Hacker News Reader

Comment by jajag | Hacker News Reader

Yes: https://en.wikipedia.org/wiki/Extracorporeal_membrane_oxygen...

Interesting, thanks. Apparently it's already being used to treat covid-19. But the potential side effects seem to be nasty and the survival rate not great, so I guess it's only really a last resort treatment.

Human bodies are just too damn complex. Even when the problem is clearly identified--a lack of oxygen in the blood--there's no way to just insert more that doesn't cause a hundred new problems :-/ Natural selection seems to be really bad at modular design.

> Natural selection seems to be really bad at modular design.

Or the other way around, modular design does not appear to be a requirement for "success.". We like it for the reduction in complexity but that is ultimately a tradeoff.

Well, it’s successful until it’s not. Nature’s primary approach to debugging seems to be: you die. We clearly weren’t intended to open up the hood and fix things.

Biology doesn't care about individuals. It cares about populations. The main mechanism is exponential growth, capable of quickly filling carrying capacity boundaries of arbitrary size. Individual death is a small setback, as another branch of the exponential growth tree takes over shortly.

Biology is layers and layers of legacy code and technical debt and it never gets refactored.

There's also a concern that the filters in an ECMO machine damage or eliminate the antibodies that are trying to fight the virus.

So putting someone on an oxygenator might condemn the patient to stay on one until the virus kills them.

Reply on news.ycombinator.com