They didn't claim evolution got it perfectly right.

They speculated that immune systems evolved to avoid being continuously on alert. And that's exactly right- our immune systems have an extremely complicated system for detecting foreign invaders that is tightly regulated. And a failure to regulate that is often associated with autoimmune disorders, which remain very poorly understood.

I've studied biology from the perspective of engineering better drugs for decades now and I can say with confidence that I simply don't understand how the immune system works, and I don't think anybody else really does either (compared to, say, the heart, or many biological systems like protein production). We have identified many players, and observed a great deal of actions, and have speculative models for many of the underlying processes, but we don't really have an "understanding" of the immune system. I skimmed this paper and frankly, it has a very long way to go before people are convinced to try this in human clinical trials.

I look forward to innovations, but to a first order approximation: evolution found model parameters that exceed the best human science and engineering.