“It doesn’t feel planned.”
Alright, exercise time: rather than picking out the individual things wrong with this scenario, can you describe (at a relatively high level) what your plan would be, if you were to plan this? From scratch, with today’s technology, what would be the optimal way to run this sort of study?
Many of the pain points listed have a common trait: the decision would seem easier with less information. For example, the PhD decision is easier if you didn’t know about the costs which have been sunk, the identity decisions are easier if you’re not sure of your own identity, cached thought problems are easier without having that thought cached, etc…
But we know that information should never have negative value. So why not highlight that dissonance? Imagine the following exercise:
Handout: “You spent the last 3 years working toward a PhD. You passed up a $90k job to stay in the program. Now you have 2 years left, and another $90k job offer has come your way. Do you take it?” (I don’t know much about PhD programs, so feel free to imagine more plausible numbers here and add narrative).
Exercise 1: Is there any information you would prefer to not know? Exercise 2: How much would you pay to not know it?
If you really want to have fun, give people monopoly money and let them bid to remove information from a range of scenarios. Note that we’re not offering to change the facts, just to not know them.
Personally, I think this would be a lot easier if I could just forget about all that time spent in the PhD program.
At least in this case, the exercise highlights the difference between consequentialist and non-consequentialist reasons/excuses for doing things. The “how much would you pay to not know it” is especially handy, since it puts a number on that mental pain. Then we can ask whether the mental pain is worth all the money you’d lose by, in this example, staying in the PhD program.
Thanks, Joachim. Like I said, I don’t know much about PhD programs. What would be some better numbers to make the point?
Charades! If Paul Graham’s entrepreneur had to explain the website in charades, I bet he’d be acting out a use case (at least, that’s the first thing I’d think of). Outlaw all that “first word, 2 syllables...” junk and have people play charades. It’s really hard to be abstract without words.
There’s also the problem of an infinite number of possible solutions.
I would like feedback on a theory.
First, some groundwork. Voldemort’s first “destruction” as a result of Lily Potter’s sacrifice does not seem right. Even if we accept that love and sacrifice somehow block the killing curse, she was trying to kill Voldemort when he killed her. That doesn’t add up to burnt-out husk of Voldemort. Now, before I suggest a solution, you might want to consider alternatives for yourself.
I suggest that Voldemort was intentionally turning Harry into a Horcrux. There are at least two reasons for the burnt-out husk result. (1) Using a human as horcrux was probably experimental, and prone to disaster. (2) Voldemort’s original body served as the human sacrifice to create the horcrux. There are many other possibilities, and the details are not critical. What matters is that Harry, as a horcrux, is at least partly a copy of Voldemort himself (thus the mysterious dark side). Interestingly, this means a copy of Voldemort grew up with a loving and supportive family.
Now for the important part: what the heck is Professor Quirrel up to? Acquiring political power for Harry seems to be high on his agenda. If Harry contains a copy of Voldemort, then acquiring political power for Harry is, in some sense, acquiring power for Voldemort.
Summary: Harry is (partly) Voldemort. Voldemort/Quirrel’s plan (or at least one of them) was to make a copy, put it in Harry, and make Harry the new Lord.
The events of the recent chapters have provided significant evidence against this theory. Taking out Harry’s best friends is not consistent with putting Harry in a position of political power, assuming that it was Voldemort/Quirrel behind the incident. But that’s a problem regardless of theory. Professor Quirrel’s obvious agenda is to put Harry in power. If he really wants to destroy Harry, why this agenda? Note that far less would suffice to make Quirrel look innocent in Harry’s destruction. But if he really does want Harry in power, why remove his friends?
If someone else was behind it, who and why? Dumbledore should not be ruled out yet, although he doesn’t have an obvious motive and he’s putting up a damn good show if it was him. Lucius does have an obvious motive, but also put up a damn good show, and it seems overly clever for him. Postulating an unknown third party would take a lot more evidence to justify the increase in complexity and provide little predictive power, though it’s not impossible.
So, three questions: (1) Any other evidence for/against the “Harry is the new Voldemort” theory? (2) If Quirrel was behind the attack on Draco/Hermione, what is his real agenda? (3) If someone else was behind it, who and why?
I don’t think the “spoil the setup” experiment distinguishes TDT from the belief in fairness. Just because the second person’s decision comes after the first doesn’t mean it has no effect on the first. It’s very much like Newcomb’s problem in that regard, and one of the main points of TDT was to account for that effect. Depending on the details of the rewards and how strongly you think the other player’s decisions correlate with your own, it may make sense to precommit to cooperation even if you’re told the other person’s choice. And if it makes sense to precommit to cooperation, that’s what TDT will do (unless I’m missing something).
