= 27d567bc2d48d9f40e9ccc20ea370b15)
Anyone care to elaborate on Why a Bayesian is not allowed to look at the residuals?
I got hunches, but don't feel qualified to explain in detail.
In response to
Open Thread for February 3 - 10
A political question:
Our recently elected minister for finance just did something unexpected. She basically went:
“Last autumn during the election campaign, I said we should do X. After four months of looking at the actual numbers, it turns out that X is a terribad idea, so we are going to do NOT X”
(She used more obfuscating terms, she’s a politician after all.)
The evidence points to her actually changing her mind rather than lying during the election.
The question:
Would you prefer a politician sane enough to change her mind when presented with convincing evidence or one that you (mostly) agree with?
This math only works if I value a year of someone else's life approximately the same as a year of my life.
If instead I value a year of someone else's life (on average), say, a tenth as much as I value a year of my own life, then if I use your numbers to compare the EV of cryonics at 4 GDY (giga-Dave-years) to the EV of life-extension research at 1.75 GDY, I conclude that cryonics is a better deal.
Approached the other way... if I don't value any given life significantly more than any other, there's no particular reason for me to sign up for cryonics or research life extension. Sure, currently living people will die, but other people will be alive, so the total number of life-years is more or less the same either way... which is what, in this hypothetical, I actually care about. The important thing in that hypothetical is increasing the carrying capacity of the environment, so the population can be maximized.
It turns out to matter what we value.
Your first point is of course valid. My algorithm for determining value of a life is probably a bit different from yours because I end up with a very different result. I determine the value of a life in the following manner:
Value = Current contribution to making this ball of rock a better place + (Quality of life + Unrealised potential) * Number of remaining years.
If we consider extended life spans, the first element of that equation is dwarfed by the rest so we can consider that to be zero for the purpose of this discussion.
Quality of life involves a lot of parameters, and many are worth improving for a lot of people. Low hanging fruit includes: Water supply and sanitation in low-income countries, local pollution in the same countries, easily treatable diseases, Women's lib. All of these are in my opinion worthy alternatives to cryonics, but maybe not relevant for this particular discussion.
The remaining parameter is Unrealised Potential which I think of as (Intelligence * conscientiousness). I am brighter than most, but more lazy than many, so the result, if interpreted generously, is that I may be worth somewhat more than the median but certainly not by a factor of 10, so if we still go with the numbers above (even if Eliezer pointed out that they were crazy), my stance is still that cryonics is a poor investment. (It may be fun but not necessarily productive to come up with some better numbers.)
Also: I have absolutely no problem accepting that other people have different algorithms and priors for determining value of life, I am just explaining mine.
Your other point was more of a surprise and I have spent a significant amount of time considering it and doing rudimentary research on the subject, because it seems like a very valid point. The main problem is that it does not seem that the total number of high quality life-years is limited by the carrying capacity of the planet, especially if we accept women's lib as a core requirement for attaining high quality.
Declining fertility rate seems to be extremely well correlated with higher quality of life so, once we sort the poverty problem, the planets population will decline. Singapore already has a fertility rate of less than 1 child per woman and chinas population is expected to peak in 2020.
In the short term however, the carrying capacity may very well be a limiting factor and may be worth increasing. Also because a larger carrying capacity will indirectly help with sorting the poverty problem so it's a double win. In fact I am seriously considering moving some of my retirement savings from index funds to aquaculture because that seems to be where the most low-hanging fruit seems to be. Suggestions are welcome.
Again thanks for the pushback, having to actually think through my arguments is a new and welcome experience.
How do you invest $50,000 to get a 25% chance of increasing everyone's lifespan by 10 years? John Schloendorn himself couldn't do that on $50K.
Reviewing the numbers you made up for sanity is an important part of making decisions after making up numbers.
You're right. Those numbers weren't just slightly coloured by hindsight bias but thoroughly coated in several layers of metallic paint and polished. They need to be adjusted drastically down. The reasons I originally considered them to be reasonable are:
The field of cancer research seem to be a lot like software in the 80s in that our technical ability to produce new treatments is increasing faster than the actual number of treatments produced. This means that any money thrown at small groups of people with a garage and a good idea is almost certain to yield good results. (I still think this and am investing accordingly)
I have made one such investment which turned out to be a significant contribution in developing a treatment for prostate cancer that gives most patients about 10 extra years.
