Work Hard

How much should you work? Sparked by the decisions and shifts in cultural expectations arising out of twitter, the question has been popping up lately and 140 characters isn’t enough room, so here we are?

In my opinion this question is defined by four broad concepts:

  1. Problem space
  2. Incentive alignment (see: principal vs agent problem)
  3. Internal vs extrinsic motivation
  4. Unlimited PTO

Problem Space

The first and arguably most important question, what is the problem space you’re trying to solve? A problem space does not need to be extremely difficult and require a high skill level to be useful to society, function as a business or run as a company.

What are the bounds? There is insane variance when it comes to writing software.

Writing a game engine from scratch for example isn’t building a distributed system serving millions doing complex logic. A small node.js app running on heroku serving 10k requests a month can be doing something extremely useful in society unlocking value for a lot of people.

Not all problem spaces demand a serious work ethic or mode of operation.

Incentive Alignment

If you want to delegate work to someone else, and you want them to work as hard as they can, there’s a level of sacrifice in ownership and control required. It requires trust and respect that the person on the other side will execute with freedom to do as they see fit.

This is explained best in my experience by Graham Duncan with the idea of a decision space.

There’s a large spectrum on different agreements that solve the agent vs principal problem, insanely varied levels of autonomy, decision spaces and incentive models. The schelling point is different from one company and industry to the next.

Internal vs extrinsic motivation

It doesn’t matter that you’re giving ownership or are solving an important hard problem. Whoever is on the other end has to want it for their own reasons. You can’t ask someone to sacrifice their life in service of building something no matter how great. It often requires painful sacrifices.

You can’t really threaten someone into productivity. Forcing people to work under threat of losing their jobs may work temporarily but any rational actor will escape that situation eventually.

Unlimited PTO

I don’t know about making anything else, but in my experience/opinion I cannot operate at a fixed rate of production. Knowledge work is not an assembly line where you can see consistent output.

If I had to give arbitrary figures think 60-70% over a 40 hour week is a bad model. It’s more like -30% - 5000% over a sporadic period of time. I sprint intensely and take breaks. Without these breaks my mind breaks down into a fog and it begs me to do something anything mindless.

This rate of production depends again on what the output is. Is it something I already know how to do? Is the output hundreds of api endpoints doing a different version of CRUD?

I don’t think it’s a question of sitting for 6 to 9 hour stretches a day churning out code.


I believe in working as hard as is humanly possible, I must, I’m not nearly smart enough to think otherwise. The real world is messy and to create almost anything even of little value you have to give it your all to stand a chance.

Sometimes that means working really hard for longer streches of time than I am comfortable with, these are external constraints. I take what is given and do what I can.