The superrationality explanation still makes sense. If the other player’s choice is known, then symmetry is broken, so the superrational agent should defect.
Other than that, I’m not really sure what you mean by “explanation”. The “folk version of superrationality” sounds plausible, but the underlying causes of the experimental results still feel pretty mysterious. Demystifying them is well beyond my capability, but it’s certainly an interesting question.
Does anyone know if the probabilities output by Solomonoff Induction have been proven to converge? There could be as many as O(2^n) hypotheses of length n, each of which get a probability proportional to 2^-n. Once you sum that over all n, it doesn’t converge, therefore the probability can’t be normalized unless there’s some other constraint on the number of hypotheses of length n consistent with the data. Does anyone know of such a constraint?
Taking the Solomonoff Induction route, making a highly compressible version of Windows is EXACTLY what it would mean to have “discovered deep insights into how to build powerful operating systems”. Similarly, getting a lot of power out of simple designs is exactly what it means to have insights, at least in the context of Solomonoff Induction (and from there to science in general).
That said, great article. The contribution of complexity from the environment is a major issue, even as early as the womb, and that was definitely an important oversight.
I found the use of multiplication particularly useful, since it forced the reader to pay attention to the physical/logical distinction. If, say, addition had been used, then a determined reader could try to use physical constraints alone (though they would be cheating).
I still feel confused. I definitely see that, when we talk about fairness, our intended meaning is logical in nature. So, if I claim that it is fair for each person to get an equal share of pie, I’m trying to talk about some set of axioms and facts derived from them. Trying.
The problem is, I’m not convinced that the underlying cognitive algorithms are stable enough for those axioms to be useful. Imagine, for example, a two-year-old with the usual attention span. What they consider “good” might vary quite quickly. What I consider “just” probably depends on how recently I ate. Even beyond such simple time dependence, what I consider “just” will definitely depend on context, framing, and how you measure my opinion (just ask a question? Show me a short film? Simulate the experience and see if I intervene?). Part of why friendly AI is so hard is that humans aren’t just complicated, we’re not even consistent. How, then, can we axiomatize a real human’s idea of “justice” in a useful way?
Informally speaking, it seems like superexponential numbers of people shouldn’t be possible. If a person is some particular type of computation, and exactly identical copies of a person should only count once, then number of people is bounded by number of unique computations (exponential). It does not seem like the raw Kolmogorov complexity of the number will be the right complexity penalty if each person has to be a different computation.
tl;dr: Excess body fat and obesity are an immune response to gram-negative gut bacteria, not a metabolic problem. Fix it by taking oral polymyxin, or a comparable antibiotic. Further research into good antibiotics for this purpose would be appreciated.
Earlier this year, an article found that bacteria from an obese human could cause obesity in mice. They isolated the bacteria, put it in some randomly chosen mice, and after a few months the mice with the bacteria were fat and had diabetes problems while the control group was healthy. With a second experiment they found that the mechanism is the molecule lipopolysacharride (LPS aka endotoxin), found in the membrane of all gram-negative bacteria. When gram negative bacteria become established in the gut, the LPS triggers a inflammation response from the immune system which causes both fat accumulation and diabetes in the long run. So they’ve established very firmly that gut bacteria are sufficient to cause excess body fat, but whether that’s the main source in the general human population is unknown. (source: http://www.nature.com/ismej/journal/v7/n4/pdf/ismej2012153a.pdf, apologies if it’s behind an academic firewall)
So how does one get rid of gram-negative bacteria? It turns out that there is a common antiobiotic, polymyxin, which specifically targets LPS itself and kills bacteria which produce LPS. Polymyxin is among the most common topical antibiotics (along with neosporin), and can also be taken intravenously or orally. Intravenously it is a mild neurotoxin, but this is not an issue if taken orally.
Finally, it turns out that a study published back in 2006 administered polymyxin intravenously to rats. They found a 46% drop in adipose fat mass in rats given polymyxin. They had no idea what the mechanism was, hypothesized some vague connection to insulin signalling, and it just went down as one of those weird results. But now, in light of the more recent results, we can be pretty sure that gram-negative gut bacteria were the issue. The importance of this study is that it suggests gram-negative bacteria are a major cause of excess body fat in the general rat population, not just a special case of the 2013 study. So, it’s reasonable to suspect that polymyxin would fix most human obesity too. (source: http://link.springer.com/article/10.1007/s10989-005-9009-9)
Ah, but the cool thing is that you don’t need an experimental treatment. We’re talking bog-standard antibiotics, nothing unusual. The only unusual thing is what you’d be trying to do with those antibiotics.
