It's imper-AI-tive
We've established the clear benefits of declarative rather than imperative leadership.
There are exceptions, such as when working with juniors or in times of crises, but even here, there is nuance, and these exceptions can often end up proving the rule.
But I want to offer another way of thinking about imperative vs declarative that takes on more importance each day: working with AI.
I'm not talking about working directly with LLMs here - this is another case where often declarative wins (especially with reasoning models).
Rather, how do we get our teams to work effectively with AI?
In conversations and what I see on the ground in my fractional CTO role, there is still a lot of resistance and fear when it comes to engineers leveraging AI.
Often, this is precisely because of overly declarative statements from senior leadership - "we must be an AI-first company", "I want X% of code to be written by AI" - without sufficient grounding in the realities of AI tooling and capabilities.
Engineers then dismiss AI as a new crypto fad, or something that is being forced on them to put them out of a job, or both, and are thus understandably reluctant to experiment.
So, what does being imperative with your team look like when it comes to using AI?
We'll wrap up this mini-series tomorrow.