Less Wrong is a community blog devoted to refining the art of human rationality. Please visit our About page for more information.

Stop Using LessWrong: A Practical Interpretation of the 2012 Survey Results

-37 aceofspades 30 December 2012 10:00PM

Link to those results: http://lesswrong.com/lw/fp5/2012_survey_results/

I've been basically lurking this site for more than a year now and it's incredible that I have actually taken anything at all on this site seriously, let alone that at least thousands of others have. I have never received evidence that I am less likely to be overconfident about things than people in general or that any other particular person on this site is.

Yet in spite of this apparently 3.7% of people answering the survey have actually signed up for cryonics which is surely greater than the percent of people in the entire world signed up for cryonics. The entire idea seems to be taken especially seriously on this site. Evidently 72.9% of people here are at least considering signing up. I think the chance of cryonics working is trivial, for all practical purposes indistinguishable from zero (the expected value of the benefit is certainly not worth several hundred thousand dollars in future value considerations). Other people here apparently disagree, but if the rest of the world is undervaluing cryonics at the moment then why do those here with privileged information not invest heavily in the formation of new for-profit cryonics organizations, or start them alone, or invest in technology which will soon develop to make the revival of cryonics patients possible? If the rest of the world is underconfident about these ideas, then these investments would surely have an enormous expected rate of return.

There is also a question asking about the relative likelihood of different existential risks, which seems to imply that any of these risks are especially worth considering. This is not really a fault of the survey itself, as I have read significant discussion on this site related to these ideas. In my judgment this reflects a grand level of overconfidence in the probabilities of any of these occurring. How many people responding to this survey have actually made significant personal preparations for survival, like a fallout shelter with food and so on which would actually be useful under most of the different scenarios listed? I generously estimate 5% have made any such preparations.

I also see mentioned in the survey and have read on this site material related to in my view meaningless counterfactuals. The questions on dust specks vs torture and Newcomb's Problem are so unlikely to ever be relevant in reality that I view discussion about them as worthless.

My judgment of this site as of now is that way too much time is spent discussing subjects of such low expected value (usually because of absurdly low expected probability of occurring) for using this site to be worthwhile. In fact I hypothesize that this discussion actually causes overconfidence related to such things happening, and at a minimum I have seen insufficient evidence for the value of using this site to continue doing so.

Comment author: Abd 14 November 2012 03:16:46AM *  2 points [-]

I pointed to sources that contain huge lists of sources, including such studies. Some of what I pointed to is free. There is no need to reproduce this here. The relevance here is to cascades, which occur without "conspiracies."

A common response to a cascade being pointed out is to call the observer a "conspiracy theorist," and that happens even if no conspiracy has been alleged. That people might be unconsciously motivated by issues of reputation and "face" is just what's so for human beings.

I mentioned funding and was explicit that I did not know if this had an actual effect on recommendations.

Taubes has laid out the history of the "official dietary recommendations," and he makes a persuasive case that some serious errors were made, and that some are persisting in beliefs that are not consistent with what is scientifically known.

Anyway, aceofspades asks for studies. He didn't specify the context, but it was that he had written

Would you mind linking to this research that shows low carb diets lower cardiac risk factors?

I linked to extensive coverage of that research, by science journalists. However, specifically, and just what I picked up quickly:

Weight loss with a low-carbohydrate, Mediterranean, or low-fat diet. (Blood lipids, i.e., cardiac risk factors, were studied.)

Comparison of the Atkins, Zone, Ornish, and LEARN diets for change in weight and related risk factors among overweight premenopausal women: the A TO Z Weight Loss Study: a randomized trial. (Lipid profile was studied.)

Systematic review of randomized controlled trials of low-carbohydrate vs. low-fat/low-calorie diets in the management of obesity and its comorbidities. (This is a "systematic review," very much on point as to cardiac risk.)

Part of my own experience:

I was under forty when my doctor, whom I trusted greatly, recommended that I go on a low-fat diet because I had mildly elevated cholesterol. Over 20 years later, the results: I'd gained about 30 lbs, my cholesterol levels were a lot higher. Sure, I wasn't terribly compliant, but I'd shifted the balance greatly toward low-fat. Turns out my experience was typical. Compliance with low-fat diets is commonly poor, and the effect of the recommendation is often weight gain and worsening lipids. So then statins are prescribed....

