Comment by petermccluskey on How Can Donors Incentivize Good Predictions on Important but Unpopular Topics? · 2019-02-06T17:57:10.039Z · score: 1 (1 votes) · EA · GW

Regulations shouldn't be much of a problem for subsidized prediction markets. The regulations are designed to protect people from losing their investments. You can avoid that by not taking investments - i.e. give every trader a free account. Just make sure any one trader can't create many accounts.

Alas, it's quite hard to predict how much it will cost to generate good predictions, regardless of what approach you take.

Comment by petermccluskey on Disentangling arguments for the importance of AI safety · 2019-01-24T05:58:34.145Z · score: 3 (3 votes) · EA · GW

Drexler would disagree with some of Richard's phrasing, but he seems to agree that most (possibly all) of (somewhat modified versions of) those 6 reasons should cause us to be somewhat worried. In particular, he's pretty clear that powerful utility maximisers are possible and would be dangerous.

Comment by petermccluskey on Pursuing infinite positive utility at any cost · 2018-12-12T02:00:06.230Z · score: 6 (3 votes) · EA · GW

I think it's more appropriate to use Bostrom's Moral Parliament to deal with conflicting moral theories.

Your approach might be right if the theories you're comparing used the same concept of utility, and merely disagreed about what people would experience.

But I expect that the concept of utility which best matches human interests will say that "infinite utility" doesn't make sense. Therefore I treat the word utility as referring to different phenomena in different theories, and I object to combining them as if they were the same.


Similarly, I use a dealist approach to morality. If you show me an argument that there's an objective morality which requires me to increase the probability of infinite utility, I'll still ask what would motivate me to obey that morality, and I expect any resolution of that will involve something more like Bostrom's parliament than like your approach.

Comment by petermccluskey on Pursuing infinite positive utility at any cost · 2018-11-15T00:26:03.939Z · score: 6 (3 votes) · EA · GW

>For all actions have a non-zero chance of resulting in infinite positive utility.

Human utility functions seem clearly inconsistent with infinite utility. See Alex Mennen's Against the Linear Utility Hypothesis and the Leverage Penalty for arguments.

I don't identify 100% with future versions of myself, and I'm somewhat selfish, so I discount experiences that will happen in the distant future. I don't expect any set of possible experiences to add up to something I'd evaluate as infinite utility.

Comment by petermccluskey on Thoughts on short timelines · 2018-10-24T18:06:59.250Z · score: 4 (4 votes) · EA · GW

I disagree with your analysis of "are we that ignorant?".

For things like nuclear war or financial meltdown, we've got lots of relevant data, and not too much reason to expect new risks. For advanced nanotechnology, I think we are ignorant enough that a 10% chance sounds right (I'm guessing it will take something like $1 billion in focused funding).

With AGI, ML researchers can be influenced to change their forecast by 75 years by subtle changes in how the question is worded. That suggests unusual uncertainty.

We can see from Moore's law and from ML progress that we're on track for something at least as unusual as the industrial revolution.

The stock and bond markets do provide some evidence of predictability, but I'm unsure how good they are at evaluating events that happen much less than once per century.

Comment by petermccluskey on A model of the Machine Intelligence Research Institute - Oxford Prioritisation Project · 2018-09-24T15:30:40.619Z · score: 1 (1 votes) · EA · GW

I'm a little unclear on what you are asking.

How strictly do you mean when you say "provably safe"? That seems like an area where all AI safety researchers are hesitant to say how high they're aiming.

And by "have it implemented", do you mean fully develop it own their own, or do you include scenarios where they convey keys insights to Google, and thereby cause Google to do something safer?

Comment by petermccluskey on Open Thread #40 · 2018-07-17T15:13:02.230Z · score: 3 (3 votes) · EA · GW

I don't trust the author (Lomborg), based on the exaggerations I found in his book Cool It.

I reviewed that book here.

Comment by petermccluskey on Open Thread #39 · 2018-05-30T01:09:45.099Z · score: 5 (1 votes) · EA · GW

I suggest starting with MAPS.

Comment by petermccluskey on Against prediction markets · 2018-05-13T16:36:47.071Z · score: 1 (1 votes) · EA · GW

I think markets that have at least 20 people trading on any given question will on average be at least as good as any alternative.

