Yes, it won’t work for people who can’t manage a day without eating at least from time to time, although you can also try slowing down the rate of change.
As I said in another comment, changes in water retention (and scale flucuations etc.) don’t really matter because it will come out the same on average.
It doesn’t matter. Fluctuations with scales and with water retention may mean that you may end up fasting an extra day here and there for random reasons, but you will also end up eating on extra days for the same reason. It ends up the same on average.
Technology frequently improves some things while making other things worse. But sooner or later people find a way to improve both the some things and the other things. In this particular case, maybe they haven’t found it yet.
“But actually they weren’t aware that they were not eating less.”
This is why I advocate the method of using a Beeminder weight goal (or some equivalent), weigh yourself every day, and don’t eat for the rest of the day when you are above the center line. When you are below it, you can eat whatever you want for the rest of the day.
This doesn’t take very much willpower because there is a very bright line, you don’t have to carefully control what or how much you eat, it’s either you eat today or you don’t.
Ok, that makes sense.
I don’t think I understand. What is the third possible environment? And why exactly is the behavior stupid? It sounds like it might be actually true that it is too dangerous to test whether you are in Heaven or Hell in that situation.
“Imagine a world in which no one was living below the average income level.”
This is a world where everyone has exactly the same income. I don’t see any special reason why it would be desirable, though.
Probability that there are two elephants given one on the left and one on the right.
In any case, if your language can’t express Fermat’s last theorem then of course you don’t assign a probability of 1 to it, not because you assign it a different probability, but because you don’t assign it a probability at all.
Basically the problem is that a Bayesian should not be able to change its probabilities without new evidence, and if you assign a probability other than 1 to a mathematical truth, you will run into problems when you deduce that it follows of necessity from other things that have a probability of 1.
Any program that reads this post and these articles wasn’t stuck in a sandbox anyway.
I would agree with the social norm of never ever going swimming. In fact, I have a very hard time understanding why people are so willing to basically immerse themselves in an environment so deadly to human beings. I certainly never do it myself.
You are assuming that human beings are much more altruistic than they actually are. If your wife has the chance of leaving you and having a much better life where you will never hear from her again, you will not be sad if she does not take the chance.
If the other player is choosing randomly between two numbers, you will have a 50% chance of guessing his choice correctly with any strategy whatsoever. It doesn’t matter whether your strategy is random or not; you can choose the first number every time and you will still have exactly a 50% chance of getting it.
That is not a useful rebuttal if in fact it is impossible to guarantee that your AGI will not be a socialpath no matter how you program it.
Eliezer’s position generally is that we should make sure everything is set in advance. Jacob_cannell seems to be basically saying that much of an AGI’s behavior is going to be determined by its education, environment, and history, much as is the case with human beings now. If this is the case it is unlikely there is any way to guarantee a good outcome, but there are ways to make that outcome more likely.
If you are “procrastinate-y” you wouldn’t be able to survive this state yourself. Following a set schedule every moment for the rest of your life is very, very difficult and it is unlikely that you would be able to do it, so you would soon be dead yourself in this state.
Ramez Naam discusses it here: http://rameznaam.com/2015/05/12/the-singularity-is-further-than-it-appears/
I find the discussion of corporations as superintelligences somewhat persuasive. I understand why Eliezer and others do not consider them superintelligences, but it seems to me a question of degree; they could become self-improving in more and more respects and at no point would I expect a singularity or a world-takeover.
I also think the argument from diminishing returns is pretty reasonable: http://www.sphere-engineering.com/blog/the-singularity-is-not-coming.html
Human beings are not very willing to be rational, and that includes those of us on Less Wrong.
If you’re really honest about your willingness to be rational, it seems like this could be kind of depressing.
I actually meant it more generally, in the sense of highly unusual situations. So gjm’s suggested path would count.
But more straightforwardly apocalyptic situations could also work. So a whole bunch of people die, then those remaining become concerned about existential risk—given what just happened—and this leads to people becoming convinced Zoltan would be a good idea. This is more likely than a virus that kills non-Zoltan supporters.
Eliezer isn’t arguing with the mathematics of probability theory. He is saying that in the subjective sense, people don’t actually have absolute certainty. This would mean that mathematical probability theory is an imperfect formalization of people’s subjective degrees of belief. It would not necessarily mean that it is impossible in principle to come up with a better formalization.