My new doctor suggested the South Beach diet (kind of a compromised lower-fat or lower-sat-fat Atkins diet, also by a famous cardiologist), but I did the research this time, and found that the science was stronger behind Atkins. I told him, and he led me into his office and handed me the standard textbook on diabetes, written in the 1920s, that described what was then the standard treatment for type II diabetes. A low-carb diet. Insulin had just been discovered, and insulin was considered a miracle drug for the rest, who didn't respond to low-carb diets. Fast forward, the American Diabetes Association discourages low-carb diets. Why? It's really a good question!

Well, why hadn't this doctor told me straight out about low-carb, that my high cholesterol was not necessarily a problem? It's a little thing called "standard of practice." He could lose his job and/or his license. However, he could smile at me and tell me "whatever you are doing, keep it up." (Because my lipids and other indicators of heart health improved greatly.)

And then I found from a biopsy that I have prostate cancer. Taubes describes a plausible mechanism for how high-carb diets can increase the incidence of prostate cancer.

My story is anecdotal, and there is much we don't know about diet, but "experts" still confidently tell us what to eat and what not to eat, and it's entirely possible that the advice given to me, in full good faith, 30 years ago, led me into a potentially fatal disease. And similar may be true for many others. And it is still going on.

I was referred to a radiation oncologist who advised radiation treatment, if not surgery. So, again, I did the research, and found that the latest advice for someone exactly my age and situation was "watchful waiting." I'm still more likely to die from something else than prostate cancer.

So why the recommendation from the oncologist? Well, it's what he does. Go to a carpenter, you are likely to get some advice that involves a hammer. But is he aware of the latest research? Probably, though possibly not. But he's not about to recommend something based on that, because it is not yet the "standard of practice," and he can get his ass sued. Even if the advice was right as to risks.)

Cascades are a real problem that dumb down social structures, and especially when they create a "scientific consensus" that isn't rooted in science and the scientific method. Cascades, however, occur in all kinds of social situations.

Comment author: aceofspades 19 November 2012 08:15:59PM 0 points [-]

I agree that low carb diets are an effective means of weight loss relative to low fat diets for people in the aggregate. I do not agree that they are in the aggregate better for reducing mortality than DASH, and I think my personal health is optimized by eating whole grains, fruits, vegetables, low fat dairy, and lean protein and avoiding all else.

Comment author: Abd 11 November 2012 03:30:20AM *  4 points [-]

Okay, read Taubes' article in the New York Times, "What if it's all been a big fat lie?". That's ten years old, there has been research published since then, but nothing to change the basic conclusions.

I suggest reading it before the rest here!

The organizations are not "scientific." They are largely political creatures, and how they are funded can be an issue. If cholesterol is not the problem, what happens to the statin drug market? But I don't know that recommendations are driven by funding.

Taubes is a thorough science writer, a skeptic, and it is indeed science that he's interested in. He is not selling a diet.

Taubes covers the history of diet recommendations in the U.S. It's shocking.

Something brief: In 1957, the American Heart Association opposed Ansel Keys (the author of the epidemiological study that got the whole fat=bad thing going), with a 15-page report, saying there was no evidence for the fat/heart disease hypothesis. Less than four years later, a 2-page report from the AHA totally reversed that, and, according to Taubes, that report included a half-page of "recent scientific references on dietary fat and atherosclerosis," many of which contradicted the conclusions of the report, which recommended reducing the risk of heart disease by reducing dietary fat..

What happened? Did the science change that quickly? Read Taubes! (i.e, read the book, "Good Calories, Bad Calories." Taubes also has a recent book, less technical, more popular, I think, but I haven't read it.)

I could point to studies; the Atkins diet in particular has been studied independently, and it improves cardiac risk factors, it does not make them worse. Yet it's a high-fat diet. So what is the risk?

Yes. I'm arguing against a commonly-recommended diet. I'm suggesting that relying on these agencies and their recommendations, without understanding the science, is very dangerous.

Taube had written a book about salt, and when he was doing the research, he noticed nutritional "expert" after "expert" who had no clue how science works, who used extremely poor reasoning, conclusion-driven. And he noticed the same when he started working on fat.

When I started reading in the field, out of personal necessity, I could see it myself, really poor "science" being commonly asserted as if it were simple fact.

Such as "a calorie is a calorie." I.e., it's said there is no difference between fat calories and carb calories, and the claim of Atkins that fat had a "metabolic advantage" was allegedly preposterous, this would supposedly violate the laws of thermodynamics.


  1. various foods take different amounts of energy to metabolize, and some calories are excreted.

  2. food calories are not thermodynamic calories, and this is not merely the "kilocalorie" thing, they are modified according to metabolic factors estimated from studies that were done about a century ago, and that may not be accurate under various dietary conditions.