There are far too many similarities between cancer cells and ageing cells for me to readily accept that it is a coincidence. This means that investing in cancer research startups has the added bonus of a tiny but non-zero chance of someone solving the entire problem as a side effect.
In retrospect, I also went: "prostate cancer patients == many => everyone == many" (I know, scope insensitivity :( )
On the other hand, my numbers for cryonics were also absurdly optimistic, so I'm not yet convinced that the qualitative point I was trying (ineptly) to make is invalid. The point was: Even a large chance of extending one life by a lot should be outweighed by a smaller chance of extending a lot of lives by a little, especially if the difference in total expected number of years is significant.
Also: Thanks for the pushback. I am far too used to spending time with people who accept whatever I say at face value and the feeling I get on here of being the dumbest person in the room is very welcome.
This math only works if I value a year of someone else's life approximately the same as a year of my life.
If instead I value a year of someone else's life (on average), say, a tenth as much as I value a year of my own life, then if I use your numbers to compare the EV of cryonics at 4 GDY (giga-Dave-years) to the EV of life-extension research at 1.75 GDY, I conclude that cryonics is a better deal.
Approached the other way... if I don't value any given life significantly more than any other, there's no particular reason for me to sign up for cryonics or research life extension. Sure, currently living people will die, but other people will be alive, so the total number of life-years is more or less the same either way... which is what, in this hypothetical, I actually care about. The important thing in that hypothetical is increasing the carrying capacity of the environment, so the population can be maximized.
It turns out to matter what we value.
Actually, that last bit was an entirely new thought to me, thanks
In response to
The Third Alternative
The only lies to children should be Lies to Children. Any other lies, including Santa, creationism or any other fiction presented as facts should be considered child abuse.
(My ex tried to bring up our children as YECs after being 'born again' and our courts ruled this to be child abuse which is why I'm a single dad. I may be a bit more than the average fanatical about this particular point.)
In response to
Why I haven't signed up for cryonics
I have a view on this that I didn't find by quickly skimming the replies here. Apologies if it's been hashed to death elsewhere.
I simply can't get the numbers to add up when it comes to cryonics.
Let's assume a probability of 1 of cryonics working and the resulting expected lifespan to be until the sun goes out. That would equal a net gain of around 4 billion years or so. Now, investing the same amount of money in life extension research and getting, say a 25% chance of gaining a modest increase in lifespan of 10 years for everyone would equal 70bn/4 = 17.5bn expected years total.
If you think (like I do) that even if cryonics end up working, there will probably be a hard limit significantly less than 4bn years(my uneducated guess is that it will be around the lifespan of a ponderosa pine, or 6000 years or so), or that cryonics have a probability of less than one, the figures only get worse.
Of course, there may be a lower chance of extending everyones life by even a little, but on the other hand, my grandfather has already gained a decade partly as a result of my investment in a cancer research startup some years back, so I'm not willing to lower those odds by enough to make cryonics come out on top :)
Cycling does not uniquely reduce risk of coronary heart disease. Even if cycling beats driving because of reduced risk of heart disease, driving + non-dangerous exercise would still beat cycling.
At roughly double the time investment. I prefer to commute by bicycle whenever possible (I live in a city where about 20% of people bike to work during summer and about 5% during the winter, so I suspect risk is lowered by bikes being more common on the road). The commute by bike takes about 80 minutes (including return), sitting in rush-hour traffic takes about the same, as would "non-dangerous" exercise. Discounting the negative effects of commuting by car, I would still be losing about 400 hours per year by "exercising safely".
So in order to make up for the lost time, the increased risk of commuting by bicycle should reduce my life expectancy by roughly 0.4%. It doesn't.