Your comments about superforecasters suggest that you think what matters is hiring the right people. What I think matters is the incentives the people are given. Most organizations produce bad forecasts because they have goals which distract people from the truth. The biggest gains from prediction markets are due to replacing bad incentives with incentives that are closely connected with accurate predictions.

There are multiple ways to produce good incentives, and for internal office predictions, there's usually something simpler than prediction markets that works well enough.

Comment by petermccluskey on A case for developing Aldehyde Stabilized Cryopreservation into a medical procedure (1/4) · 2018-05-12T19:37:02.910Z · score: 1 (1 votes) · EA · GW

I object to the idea that early stage Alzheimer's is incurable. See the book The End of Alzheimer's.

Comment by petermccluskey on Against prediction markets · 2018-05-12T16:59:59.436Z · score: 7 (7 votes) · EA · GW

Who are you arguing against? The three links in your first paragraph go to articles that don't clearly disagree with you.

I’d also be curious about a prediction market in which only superforecasters trade.

I'd guess that there would be fewer trades than otherwise, and this would often offset any benefits that come from the high quality of the participants.

Comment by petermccluskey on [Draft] Fighting Aging as an Effective Altruism Cause · 2018-04-18T16:54:40.836Z · score: 2 (2 votes) · EA · GW

It is even known to extend the life of diabetics so they live longer than healthy people.

No, it is known to correlate with living longer. But some or all of that correlation seems to be due to the sickest diabetics being switched from metformin to other drugs.

Comment by petermccluskey on On funding medical research · 2018-02-23T23:04:43.098Z · score: 1 (1 votes) · EA · GW

It does seem like there are important areas where medical research is inadequate. I'll suggest that part of the problem is inadequate effort devoted to treatments that aren't protected by patents.

It looks like some unknown fraction of ME/CFS is caused by low thyroid hormone levels. "Subclinical" hypothyroidism has symptoms that are pretty similar to those of ME/CFS. They are usually distinguished by TSH tests. [TSH is the standard measure of thyroid levels; there are a number of other options, none of which are ideal].

Here's speculation that we should distrust TSH results. (There's a more detailed and very verbose version of that speculation here).

There's plenty of confusion about when it's wise to increase a patient's thyroid hormone. E.g. this small RCT study which gave a standard T4 dose, rather than adjusting the dose to achieve some measure of optimal hormone levels. The reported TSH levels of 0.66 in patients receiving T4 suggest that many patients got more than the optimal dose, and/or didn't convert T4 to T3 well.

In contract, two smaller uncontrolled studies (here00014-0/abstract) and here) reported good results from T3 treatment for treatment-resistant depression (H/T Sarah Constantin). Plus there are lots of anecdotal reports of benefits (see mine here).

There are real dangers from overdoses, and it's unclear how well researchers have measured the benefits, so it's easy to imagine that most doctors are erring on the side of inaction.

My intuition says that there's plenty of room for making protocols that more safely determine the optimal dose. I don't have enough expertise to estimate how tractable that is.

Another area where EAs might possibly provide an important benefit is Alzheimer's. There have been some recent claims that there are strategies which substantially prevent Alzheimer's or reverse it in early stages. As far as I can tell, these claims aren't prompting as much research as they deserve.

Some parts of those strategies are backed by small RCTs published in 2013 and 2012, and yet the first Google search result for Alzheimer's is still a page that says Alzheimer's "cannot be prevented, cured or even slowed".

I expect good research about Alzheimer's to be too expensive for EAs to fund directly, but it seem like we should be able to do something to nudge existing research funding into better directions.

Comment by petermccluskey on A model of the Machine Intelligence Research Institute - Oxford Prioritisation Project · 2017-05-23T19:11:03.918Z · score: 3 (3 votes) · EA · GW

colonisation of the Supercluster could have a very low probability.

What do you mean by very low probability? If you mean a one in a million chance, that's not improbable enough to answer Bostrom. If you mean something that would actually answer Bostrom, then please respond to the SlateStarCodex post Stop adding zeroes.

I think Bostrom is on the right track, and that any analysis which follows your approach should use at least a 0.1% chance of more than 10^50 human life-years.

Comment by petermccluskey on A model of the Machine Intelligence Research Institute - Oxford Prioritisation Project · 2017-05-22T14:41:16.122Z · score: 5 (5 votes) · EA · GW

Can you explain your expected far future population size? It looks like your upper bound is something like 10 orders of magnitude lower than Bostrom's most conservative estimates.

