Brian Armstrong sent a tweet yesterday to announce his decision to lay off 14% of the people that work at Coinbase.
He claimed two main reasons, the first being that Coinbase isn't doing as well as he thinks it should in the market, and the second being AI.
To back up that second point, he mentioned that
Non-technical teams are now shipping production code and many of our workflows are being automated.
He followed up with some thoughts specifically around management, saying
The future is small, high context teams that can move quickly. Leaders will own much more, with as many as 15+ direct reports.
and
No pure managers: Every leader at Coinbase must also be a strong and active individual contributor. Managers should be like player-coaches, getting their hands dirty alongside their teams.
What is a Manager's Role in a world rapidly adopting AI?
Brian's understanding of a manager's role in a world rapidly adopting AI is fundamentally flawed. When AI can automate many tasks, the role of a manager becomes even more crucial in providing human oversight, strategic direction, and emotional intelligence. This need is only amplified when you consider the impact of AI-driven layoffs on employee morale at a company like Coinbase.
15+ Direct Reports doesn't work
For a period in my management career, I had nearly 15 direct reports. It was a nightmare.
At that number, I felt like I was their manager in name only. I had limited time to provide meaningful support, mentorship, or guidance to any of those I managed. I was constantly putting out fires and reacting to issues instead of proactively leading my team. Quarterly and annual reviews felt superficial because I couldn't give each person the attention they deserved throughout the quarter. My direct reports were empathetic to the situation, but it was clear I couldn't be the leader they needed during that time. This led to wider issues for my role like not having capacity to effectively manage up, advocate for my team, or provide strategic input to the company. This, in turn, led to business continuity risks like unrealistic timelines being committed to, an increase of incidents for the platform, and a general decline in team morale. Keep in mind, this was all before AI adoption exploded.
At 15+ direct reports you're not a manager, you're a coordinator. You're coordinating, at a high level, the work of your team, but you're not providing the leadership and support that they need to thrive. Your direct reports will rarely have your full attention, and they will likely feel neglected and unsupported. You won't be able to provide the mentorship and guidance that they need to grow and develop in their careers. It's unsustainable and will lead to high turnover and low morale long-term.
On Player-Coaches
I can understand the appeal company leaders see in everyone, even outside of Engineering Managers, being able to build things while juggling their existing daily duties. To leaders, they see promise of faster feature and product releases. Faster releases mean more sales opportunities. More sales opportunities mean more revenue and market share.
Personally, I act as a player-coach from time to time already, but I don't think it's a good idea for everyone. It's a tightrope to walk, and it can lead to mistrust in an organization if not done with care.
- If a manager is too focused on being a player, they may neglect their managerial duties. The team will suffer.
- If they're too focused on being a coach, they may neglect their individual contributions. The team will suffer.
- If they've taken on individual contributor work that's a dependency for their team in some way, they're gambling with the success of their team for the sake of their own ego or the ego of those demanding it.
The only way Player-Coaches can work is if the company has a strong culture of trust and accountability, and if the manager is able to balance their dual roles effectively. Meaning they don't have 15+ direct reports, and the IC work they take on is not critical-path for their team.
Conclusion
Brian's vision for management at Coinbase is misguided and will likely lead to a toxic work environment for those 86% that remain as of writing this post.
The issue at hand for the wider industry is that Brian's ideas are not unique. Leaders across the tech industry are faced with deciding how to adapt to the rapid adoption of AI, and many are likely to make similar mistakes. It's crucial for leaders to understand the evolving role of managers in this new landscape and to prioritize the well-being and development of their teams over short-term profit.
AI isn't going anywhere. The leaders worth following will be the ones who use it to amplify their teams without losing sight of a basic truth: people need other people to thrive. Pretend otherwise and you'll hurt both the people you lead and the work you're trying to build together.