One reason any dietary change can cause weakness is that the body doesn't produce enzymes it doesn't need. When you're eating mostly carbs, it doesn't produce as much of the fat-burning enzymes, and vice versa.

This is why they tell you to eat a lot of carbs before taking a Glucose Tolerance Test. If you haven't been eating carbs, you won't be producing carb-metabolizing enzymes and you may test diabetic when you're not.

The ultimate low-carb diet is starvation, and people who have been starving will test diabetic on a GTT.

I once read a book by some minister who was preaching to eskimos. At one point his guide, who was on a high-meat, low-carb diet, had planned for them to stop at a village to buy caribou from the residents. When they got there, they found the village had been abandoned, and they were forced to survive on the starchy food they'd brought for emergencies. The eskimo complained that he had no energy on the high-starch diet. It was probably the same phenomenon. He had a lot of protein- and fat-metabolizing enzymes but not a lot of starch-metabolizing enzymes.

BTW, Cordain is wrong if he says the brain can use only glucose. They've found that the brain can use up to 75% ketones after it's given a chance to adapt. The "only uses glucose" is one of those old saws that refuses to die, like the idea that carbohydrates are essential or low-carb diets damage your kidneys and leach calcium from your bones. Like adaption to a LC diet, the brain takes several days to adapt to using ketones, so some people don't feel like themselves for a few days if they go abruptly from a high-carb to a low-carb diet.

Michael Eades mentions this in one blog post, although he doesn't give the percentage. I think I got that from a biochemistry text, but I'm not sure.