  3. carb metabolism (burning glucose) runs the body in a different way, and has behavioral effects, compared to fat metabolism. Appetite shifts (fat suppresses appetite, generally).

There never was good evidence that saturated fats increased cardiovascular risk, that was speculation from the highly flawed Keys study. It was thought "well, to really know will take very expensive trials, we can't do that, so why not reduce fat? It can't hurt!"

But it could and probably did hurt. Lower fat in the diet, you almost certainly raise carbs, and quite possibly increase obesity, diabetes, heart disease, and there is an effect on cancer, apparently.

Bottom line, the officially-recommended diets have very little science behind them.

This really is not the place to debate the issue. Read the literature! Taubes is an excellent door into it, the book GCBC is about a fourth footnotes.

Or look at the Wikipedia article Saturated fat and cardiovascular disease controversy, (Do not trust Wikipedia articles to be neutral. They frequently are not. Use them to find other sources.)

It's tempting to sit back and trust the official organizations. It's a lot of work to actually read the evidence. However, is this important?

I thought it was, like, my life depends on it.

The AHA is a $600 million/year organization. If fat/heart disease hypothesis is as wrong as it appears to be, they may have cost Americans, in damage to health, a great deal more than that. Now, consider what we know about human organizations. When they get it spectacularly wrong, but before there is absolute proof, do they back up easily?

No. Their business is to be the experts, remember that $600 million per year.

Comment author: aceofspades 14 November 2012 01:33:10AM 0 points [-]

This pattern-matches exactly to everything else conspiracy theory related I have ever read, and by that I mean it misinterprets the relative incentives. You speak of organizations that apparently face financial loss if they turn out to be wrong, but you provide no convincing reason for why they would lose funding if they revised their positions due to new evidence. You also don't mention the huge profits an organization would surely make if it provided compelling evidence for how to actually lower the risk of the largest cause of death in the United States. In particular:

-I'm not going to read a book rather than reading the results of randomized, controlled trials or meta-analyses of many such studies.

-You say you "could point to studies." Then do it.

In response to Belief in Belief
Comment author: aceofspades 11 November 2012 12:59:40AM -1 points [-]

It doesn't seem to me that this post actually makes any coherent argument. It spends a fair amount of words using seemingly metaphysical terms without actually saying anything. But that's not even the important thing.

Is this post supposed to increase my happiness or lifespan, or even that of someone else?

Comment author: aceofspades 11 November 2012 12:53:49AM -1 points [-]

If this is article is actually correct, representative, etc. then the only thing it says to me is that the entire field of self-help is completely worthless, so I am going to actually operate under that assumption and just do what I want.

Comment author: Swimmer963 09 November 2012 03:24:35AM 4 points [-]

In fact constructing this abstract system does not seem to convincingly help me further its purported goal

I think this is a common problem. That doesn't mean you have to give up on having your second-order desires agree with your first-order desires. It is possible to use your abstract models to change your day-to-day behaviour, and it's definitely possible to build a more accurate model of yourself and then use that model to make yourself do the things you endorse yourself doing (i.e. avoiding having to use willpower by making what you want to want to do the "default.")

As for me, I've decided that happiness is too elusive of a goal–I'm bad at predicting what will make me happier-than-baseline, the process of explicitly pursuing happiness seems to make it harder to achieve, and the hedonic treadmill effect means that even if I did, I would have to keep working at it constantly to stay in the same place. Instead, I default to a number of proxy measures: I want to be physically fit, so I endorse myself exercising and preferably enjoying exercise; I want to have enough money to satisfy my needs; I want to finish school with good grades; I want to read interesting books; I want to have a social life; I want to be a good friend. Taken all together, these are at least the building blocks of happiness, which happens by itself unless my brain chemistry gets too wacked out.

Comment author: aceofspades 11 November 2012 12:35:26AM 0 points [-]

So the normal chain of events here would just be that I argue those are still all subgoals of increasing happiness and we would go back and forth about that. But this is just arguing by definition, so I won't continue along that line.

To the extent I understand the first paragraph in terms of what it actually says at the level of real-world experience, I have never seen evidence supporting its truth. The second paragraph seems to say what I intended the second paragraph of my previous comment to mean. So really it doesn't seem that we disagree about anything important.

Comment author: Abd 09 November 2012 03:48:35AM *  2 points [-]

There is a lot of really bad "science" out there on diet, there was a political decision made in the 1970s to promote low-fat diets, in spite of what most scientists thought. For a detailed story on this, and on what is known about fat and carbohydrates in diet, I suggest Gary Taubes' Good Calories, Bad Calories.

