Diet and exercise generally do not cause substantial long term weight loss

There is a pretty serious selection bias in that study.

I know some people who lost a noticeably amount of weight and kept it off. These people did NOT go to any structured programs. They just did it themselves.

I suspect that those who are capable of losing (and keeping it off) weight by themselves just do it and do not show up in the statistics of the programs analyzed in the meta-study linked to. These structured programs select for people who have difficulty in maintaining their weight and so are not representative of the general population.

Diet and exercise generally do not cause substantial long term weight loss.

"Healthy diet" and dieting are often two different things.

Healthy diet might mean increasing the amount of vegetables in your diet. That's simply good.

Reducing your calorie consumption for a few months and then increasing it in what's commonly called the jo-jo effect on the other hand is not healthy.

Why is this surprising? You give someone a major context switch, put them in a structured environment where experts are telling them what to do and doing the hard parts for them (calculating caloric needs, setting up diet and exercise plans), they lose weight. You send them back to their normal lives and they regain the weight. These claims are always based upon acute weight loss programs. Actual habit changes are rare and harder to study. I would expect CBT to be an actually effective acute intervention rather than acute diet and exercise.