Ask HN: What Makes a Great CTO?
Once a company gets to a point where the first coder is no longer doing hero work on the platform and has to lead - what habits, attitudes, practices, and skills lead to greatness? Books or talks in the subject?
2 comments
A startup I've worked for in the past had one of the smartest cofounder/CTOs I've ever worked with. He was also arrogant as heck. The engineering team inherited his arrogance and it was quite an insufferable place to work. I left that company for a number of reasons, this was certainly one of them.
I get it. When I was young in startups a decade ago I was often called the rockstar or ninja. I was arrogant then too. It felt good. But it doesn't do good for people who come into the company later. The culture is already toxic by the time the team is growing. Now that I'm older, I'm extremely mindful of this behavior now and make sure I don't repeat these mistakes as a team lead.
There are tons of other tips and advice for this position too, but this was definitely the first thing to come to mind.
I get it. When I was young in startups a decade ago I was often called the rockstar or ninja. I was arrogant then too. It felt good. But it doesn't do good for people who come into the company later. The culture is already toxic by the time the team is growing. Now that I'm older, I'm extremely mindful of this behavior now and make sure I don't repeat these mistakes as a team lead.
There are tons of other tips and advice for this position too, but this was definitely the first thing to come to mind.
"If I disappear, will the company survive at worse and thrive at best?". If the answer is "No", asking "Why?" recursively will give pointers on what's wrong and what needs fixing.
Working hard to put yourself out of a job can lead to very interesting results in my opinion.
"If I disappear will they be okay?". "No.". "Why?", "Because they don't know how to build product yet?". "What does that mean?", "They can't focus on what's a priority and what's not?". "How so? Can you? How do you do it? How can you make sure everyone knows what's important and what's urgent, and how to shift, and based on which criteria? Can you put that in writing so eveyrone knows how to make these decisions? Do it, share it, have people poke holes into it".
Building a team where if something were to happen any one of the members could step in to fill in the role, at least temporarily avoiding a panic, or even better, run with the team better than I did, is a goal.
This requires explaining everything you do. The reasons, process, tradeoffs, and thought process behind your decisions. Systematically demystifying and removing even the slightest chance that what you're doing could be attributed to "talent" or being "special" or "smart", and working to make it learnable, systematic, repeatable, transferable, and institutionalized so it becomse the baseline at your organization. Doing this gives many instances of decision making, learning and thought processes, work habits, communication.
Examples people can learn from and hopefully poke holes through to expose weaknesses and flaws so that you too learn from them.
Doing this also instills a great conversation style in the team where someone does something, and the other person saying: "Stop right there. How did you know you had to do that to fix this?". I have colleagues who'll ask that kind of questions where it's not enough that you do something, but they're interested in the meta skill that drove you to make a decision. Good or bad, we often back-track.
We share steps that lead us to an action or conclusion. For example, we might look at our hiring. People who quit or whom we let go, and then dissect every case to amortize that experience and make it more humane, or avoid the events that lead to it. For example, asymmetry of information that lead to weird conversations. Spotting and stopping unproductive behavior earlier and correcting.
We might look at all the machine learning products we built for our clients and think about why they took so long or cost that much, extract common functionality, find common problems, detect that the barrier to entry is high, and abstract these. This is how we decided to build our ML platform.
I come from a control theory/optimal control background, and this is one of the ways I look at things. Robust against peturbations, stable, adaptive, optimal systems in everything, while being conscious of the tradeoffs of these optimizations.
When you see someone doing something that's clearly working, think of adopting it for a while. If it gives good results, think of ways to make it systematic and increase the body of knowledge of the team/organization.
When you learn something, do the same.
Model the habits you want to see in others. The team will take note it's clearly working for you and will give them a try because they want to have those results as well, and then they'll adopt them and improve them so you get to adopt the better version.
Have a list of mistakes you have made, how you've solved them, and how you fixed the thought process that lead to them. When one makes that mistake, you correct it and share the instance of when you made that mistake, and how you went about solving it. People learn way faster and the whole experience is more agreeable. It also instills intellectual honesty and once it's there, it's embedded in the team. It avoids un-necessary frictions and ego trips.
Being a gate keeper when it comes to hiring. The standard is not hiring someone I wouldn't accept as my team lead/CTO/CEO/COO some day. We hire people with the possibility in mind that they might run the company some day, and thus we're careful. The people we hire will lead their own teams in the near future. We weed out behaviorally flawed people early on: dishonest technically, not taking responsibility for things, changing facts as we wouldn't trust them, and we can't really work with people we don't trust. Example: if something breaks, I don't want someone to cover their tracks and waste us precious time we could be spending fixing what broke.