Also, statistics and my personal experience indicates that the most effective way of avoiding traffic accidents is to live in a western country other than the United States
In response to
Tell Your Rationalist Origin Story
I never had a watershed moment when I ‘discovered' rationalism. For those of you who grew up with religion and take faith as a more or less given part of society, I must have had a rather peculiar childhood; When I was little, I spent quite a lot of time with my grandfather who was an uneducated farmer and had never heard of Bayes’ Theorem. (But loved it when I recently explained the basics to him.) I remember starting sentences with "I believe…" and I never got any further before being interrupted with "If you want to believe, you can go to church. Around here we go with the evidence”. Upon being asked about the possibility of an afterlife, he went: “If there is an afterlife, it can not be observed, and if it can not be observed it is certainly not important enough to talk about”. He is a rather abrupt old man. Another nugget of wisdom was “There is an infinite supply of possible mistakes to choose from, so there is no point in making the same one twice.”
I also live in a country where about 80% of the population are Atheists/Agnostics, which makes religion something the crazies are talking about. Empiricism and rationalism has simply been the default position for my entire life, even if not expressed explicitly or in mathematical terms.
Even atheists who are taught from the age of five that ‘belief’ is not a permissible word are not exempt from biases of course, and the first time I consciously noticed that particular problem was right after I died (which could count as a watershed moment.) In my early twenties, I went through a bit of a mental/emotional crisis (for reasons that I may end up covering in another post) that I handled through extreme amounts of exercise and caused me to lose weight. During my yearly checkup at the doctors, he noticed that I was severely underweight. (we’re talking a BMI of about 12 here, so severely is not an understatement) He spent quite some time going over my eating habits and couldn’t find anything wrong. Two months later, I was found dead by the roadside and revived in the ambulance. Random passersby who know CPR is a good thing (TM). My body had simply shut down due to spending more energy than I had absorbed over an extended period of time. After recovering for a bit at the hospital, I slowly realised that the doctor had observed something wrong (a very low BMI) and upon finding that the normal explanation (anorexia) didn’t apply, he had discarded the observation rather than looking for other possible explanations. I was quite upset about this, and I also started examining myself with respect to why I hadn’t noticed that something was wrong either.
This experience made me rather obsessed with cognitive science and nutrition / exercise (which may be worth another post, this community seems interested in the topic); enough to take a year of pre-med, a year of psychology (which didn’t help much) and to read all I could find about cognitive theory (which did, especially Gödel, Escher, Bach which is also a wonderful book when considered only on it’s merits as literature.) This was in 1994, so quite a time before OvercomingBias and LessWrong, or I would most certainly have found this place back then. (btw, we have free tuition here, so I take at least one college course per semester, hence the odd academic choices. This year I’m doing introduction to quantum mechanics which led me to stumble upon the quantum physics sequence which led me to read HPMOR, which made me lurk around this site for the last five months and reading the rest of the sequences before creating an account.) (this comment counts both as a 'hello' post and a 'Origin story' post :)
= b715d02e85f7b3b4ae65ca3d3e450868)
Subscribe to RSS Feed
= f037147d6e6c911a85753b9abdedda8d){kind=logo}
Having something to care for creates motivation. A dog will bug you if you don't care enough for him and if he's stronger willed than you are, he just might get you to take action instead of doing nothing by looking at you with a sad face.
It easy to give your friends an excuse why you want to stay at home. When you however have a dog who looks at you and expects you to care for it the dog won't take any bullshit excuses.
It also an animal. Having a dog die because you don't care enough for it is tragic but less tragic than committing suicide because you took the wrong antidepressant.
If you really fail to develop a relationship with the animal you can also give it away.
Also, pets trigger oxytocin release.
A dog usually also "forces" you to exercise moderately by walking it and exercising while depressed is notoriously difficult to follow through on.
I vaguely remember reading a study that found talking to your dog gave the same result as talking to a therapist, (indicating that talking is more important than having someone listen) but I can't find the link at the moment, maybe I misremember.
Owning a dog is not the only option either, I have fairly regularly lent out some of mine to friends and family going through a rough patch with good results.
I suspect it depends on severity and cause though. If the depression is bad enough you may need chemicals just to get you to a level where you can benefit from other treatment.