Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I think a key here is recognizing when the expected result is catastrophe and acting accordingly.

There's a whole lot of procedures in the hospital that are designed for normal times and are _good_ in normal times(*) - they enforce people double-checking their results, they give radiologists time to make the right call, they make sure the right drugs are going into the right people, they make sure patients are prepped correctly for a procedure. All of that takes time, but it moves your expected outcomes from, say, 95% to 99%.

In the case described, the expected outcome is 0% - it's a fucking catastrophe, there's 250 people on their way, and everyone who comes in is a corpse waiting to happen. At that point, absolutely, flow is king - hitting an expected outcome of even 50% (looks like they got somewhere around 65%) is a massive upgrade, so yeah, shove as many people through the CT scan bay as you can, let your radiologist flip a coin, and if the nurse gives the wrong drug, fuck it, that's probably not what killed the patient - move on and hope you do better next time.

I think the real takeaway here is recognizing what the stakes of your situation are and acting accordingly - what are the outcomes that matter, and what's the set of activities that get you closer to your desired ones. Sometimes that's care and caution, and sometimes that's taking your best guess, committing to it, and accepting the outcomes.

Huge kudos to the author and to whoever set up the system that enabled them to act as needed and supported them in doing so, though - I'm sure they broke every regulation in the book over the course of those six hours. Plenty of other disasters have been made much worse by the lack of flexibility in a system to recognize extreme circumstances and act accordingly.

(* take "good" with all the caveats you want here. I get it.)



"Move fast and break things" is absolutely the right method in some scenarios. The trick is, as you say, identifying if you're in one.


“It depends” is certainly the answer more often than people want to believe.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: