The premise of this company is that we have taken on a project that some might regard as impossible. We’re building a computer that will do things that are impossible today. We take the absurdity of this project as a given. We believe that finding pragmatic, practical ways to do things that others regard as impossible is the purest, most impactful form of engineering.

We encourage debating ideas based on technical merit, intuition, calculation, analysis, experience, or vision. It’s critical that, when there is disagreement, we articulate the basis of our positions, and be prepared to defend them without turning business and technical discussions into personal conflict. It’s critical that even when we disagree, we understand the value of disagreeing but still committing to a single path forward. We do not have room for shutting down debate because problems look too hard or too weird. The best ideas often come out of the tension between creative new ideas and built-up expertise on what has worked in the past. In the end, decisions need to be made on the basis of analysis and data, not personality.

Our competitors are the semiconductor giants, like NVIDIA, Intel, and AMD, and they are outspending and outmanning us 1,000:1. This is OK -- the big companies will slow themselves down with bureaucracy, and their employees are often incentivized not to innovate. However, we don't- and won’t- have the same resources as a big company, or the luxury of incumbency. We'll seek to give you the money and resources you need to accomplish incredible things here, but relying on yourself means that when you need things, you directly put in the work to solve the problem or alert the appropriate team member, instead of letting the system handle it. If something is blocking you, figure out how to get unblocked, and escalate until you get what you need. Don’t just go silent and give up.

There is no work that is beneath any of us. This project requires grit, hard work, and a willingness to get your hands dirty. Everyone on this team needs to be prepared to put in the work to get things done in the face of adversity. Sometimes what must be done is to optimize speed over perfection, or to make a decision on partial information. What must be done is sometimes the 80 percent solution, so you can turn to the next problem. Part of this is the need to figure out how to make efficient use of your time - sometimes what must be done is to delegate, or to declare that you can’t do something that you’re being asked to do, because you have an overriding priority on something else, even when it feels unnatural to do so. The role of leadership is to help deconflict these priorities, but they can’t do their job without data from you.

We are building an incredibly complex system, with dozens of interconnected dynamic parts. On a daily basis, you will make decisions that affect other projects and your teammates. When you make those decisions, you need to think consciously about who needs to be part of the decision process, who needs to be informed that the decision has been made, and who will be affected. Communicating clearly who is part of the decision, and making sure that everyone who needs to know is informed, is critical. This doesn’t mean that all decisions should be made by committee. What it does mean is that when you make a decision, you need to be cognizant of whether you’ve gotten the right input from your colleagues - whatever that amount is - and that you make sure that your colleagues are aware of the decision. You are responsible for making good decisions, and part of making good decisions is getting the right inputs.