I just read an exciting new book called "Why we get fat and what to do about it."  Exciting to me anyway because he agrees with me 100%, 67% of the time. Since I'm right 98% of the time that's pretty good. So read this book. (Even though 91.2% of statistics are just made up)
Gary Taubes is a science writer of some acclaim and has done lots of important homework on obesity. He rightly points out that a lot of cogent, mostly German, observations we need to understand obesity are not new. Some are centuries old. This book is a more accessible sequel to his "Good calories, bad calories." (2007)
I agree with Taubes most when he says that the calories in - calories out explanation of obesity is fundamentally flawed, to put it more mildly than he does. He finesses the first law of thermodynamics, that dictates the calories-in calories-out rule, mostly by saying it's besides the point. He cites for example the statement that it doesn't explain obesity any better than saying that alcoholism is caused by over-drinking. So he mostly concedes that fat people do eat more. "We don't get fat because we overeat; we overeat because we're getting fat." Elephants eat a lot because they're turning into elephants, not because they eat too much, to quote another example he uses. However, there is good evidence, for example in genetically obese rats, that the fat rats don't eat more than lean rats. I think he overlooks another possible problem with the energy in - energy out rule. That is, organisms can vary the efficiency with which they burn calories.

He makes another good point here that we are talking about minuscule "imbalances" amounting to one bite of food a week for example, that are impossible to monitor in rats, much less in free range humans. Taubes comes tantalizingly close to finding out that you eat less when you exercise and get "in shape." He describes two classic experiments by Jean Mayer, an influential nutritionist way back in the 1950's. Mayer did an experiment in rats wherein he found that rats that exercised ate less than rats who didn't. Then Mayer made observations on workers in West Bengal. Mayer found that the workers who had desk jobs weren't any fatter and ate just as much as the the ones who did heavy work. Taubes proceeds to say that Mayer drew the wrong conclusions from his findings. That may be so but then so does Taubes when he repeats "exercise increases your appetite." I've said that exercise makes you eat less, not because it changes your basal metabolism,(1) although this is a little bit difficult for me to accept because it decreases your blood pressure and heart rate, (it increases your stroke volume - the amount pumped out with each heart heart beat) but because it decreases the energy cost of the exercise that got you into shape and any daily activity.(2) Taubes presumption that exercise makes you hungry comes partly from the fact that lumberjacks, to use his example - others would be Micheal Phelps and Lance Armstrong, have to eat a lot; but they exercise way more than can be reasonably expected for the rest of us and that overwhelms the savings of the rest of their day.
I also believe with Taubes that insulin makes you fat and so carbohydrates which make you secrete insulin are poison too.
He says that we've spent two and a half million years being hunters and gatherers that eat meat and vegetables and only twelve thousand years being farmers of carbohydrates. Then we got addicted to them, like beer and cigarettes. That's why we're not engineered to deal with carbohydrates. It's also true that you can do just fine eating only whale blubber. It doesn't really hurt your heart or your cholesterol as much as eating starch does, contrary to the authorities of not so long ago, especially if you lose weight eating that whale blubber. The myth that eating fat is bad has been busted most recently by Atkins and them. There is no need to beat that dead horse. We do need to keep beating the calories-in - calories-out-horse though. That one's not dead.
So Taubes answer to the question of his title is don't eat carbohydrates. I agree with a lot of what he has to say, but where do I disagree with him? 
Books like this, and I have a whole self of books with titles like "Why we get fat and what to do about it," often start out commendably by debunking the conventional wisdom, but then they all have to come up with the one simple secret that answers all your needs, that's also wrong. I'm not sure why they can't just stay with the message that this is a complicated subject and most of what you're told by so called experts, like me, doesn't work. It might be that their publishers tell them that kind of message won't sell books. Anyway in the case of Taubes the secret is just don't eat any carbohydrates. That is not a new idea, that's the Atkins diet. There are plenty of studies that show that even though you can loose at least as much weight on this diet as any other, maybe even more, in the long run, just a year or 2, you gain back the weight. It might be that no one can stay on a no-carbohydrate diet. That would be my problem if I tried, because never eating pasta or bread or fruit would be worse than death.

But I think it's more than that. After all, quiting smoking might be worse than death too, but I'm always telling people to do it anyway. In the case of the no-carbohydrate diet, at least  you don't have to be hungry. According to these guys, if you get hungry you can always just eat another piece of salami. Also the answer, when we finally do get it may indeed be something simple. There's nothing intrinsically wrong with the idea that there might be a simple solution. Still the evidence is that the no-carbohydrate diet doesn't work even if you stick to it.
On page 181 Taubes says that the incidence of heart disease has not diminished recently. That's not true. Also, Taubes' caviling to the contrary not-with-standing, the best evidence shows that the order from baddest to least bad is high LDL, low HDL and high triglycerides.
In the last chapter "Following Through" Taubes does some back peddling. This might not work for everyone. Can some people get away with eating just a little carbohydrate? Are some people so addicted to carbohydrate that even one bite will lead to falling off the wagon?
Finally, Taubes' doesn't satisfactorily speak to the 5 conundrums that you have to address with any answer to the obesity problem. 1. Why did we get so fat in the last 30 years? 2. How does biology defy the 1st law of thermodynamics. 3. How does the set point in your brain get reset 4. Why does obesity surgery work so well? 5. Why are we living longer and having fewer heart attacks if we're all fatter?