And unlike the usual vaguely promising substances, the effect sizes were huge in both experiments. You don’t get a 46% drop in adipose fat from random noise.
Thanks for pointing out the journal error, that has been corrected. Also big thanks for the working link.
The “experiment” with the human subject in the ISMEJ article was stupid, which was why I didn’t mention it. Everything I’m saying is based on the mouse experiments.
I do think your interpretation of these experiments is way too restricted. In a frequentist sense, everything you’re saying is reasonable, since we don’t know how well various results generalize (mouse to human, intravenous to oral, etc...). But in a Bayesian sense, this is pretty good evidence. I guessed that polymyxin would reduce body fat (regardless of how it’s administered) just based on reading the ISMEJ article, which never mentioned antibiotics. That means the first article alone was enough to promote the hypothesis out of entropy. I then found the second article by searching for papers discussing polymyxin and obesity, and the result was what I expected (large drop in fat after polymyxin administration), so that’s a big evidential boost in favor.
Neither of those gives evidence for the mouse result generalizing to humans. However, we do know that gram-negative bacteria are pretty ubiquitous and do trigger an immune response in humans similar to that in mice, so based on the physical systems we should expect a similar response to antibiotics.
There was a special issue of Nature focusing on human microbiota a few months back which seems to have a lot more relevant research with humans, but I haven’t had time to go through them in depth yet (which is why this was a poorly-researched comment rather than a full discussion post).
I think it would be helpful to explicitly write out the what you’re calculating mathematically, e.g. \sum{k=1}^{\infty} (\frac{2^k}{k})*P[halting~by~step~k] = \sum{k=1}^{infty} (\frac{2^k}{k})(2^k) = \sum_{k=1}^{infty} \frac{1}{k} just to make it clear that this is the harmonic series. I know it’s described in English, but the math is often helpful. (For posting latex, see also: http://lwlatex.appspot.com/. I don’t know if that’s the most up-to-date method, but it works.)
A few starting points for relevant and rigorous research:
http://www.ted.com/talks/bruce_bueno_de_mesquita_predicts_iran_s_future.html (this guy’s stuff is more targeted at popular audiences, but he’s a good gateway to a directly relevant field)
Anything by Thomas Schelling (great introduction to nonzero sum game theory in general)
http://www.amazon.com/Solving-War-Puzzle-Beyond-Democratic/dp/0890894825 (statisticians have done some decent work on the real causes of war; tangentially relevant)
Who knows what amazing research teams might be borne of the intellectual equivalent of OK!Cupid.
This sounds like an awesome idea. Does anyone know if such a thing already exists? If not, I would be willing to commit $500 to someone who wants to make it (assuming they already have the appropriate skill set, of course).
I have also noticed this ability-to-react for several years. I am definitely good at it. Unfortunately, I cannot offer any advice on developing the skill, but I do have some observations:
-I tend to get nervous after doing something. For instance, when presenting to a large group, I feel completely calm. Afterward, however, I often start shaking as I go over it in my head.
-Regarding public speaking, I don’t generally need to write out what I’m going to say, but I do need to think it out in advance. Usually, I have a 5-10 line summary for a 10 minute presentation. Otherwise, I ramble.
-This skill is especially useful in dancing. I dance blues, salsa, and some swing socially. I almost never plan out moves; dance partners have asked me many times what that cool move was, and I have no idea. Apparently this is very unusual. The flip side is that I cannot dance pre-choreographed moves at all. Usually I forget the sequence within 3 measures and just do something else.
-I also love simulations like those you mentioned. We used to do emergency simulations in the Scouts, and I felt very in-my-element. I just go into this alert problem-solving mode and everything seems to flow smoothly. It feels great.
Two hypotheses:
Is this skill related to left/right brain dominance? I believe my ability-to-act is unusually good, and I know that I am extremely right brained. Intuitively, this ability-to-act sounds like a right-brain skill. Any other observations for/against this hypothesis?
I would bet money under long odds that this skill correlates closely with parallel thinking. Parallel thinking is the ability to consider multiple approaches to a problem simultaneously; I do it constantly. If you’ve ever looked at a problem,immediately seen three ways of doing it, estimated the difficulty of each, and gone with the best, that’s parallel thinking. When I’m in ability-to-act mode, it becomes a subconscious process. Any one else want to submit evidence for or against for this one?