• Having good delusions without being delusional. Building great things with a big impact. It's easy to have 10x improvement in how food is delivered but 100x harder to improve the nutritional composition of the world's consumption. We'd much rather fail at attempting hard problems than win small ones.

• Asking for help. There's a low likelihood we're the best at what we do. Be honest in acknowledging your current limits and learn fast. Build your own support network for your role.

• Make strong distinctions between things that are important, urgent and useful, and those that are not.

• Irrespective of your role, talk to the users about the problem before you retreat into deep specialised work to build the solution. Everything you do without knowing its value to the end customer is of zero value to the company.

• Err on the side of hurting yourself than real people in the real world. Consequential problems involve a significant impact on people's lives, like if they are able to find work or not. Build robust systems that don't use ambition as an excuse for harm. Beware of your net impact on people’s lives. Fix those issues with the urgency of survival.

• Radical transparency about the financial position of the company. You'd inevitably bring your personal life to work, be open about things that affect your motivation and stability.

• Do things that you can defend fiercely. Few people within the company will ever find out you made a suboptimal technical choice, hired an incompetent person, wasted time on what didn’t matter or made a mistake. Ask for forgiveness and not permission.

• Have fun. It's unsustainable to work on things you don't enjoy for a long period of time.