Building a world-class team

This is an excerpt from a recent Company Update email I sent to the team at Kognic:

What does it take to build a world-class team? To learn and evolve, I read biographies about companies and their leaders and talk to as many CEOs of great companies as possible. A recurring theme of books and conversations is the critical importance of strong teams of excellent people. That, in turn, requires great culture and outstanding leadership. But what is a great culture and outstanding leadership? I start with a few principles:

- Transparency. Great teams share information freely. They give and collect lots of feedback.
- Performance. Great teams expect greatness. They challenge themselves and others to grow.
- Empowerment. Great teams empower each other and win together.
- Rapid Iteration. Great teams iterate faster than everyone else.

But, as we all know, culture is the sum of our daily actions, not something we write in a document. Let’s start with some theory before we dig into specific challenges.

The Principal-Agent Problem
The Principal-Agent problem describes the conflict of interest that arises when one entity takes action on behalf of another. The problem worsens the larger the discrepancy between the two entities’ interests and information. The difference between what the agent does and what is actually in the principal's interest is known as the “agency cost.” To make this more concrete: Shareholders in a company hire a CEO. How can shareholders trust the CEO to make decisions in their interest? How are you sure the CEO does not just act in her best interest? The CEO will have time, expertise, and information that shareholders do not. The same problem applies to all executives and their teams.

I usually share something one of my mentors tells me: "Smart people with the same information and goals come to the same conclusion. The problem is that people have different information (or experience) and different goals (or risk preferences)." Culture is to select people based on their commitment to similar goals, risk preferences, and determination. It's exclusive but not based on gender, ethnicity, or experience - it's exclusive based on goals, preferences, and determination. If someone isn't smart enough or driven enough, I reserve the right to terminate that person. And I think that's what most of you want. You want to work with like-minded, inspiring, and incredible people.

Handling the Principal-Agent problem is the central challenge of managing teams and companies. You have to account for a vast range of things:

=> What makes you motivated? This is the first critical divider in terms of culture. Different people are motivated by different things. Some want work-life balance, psychological safety, and predictability. My wife is an example of that: we complement each other in a great way because she favors a stable household and lots of time available for her non-work interests. She likes her job, but it's not nearly as important to her as it is for me. She'd hate working with me professionally, and I love her for that. I need to wake up in the morning and feel like I'm impacting the world. I need to feel like I'm constantly learning. Why? Well, I suppose I have my fair share of demons. At least I've learned to harness them and channel that energy to create exciting things. There is a Hunter S. Thompson quote that sums up my worldview pretty well: “Life should not be a journey to the grave with the intention of arriving safely in a pretty and well-preserved body, but rather to skid in broadside in a cloud of smoke, thoroughly used up, totally worn out, and loudly proclaiming "Wow! What a Ride!” (Caveat: HST was a gun-loving drug addict who died by suicide. But it's still a great quote!)

=> I tell everyone who is about to join Kognic: there are much easier jobs than working at Kognic. We offer a high-growth environment and are determined to be the best in the world. That's a mindset that we are looking for in candidates and ourselves. You must be smart and driven to thrive in our environment. And we must have the courage to assume we can continue hiring insanely intelligent and driven people. Sometimes, that means it will take longer to hire for specific roles. It also means hiring managers must be attractive leaders for whom great talent wants to work. Leaders cannot sit around and complain about facing a low-quality hiring pipeline. That's just a symptom they are not spending enough time out there attracting talent. Great leaders are talent magnets. Poor leadership is the main reason great talent leaves. This does not mean we expect everyone to work crazy hours - it is a marathon, not a sprint. You will have bad days and good days. We cannot always bring our best every day - but we can always be committed to becoming the best over time. I'd much rather feel the heat of high expectations than be bored, and we look for people like that when hiring. It's not for everyone, but it is what we look for at Kognic.

=> How do you balance high expectations with being friendly and social? I'm proud that we have a “zero asshole policy.” I do not tolerate team members going around being rude or disrespectful. Hearing someone talk shit about someone else when they are not around is a grave offense in my book. If you have feedback, tell it to that person’s face. If that doesn't help, bring it to your or their manager. Once you've done that, act like a professional. There is no reason we cannot combine being fair to each other, have fun together, and expect extraordinary things from each other. Sure, leaders in the company need to be prepared to make difficult decisions that might hurt people when required. But there's no reason to be an asshole about it. I think we've done an excellent job balancing high performance with a strong social bond in the team. Toxic team members can quickly spread a lot of negative energy. Everyone should welcome raising issues, but the best people can constructively raise issues. They don't just whine about it. Most people I talk to are impressed by how friendly and nice our team is, and I agree. Clearly, we can both create a high-performance culture and be nice. All that said, at the end of the day, we are colleagues first, friends second. I will always make decisions based on what is in the interest of Kognic and our mission to accelerate alignment. Sometimes, that will be at the expense of individuals. I acknowledge this, and I think all companies and leaders should be honest about this. I encourage everyone to have a solid social network outside of work. If you don't, you become very vulnerable.

