I like the “I, Pencil” essay.
Matthew Kwan

I know how big cloud computing is. The problem isn’t how much compute you can throw at it, it’s that you can’t get accurate, timely values for all the problem parameters. My preferences vary from day to day. How in the world are you going to measure what I value and optimize for it? How would you ever measure your own preferences?

How we do this today is with the price system. It’s the only way we’ve come up with to get people to really tell us what they prefer and how much something is worth to them versus something else.

Here’s another problem. How would you do hiring? That’s totally a judgement call. I can try to hire a cheap new college grad and hope they don’t slack off too much. I can hire a proven rock star coder and hope they’re not a prima donna. I can hire someone in the middle and try to negotiate a reasonable salary. Regardless of which I pick, how will you predict how productive each of these people will be?

Shoot, I have no idea how you’d measure my productivity. I don’t just crank out widgets on a production line, that’s what robots are for. I have to do creative stuff which is different every day. That’s becoming an increasing chunk of our economy. I don’t believe for an instant you can actually put this into numbers and a ginormous matrix problem.