In reality, though, how often does it need to be fixed NOW? Like, actually needs to be fixed now, or there will be serious consequences, and not just some manager thinks they want it now? I'd be willing to bet it's less than 10%.

Reality is basically never, unless it's a major system outage, or security issue. We don't even bother looking at issue priorities, because everyone raises them as critical.

Even then, is it really an emergency? What's the consequence of it being down? Death? People being fired?

IME, 90% of emergencies are "such-and-such a manager is going to look a little bad in a spreadsheet next Monday".

That doesn't mean it doesn't need fixing, but we should start being honest with ourselves. Defining P1, P2, etc. is useful.

Well in the case of Amazon, $66,000 per MINUTE. If you want to make a case for money only being money sure. But as far as business is concerned, that is an emergency.

Well to be fair, we're not talking about emergencies, obviously those get priority. And of course, as much as people would like you to believe, most things are not an emergency!

What about the 3rd option, getting two tasks done?

- Grind for 1 hour and assuming 90% success rate, finish the task 9 out of 10 times

- Sleep

- Finish another task in 15 min with, again based on your assumption, a success rate of 100%

In start-ups there is usually a near endless list of tasks to solve so this would be the most productive approach. Again, based on a lot of assumptions.

[edit: formatting]

Except that success rate was assuming that you did get the rest you needed, mainly from leaving at a reasonable time. Staying and grinding directly contradicts that, so you're not going to be rested and well in the morning, and you're not going to have the success rate of 100%.

Basically, you just advocated for crunch time and death marches. Things that are known to not work.

You haven't prepped and then slept on task 2 so I don't think you're finishing it in 15 minutes. And now you're sleep deprived.