While little about diet is certain, the bulk of the scientific evidence is that "high saturated fat intake," in the context of a low-carbohydrate diet, does not increase real cardiac risk. On the contrary, high-fat low-carb diets, like the Atkins diet, lower cardiac risk factors.

The "scientific consensus" described above isn't.

This isn't about "paleo diet," as such, except that paleo diets do tend to be high-fat and low-carb. We did not evolve eating grain, and then the grain may be highly processed to remove most fiber, creating rapid absorption of glucose into the bloodstream, requiring, then, fast insulin release to avoid toxic levels.

We can eat carbohydrrates, but they were a small part of our diet, generally mixed with fiber, which slows digestion. Fat also does this. It's being claimed with substantial evidence that the "diseases of civilization," i.e., heart disease, diabetes, and cancer, are largely caused by diets with high carb content, especially highly processed carbs. High natural fat content does not seem to be a problem, the opposite.

I did the research and am bettiing my life on this. And I wish we knew more than we do. Taubes has started a Nutrition Science Initiative.

Comment author: aceofspades 11 November 2012 12:22:27AM -1 points [-]

Would you mind linking to this research that shows low carb diets lower cardiac risk factors? All I really know about the matter is that in the aggregate people who actually study diet generally conclude that Atkins-like diets are not optimal for health. In particular, the US Department of Health and Human Services, the Centers for Disease Control, the American Heart Association, and the World Health Organization all seem to conclude that saturated fats directly increase cardiovascular risk.

You're also arguing against anything said by these organizations when discussing highly processed carbs. DASH specifically recommends making at least half of grains consumed whole, and the implication seems to be that the ideal would be eating no refined grains.

Comment author: Swimmer963 09 November 2012 02:09:15AM *  1 point [-]

Longer? Probably not. Happier? Possible, depending on that person's baseline, since we don't know our own desires and acquiring these skills might help, but given the hedonic treadmill effect, unlikely. Achieving more of their interim goals? Possible if not probable. There are a lot of possible goals aside from living longer and being happier.

Comment author: aceofspades 09 November 2012 02:48:55AM *  2 points [-]

I have decided that maximizing the integral of happiness with respect to time is my selfish supergoal and that maximizing the double integral of happiness with respect to time and with respect to number of people is my altruistic supergoal. All other goals are only relevant insofar as they affect the supergoals. I have yet to be convinced this is a bad system, though previous experience suggests I probably will make modifications at some point. I also need to decide what weight to place on the selfish/altruistic components.

But despite my finding such an abstract way of characterizing my actions interesting, the actual determining of the weights and the actual function I'm maximizing are just determined by what I actually end up doing. In fact constructing this abstract system does not seem to convincingly help me further its purported goal, and I therefore cease all serious conversation about it.

Comment author: aceofspades 09 November 2012 02:00:33AM 3 points [-]

I have read this post and have not been persuaded that people who follow these steps will lead longer or happier lives (or will cause others to live longer or happier lives). I therefore will make no conscious effort to pay much of any regard to this post, though it is plausible it will have at least a small unconscious effect. I am posting this to fight groupthink and sampling biases, though this post actually does very little against them.

Comment author: pcm 09 October 2012 04:43:17PM 3 points [-]

When I switched to a mostly paleo diet this spring, I stopped needing willpower to prevent weight gain. I suspect I experienced other advantages, but don't have good evidence for them.

It will be hard to tell what fraction have benefited, because people who found it hard or ineffective are less likely to write about it.

Comment author: aceofspades 14 October 2012 09:02:39PM 2 points [-]

Maybe I should be more clear.

The anecdotes of a few people on this site mean very little to me in regards to the efficacy of a particular diet. There doesn't seem to be any experimental evidence with a reasonable sample size to suggest that Paleo diets actually lead to weight loss (there is evidence that DASH leads to weight loss). The paleo diet is relatively high in saturated fat (and there is a scientific consensus that high saturated fat intake causes heart disease) while DASH is not. Omitting grain and dairy eliminates the sources of some nutrients and I would hypothesize that a significant percent of people switching to a paleo diet don't actually compensate for that loss by getting those nutrients from other sources.

It just doesn't make sense to advocate for a paleo diet when there is no evidence of it performing better in the aggregate than diets which are supported by the scientific consensus. If I'm mistaken please link me to some good quality studies.

View more: Next