=> How do you balance being challenged with being trusted? There is an inherent tension between having a manager who challenges you and feeling that you are trusted. I think we all want to feel trusted. But I also firmly believe that high-performing individuals want to be challenged. I would never work for a leader who doesn't express expectations or cannot push my personal growth. I want to work for people I admire and am inspired by, and I think we need people who feel this way across the company. Steve Jobs despised "the professional-managerial class", and so does Elon Musk. I tend to agree: Kognic wants leaders who are able to jump into the arena when needed and who can inspire teams to do better based on their experience and expertise. Clearly, great ICs are not necessarily great leaders, but the best leaders are usually also great ICs in some area. We have to believe we can find those that are both. We have several great examples of people like this at Kognic: leaders who could step into most roles and do a pretty good job. I expect those who lead at Kognic to know what is going on in their team, to have a working knowledge of essential details, and to have the ability to inspire their team members with their experience and expertise. And what you lack in experience, you can always make up for with determination, grit, or raw talent. In my case, I make up for my lack of experience by working more—a lot more. I need to collect and spend way more time analyzing information to be the CEO Kognic needs since I'm 35 years old without any previous experience running a company. I'm not complaining; I'm just saying it takes a huge effort.

=> Empowered teams with context can make better decisions, but leaders need to still challenge and coach. At Kognic, we are strong believers in empowered teams. We think teams must be accountable for delivering value without leaders imposing solutions. We expect leaders to set priorities, declare intentions, and then hand them to teams to own the outcome. That's not the same as "servant leadership". Leaders are not simply there to observe and support. They are there to inspire, challenge, coach, and help when needed. They should be knowledgeable enough to analyze the situation and ask critical questions when needed. They should challenge the scope. They should trust but verify. We cannot let a desire to "get out of the way of teams" be confused with "not setting high expectations and challenging each other.” Leaders need to ask, "Can this be done faster?" or "Is this necessary?" or "Have you considered this priority?". That does not go against our intention of letting teams take ownership of their work. Empowerment is a more demanding way to work. It requires everyone on the team to take ownership of what creates value. You cannot just expect a manager to tell you what to do; that's a cop-out. We have selected a very demanding operating mode, and it only works if we have excellent people we trust can take full responsibility. I always argue leaders should coach, and ultimately replace, people if they feel compelled to micromanage them. Always resist the temptation to start micromanaging. If you don’t think your direct report can handle the responsibility you give them, you coach them as far as you can. If that doesn't work, you replace them.

=> We are morally obligated to act ethically as a company and individuals. It's not just about winning but also how we win. We would win more by lowering ourselves to Scale’s workforce compensation level. We won't as long as I'm CEO. I refuse. I don't want money gained at the expense of other people suffering. I also won't summarily fire someone the way Musk or Jobs seem to have done during their careers. I won't publicly berate someone for being an idiot, even when they are, in fact, being an idiot. It goes against my morals to incur suffering like that, regardless of how righteous I might feel. Does this limit my success? I don't care. I genuinely think we can be ethical and win; however, being ethical is non-negotiable. Again, being ethical is not the same as lowering expectations or failing to make hard decisions when required. I will terminate people who do not meet the bar for working at Kognic. But I won't publicly murder them. We will do it fairly, and we will be generous if we've failed when hiring. I think we will be more successful if we prioritize physical and mental health in the team. Sleeping under your desk is stupid. It might make you feel like a hero, but I think it just makes you look like an unstable person who has no idea how to manage time or teams. People can only do so many productive hours of work in a sustained way anyway. There are many other types of "posing" like that out there that I think are entirely irrational.

=> How do we balance psychological safety versus urgency and accountability? Fear is the mind-killer. Scared people will not take enough risk. They won't think outside the box. They won't object to the premise of misguided requests from management. Fear will make people shut up and focus on survival. Fear shuts down the flow of information, and it hurts creative thinking. Management by fear has severe adverse side effects. At the same time, urgency and accountability encourage creativity and progress. Being too relaxed is not suitable for doing a great job. Most people get more done right before deadlines. Just look at the finishing time for the London Marathon. Spoiler: it is not a normal-distribution.