Involve the team in decision making by sharing all the information you're using to reach for conclusions. We have git repositories with our meeting notes with prospects, investors, advisors, organizations. Dissemniation of financial and business information. People who weren't there read them. Even a fresh graduate can read them and learn things, ask questions, give feedback. It's a great way to discover new ways of doing things when you confront people to problems.
All in all, work systematically to make whatever put you in a certain position to become a given at a company, or at least disseminate it so you're not the only one who either can do it, or has a blueprint to how to do it.
Working hard to put yourself out of a job can lead to very interesting results in my opinion.
"If I disappear will they be okay?". "No.". "Why?", "Because they don't know how to build product yet?". "What does that mean?", "They can't focus on what's a priority and what's not?". "How so? Can you? How do you do it? How can you make sure everyone knows what's important and what's urgent, and how to shift, and based on which criteria? Can you put that in writing so eveyrone knows how to make these decisions? Do it, share it, have people poke holes into it".
Building a team where if something were to happen any one of the members could step in to fill in the role, at least temporarily avoiding a panic, or even better, run with the team better than I did, is a goal.
This requires explaining everything you do. The reasons, process, tradeoffs, and thought process behind your decisions. Systematically demystifying and removing even the slightest chance that what you're doing could be attributed to "talent" or being "special" or "smart", and working to make it learnable, systematic, repeatable, transferable, and institutionalized so it becomse the baseline at your organization. Doing this gives many instances of decision making, learning and thought processes, work habits, communication.
Examples people can learn from and hopefully poke holes through to expose weaknesses and flaws so that you too learn from them.
Doing this also instills a great conversation style in the team where someone does something, and the other person saying: "Stop right there. How did you know you had to do that to fix this?". I have colleagues who'll ask that kind of questions where it's not enough that you do something, but they're interested in the meta skill that drove you to make a decision. Good or bad, we often back-track.
We share steps that lead us to an action or conclusion. For example, we might look at our hiring. People who quit or whom we let go, and then dissect every case to amortize that experience and make it more humane, or avoid the events that lead to it. For example, asymmetry of information that lead to weird conversations. Spotting and stopping unproductive behavior earlier and correcting.
We might look at all the machine learning products we built for our clients and think about why they took so long or cost that much, extract common functionality, find common problems, detect that the barrier to entry is high, and abstract these. This is how we decided to build our ML platform.
I come from a control theory/optimal control background, and this is one of the ways I look at things. Robust against peturbations, stable, adaptive, optimal systems in everything, while being conscious of the tradeoffs of these optimizations.
When you see someone doing something that's clearly working, think of adopting it for a while. If it gives good results, think of ways to make it systematic and increase the body of knowledge of the team/organization.
When you learn something, do the same.
Model the habits you want to see in others. The team will take note it's clearly working for you and will give them a try because they want to have those results as well, and then they'll adopt them and improve them so you get to adopt the better version.
Have a list of mistakes you have made, how you've solved them, and how you fixed the thought process that lead to them. When one makes that mistake, you correct it and share the instance of when you made that mistake, and how you went about solving it. People learn way faster and the whole experience is more agreeable. It also instills intellectual honesty and once it's there, it's embedded in the team. It avoids un-necessary frictions and ego trips.
Being a gate keeper when it comes to hiring. The standard is not hiring someone I wouldn't accept as my team lead/CTO/CEO/COO some day. We hire people with the possibility in mind that they might run the company some day, and thus we're careful. The people we hire will lead their own teams in the near future. We weed out behaviorally flawed people early on: dishonest technically, not taking responsibility for things, changing facts as we wouldn't trust them, and we can't really work with people we don't trust. Example: if something breaks, I don't want someone to cover their tracks and waste us precious time we could be spending fixing what broke.
Involve the team in decision making by sharing all the information you're using to reach for conclusions. We have git repositories with our meeting notes with prospects, investors, advisors, organizations. Dissemniation of financial and business information. People who weren't there read them. Even a fresh graduate can read them and learn things, ask questions, give feedback. It's a great way to discover new ways of doing things when you confront people to problems.
All in all, work systematically to make whatever put you in a certain position to become a given at a company, or at least disseminate it so you're not the only one who either can do it, or has a blueprint to how to do it.