The problem with the question of whether this is safe in the long run for me is that once you get to the long run, causation is harder to establish. A year long intervention study in humans is pretty long as these things go, and already very expensive. Can I show that a long term ketogenic diet, talking decades here, is safe? I can point at individuals it doesn't seem to harm. The problem with this criticism of the ketogenic diet--we lack very long term studies--is also a problem with other diets. Do we know that 20 years of dietary fat restriction will do no harm? The studies really haven't been done. There are societies that have eaten more saturated fat than ours, other that have eaten less saturated fat, having a lower heart disease etc. burden. This doesn't mean that fat isn't a problem in certain contexts--it only suggests that it's not sufficient in and of itself to be damaging.

A high fat/high sugar diet causes noticeable moves in the wrong direction in real time, assuming that these will add up to problems in the long run seems prudent. A well-formulated ketogenic diet, when it causes improvements in the short run, it seems less prudent to assume that there will be a long run problem. When the only thing you can measure is short-term, assuming that the long term consequences will be the opposite of the short-term consequences doesn't seem to make that much sense to me. Especially where the ketogenic diet reverses or puts into remission hyperglycemia, hyperinsulinism and other aspects of insulin resistance and diabetes, I think the burden of proof is to show that it's harmful, since these are greater threats to health than even the purported dangers of saturated fat, even before beginning to doubt those dangers this is so.