Can you expand on this?

Sure. Let’s say you have some mundane task: you’ve got some number of people with varying dietary restrictions and they want to go to dinner together. You don’t know which places can accommodate you. Don’t do it yourself. Just ship the info over to someone you pay $25 to do the work. You can then use the time more productively.

Eventually, you can see the MTurk / Upwork folks as just slow executing programs. For some jobs, some guy in Southeast Asia will do it for $4. Don’t even think about it.

This works across the spectrum: want to clean up formatting in a doc? Ship it over. Want to transcribe a video for captions? Ship it over. Want a quick scraper for a few websites? Ship it over.

The objective is to not be the bottleneck on any task.

Get a quick idea of what the labour looks like. Southeast Asia at $5 is great. Americans at $15 are a waste of time. South Asians at $8 will break even. Eventually, every hour of your time is spent on the time that is special about you.

I suspect that when you master delegation you improve productivity by an order of magnitude. It's a game changer.

This sounds great. How can I trust the quality of the output?

Many of the things are easier to check than to do. And where it isn't, you can have multiple people do the same thing.

And last of all: if you build a working relationship with the guys who do good work you can use them for more work.

I'd be worried that multiple rounds of feedback back and forth would cost more time and money than doing it myself.

Having multiple people do the same thing, and finding people who are quality-oriented and trying to stick to them sounds clever, though.

I never really give feedback. I either keep on or end contracts but I never attempt to coach them into being useful. The cost/benefit isn't there, as you pointed out.

Sometimes, it's just that the question isn't suitable for this method and I just sink the project till I can think of better.