Anthropic just dropped Opus 4.6 and introduced Teams of Agents in beta.
Whether you realize it or not, you essentially just got a promotion.
If you were an IC, you are now a lead: managing a team of agents. If you were already a lead or manager, congratulations; you are now a leader-of-leaders.
Multiple agents working in parallel, each owning a piece of the work and coordinating autonomously. One handles the frontend; one owns the API; one tackles the migration. This should sound familiar; that is how high-performing engineering teams already operate.
However, here is what most people are missing: the skill that matters most is leadership.
The feedback loop you use with agents is the same one good leaders use with people:
Define the goal. Let them execute. Retrospective. Incorporate learnings. Repeat.
You must understand where the agent struggled and then refine your approach so it can operate more autonomously next time. You’re no longer “prompt engineering,” you’re “coaching.”
Good leaders will thrive here. They already create feedback loops for their people. They adjust their approach to the specific needs of each team member. They understand that delegation is about clarity, trust, and iteration.
Micromanagers will struggle. I’ve watched users get frustrated that an agent doesn’t do exactly what they envisioned in a perfectly deterministic way. They want a crystal ball, not a collaborator. People can’t read minds, and neither can an agent.
ICs face a different challenge, especially those early in their careers. They have never been trained to delegate, instruct with clarity, or think strategically about how work flows through a system. Interestingly, Forrester’s research shows Gen Z workers have the highest AI readiness scores of any generation. They are natural with the tools; what they lack is the leadership framework to wield them effectively.
So…
To the micromanagers: learn to coach; not just your agents, your people.
To the ICs: find leaders you admire and ask them to mentor you. Learn to delegate, communicate intent, and think about outcomes instead of outputs. There is an old saying that there is no compression algorithm for experience; however, AI agents give you the playground to accumulate it faster than any generation before you. Use it.
History has a clear message. MIT research shows roughly 60% of jobs in 2018 didn’t exist in 1940. Spreadsheets didn’t eliminate accountants. They eliminated manual calculation. In doing so, they turned accountants into analysts and advisors.
Every major technological shift follows the same pattern: a breath, then a rush.
We are in that breath right now. The wave is coming; the leaders who use this window to find raw talent and develop it will be the ones who come out ahead.
Thus, the question is whether we have the leadership skills to make this reshaping work in our favor.
Originally shared on LinkedIn, February 6, 2026.