=> I believe that leaders need to create a strong sense of urgency but without scaring people. I try to combine relentless urgency with profound optimism. We have to always be in a hurry to make progress, but we have to approach tasks with joy and curiosity. If someone gives you a timeline for six months, ask what we can do to make it take two months, but do not threaten the person who gave you the estimate. Delete everything that isn't critical to your mission, but do it with curiosity. Build excitement around speed. Faster is always better. It's so easy to fall into the trap of assuming, "This needs to take X months." Large companies have such low expectations from individual contributors that they get used to things taking a long time. If every person in a chain of five lowers the bar by 10%, you end up with a significant accumulated delay. But if each layer hunts down 10% of speed-up instead, you have a massive difference in accumulated advantage. That's the difference between mediocre and great companies. You will move slowly if you hire slow people and let them make slow plans. Instead, put impatient and relentless people in charge and eliminate all the slow people. I promise you that you will get results much faster. And talented people will be happier when things move forward quickly. My experience is that 90% of the work can be finished in 10% of the time if you are OK with 20% of things going wrong instead of 10%. That's an acceptable failure rate to be faster than everyone else. The value of being 10x faster far exceeds being right 10% more often. Most decisions can be changed anyway. As you know, I prefer being right over being consistent. So, I expect leaders to challenge timelines and hunt for speed whenever possible. You have to question the scope of projects to find just the right effort for our stage. Tuning the scope requires a deep understanding of the work being done and a fair amount of impatience. Delete as much as possible, and if you add about 20% back in later, you've probably found an decent deleting level.

=> How transparent can we be? I prefer to deal with things openly in a large group, over 1-1s, or closed groups. The more information we can push into the open, the better. The fewer closed Slack channels, the better. You can all learn to filter that which is relevant from the rest. Being exposed and tuning your filter is better than being out of touch. Cross-functional teams with direct communication between ICs far exceed hierarchy in speed and efficiency. Managers are not there to monitor and approve things but to challenge, inspire, and support. Report your conclusions and broadcast them as widely as possible. But do not slow down by going through intermediates. Rather than talk to someone's manager to figure something out, talk directly to the person you need to talk to. I expect leaders to share openly the good, the bad, and the ugly.

=> There are some limits to transparency. We do not publicly criticize a person's performance. We do not share all the details when someone is terminated. We cannot talk about sensitive aspects of negotiations with customers or investors. We cannot share if/when someone offers to buy the company. There are limits. But we set a high bar for what should be shared. We always share if we have problems in a process or if some deal is going south. We openly share our call reports, meeting reports, financials, etc. I'd much rather send a few emails too many than too few.

=> Leaders create accelerating flywheels. The most successful leaders create "flywheel teams." Teams that accelerate more and more and eventually spin really, really fast with minimum external force applied. That requires removing friction, setting a clear focus, and ensuring everyone on the team is excellent. When I read, e.g., the biography about Musk, he appears not to scale. At all. I think that's his greatest weakness. Teams appear highly dependent on his intervention. While he gets the urgency right and can attract incredibly talented people, he constantly involves himself to such a degree that the future of his companies is uncertain without him. I suspect he craves being at the center of attention more than is in the interest of his companies. But I guess you must take the good and the bad with some people.

=> Leadership is telling an engaging story. A big part of leadership is being a great storyteller. We are all part of a narrative. Kognic is a story about a team trying to develop safe and reliable embodied AI. We believe that AI is the "highest order bit." It is the most important thing we can work on during our lifetime. If we get AI Alignment right, everything else will follow. Every decision we take, every customer we sign, or every product feature we build needs to move us toward our vision of accelerating alignment. To reach our goal, we must dance with armed elephants, fight Kafka-esque complexity, and overcome many other challenges. It's an adventure. Leaders at Kognic are expected to be able to connect their work and the work of their team to our vision. To weave their thread into the story. What is your team doing to accelerate alignment? There are many ways to help: contributing to a lean data factory, helping hire fantastic talent, ensuring we get paid for our work, signing great contracts with customers, etc. We must be useful and valuable to customers now to gain the trust and resources needed to change the world. Every day, we become bigger and more capable, and everyone on the team is part of making that happen.

Kognic is not looking to be like most companies. We are the best company in the world at Accelerating Alignment. Outstanding leadership at all company levels is required for us to succeed. While demanding, the reward is that we get to enjoy a hell of a ride with ambitious and smart people.