual violence risk estimate.

And given that sexually violent people are disproportionately represented across racial categories, and given that our race ratio is very seriously skewed, race is a critical component of this sexual violence risk estimate.

Given that you believe a race adjustment should go with gender adjustment, I don’t see why you are not also advocating for all of the following:

Try and find some statistics for both EAs and sex offenders with comparable data categories on those topics and you’ll see.