That disagreement makes all the other uncertainty look extremely trivial in comparison.

Comment by petermccluskey on Rational Politics Project · 2017-01-08T19:32:58.225Z · score: 6 (6 votes) · EA · GW

You claim this is non-partisan, yet you make highly partisan claims, such as "conservatives have relied much more on lies" (you cite Trump's lies, but treating Trump as a conservative is objectionable to many conservatives).

Comment by petermccluskey on Voter Registration As an EA Group Meetup Activity · 2016-09-18T17:06:34.412Z · score: 3 (3 votes) · EA · GW

Measurability doesn't sound quite adequate to describe what this proposal is missing.

FHI and MIRI have major problems with measurability, yet have somewhat plausible claims to fit EA principles.

Voter registration has similar problems with estimating how it affects goals such as lives saved, but seems to be missing an analysis of why the expected number of lives saved is positive or negative.

Comment by petermccluskey on Voter Registration As an EA Group Meetup Activity · 2016-09-17T16:38:08.154Z · score: 4 (4 votes) · EA · GW

The obvious objection is that voters who would otherwise not vote are likely to be less informed than the average voter, so your effort causes election results to be less well informed.

You sound more concerned with whether your actions are socially approved than you are with evaluating the results.

Comment by petermccluskey on Political initiative: Fundamental rights for primates · 2016-08-12T23:31:42.329Z · score: -1 (1 votes) · EA · GW

I'll guess that the most important effects of this would be to influence which species get uploaded when, reducing the chances that the world will be ruled by uploaded bonobos, and increasing the chance of nonprimates ruling.

Comment by petermccluskey on Why don't many effective altruists work on natural resource scarcity? · 2016-02-22T17:11:16.871Z · score: 1 (1 votes) · EA · GW

On the Nymex, they currently go out to Dec 2024. That contract appears to trade less than once a week.

There might be occasional contracts for more distant years traded between institutional investors that don't get publicly reported, but the low volume on publicly traded contracts suggests people just aren't interested in trading such contracts.

Comment by petermccluskey on Investment opportunity for the risk neutral · 2016-01-25T20:06:32.111Z · score: 5 (5 votes) · EA · GW

Your use of the phrase "fair market value" is a large red flag.

I've been speculating in stocks for 35 years. One of the hardest lessons I needed to learn was to not believe that last year's prices were fairer than today's prices.

Betting on mean reversion occasionally makes sense, but I've learned to only do it after careful analysis of the fundamentals (earnings, book value, etc).

Comment by petermccluskey on Direct Funding Between EAs - Moral Economics · 2015-08-01T15:42:26.807Z · score: 0 (0 votes) · EA · GW

The goal of avoiding groupthink has the potential to be a very important reason for preferring direct funding. If the direct funding ends up substituting for donations to large, entrenched institutions, then I expect it to be valuable. But I expect that any groupthink associated with young charities that have a handful of employees comes from a broader community, not the specific institution.

Comment by petermccluskey on Direct Funding Between EAs - Moral Economics · 2015-08-01T15:20:16.503Z · score: 0 (0 votes) · EA · GW

One difference in cost comes from institutions such as Oxford requiring their employees to get prestigious wages. That makes the $75k average misleading. More obscure charities can hire employees much more cheaply.

Comment by petermccluskey on Veg*n recidivism seems important, tractable, and neglected · 2015-07-08T22:29:15.630Z · score: 0 (0 votes) · EA · GW

An average of seven years is consistent with the hypothesis that the problem is vitamin B12 deficiency. Our bodies store enough B12 that it takes anywhere from months to decades for symptoms of a severely deficient diet to become clear.

Comment by petermccluskey on When should an Effective Altruist be vegetarian? · 2014-11-27T02:16:46.818Z · score: 1 (1 votes) · EA · GW

Eggs from pasture-raised chickens are not very hard to find in Berkeley. Their nutritional advantage over grain-fed eggs was enough for me to switch to them. Yes, they cost $8 to $10 per dozen.

Comment by petermccluskey on How a lazy eater went vegan · 2014-10-08T22:39:23.692Z · score: 0 (0 votes) · EA · GW

Ambronite sounds healthier than any other vegan diet, although it's unclear whether the B12 in it is bioavailable.

I don't trust nutrition science enough to trust a fully vegan diet, but oysters and/or insects would add enough to make the diet seem safe enough for me.