Last week I wrote about Good Friction.
The idea was basically this: a good life is not a frictionless life.
Some friction is obviously bad and should be removed. But some friction is part of what makes life meaningful. It builds discipline, taste, relationships, craft, health, attention, and maybe even character, if we want to risk sounding like a my grade 6 teacher.
Last week, that idea became more practical for me.
I had a conversation with a reporter, Michael Tippett, and Steve Pratt.
The reporter was profiling Michael, and I joined because he was interested in how we collaborate together and are using AI in our work. We ended up covering a lot of ground. The future of work, agents, software, strategy, where this is all going, the usual light cocktail of existential questions for a Friday afternoon.
But one idea stuck with me.
Everyone is going to need their own rules for AI.
Not just company policies. Not generic “AI best practices.” I mean personal rules.
What do you give to AI?
And what do you keep for yourself?
I think that is becoming one of the more important questions now. Not because AI is bad. I use it constantly. Probably more than most people I know.
But because the more useful it gets, the easier it becomes to hand over parts of the process without noticing what we are losing.
Steve made a point that really landed for me. He identifies as a writer. That is the work. That is his craft. So for him, there are parts of writing he will not ever give away. Maybe AI can help edit, push, challenge, or organize. But if the actual act of writing is part of how he makes sense of the world, why would he outsource that?
That feels obvious when you say it out loud.
But I am not sure most of us have said it out loud yet.
I certainly have not said it clearly enough for myself.
For me, AI is incredibly useful once I have a mess.
A voice note. A half-formed thought. A transcript. A conversation. A pile of ideas I cannot quite organize. A phrase that keeps bothering me because I know there is something in it.
That is where AI helps. It can reflect things back. It can help me structure the mess. It can give me language to argue with.
And honestly, that last part is probably the most useful.
It gives me something to disagree with.
Which, as it turns out, is one of my primary creative processes.
But I am becoming more aware that there are parts of the process I do not want to skip.
I do not want AI to do the noticing for me.
I do not want it to decide what I care about.
I do not want it to make everything sound clean and resolved when the whole point is that I am still working it out.
I do not want it to remove all the awkwardness, because sometimes the awkwardness is where the honest part is hiding.
That is where the Good Friction idea matters.
Once you accept that not all friction is bad, the question becomes more specific:
Which friction helps me become better?
And which friction is just getting in the way?
For me, AI is great at removing the friction that drains energy before the real work begins.
The blank page.
The first structure.
The repetitive task.
The initial research.
The messy transcript.
The pile of notes and ideas I do not want to organize.
The “there is something here but I cannot quite see it yet” stage.
That kind of help can be incredibly valuable.
But there are other parts I want to protect.
The noticing.
The caring.
The judgment.
The lived experience.
The slightly weird turn of phrase that is memorable.
The part where I realize I do not actually believe the thing I just wrote.
The part where I sit with an idea long enough for it to become mine.
That last one feels important.
Because AI is very good at giving us language for things quickly. But I am not sure quick language is always the same as understanding.
Sometimes we need the slower process because the process is how the understanding happens.
That is especially true if you are trying to make something that is not just derivative.
And this is where I think AI gets a little tricky.
If everyone is using the same tools trained on the same internet to produce the same polished outputs, then the more valuable thing becomes whatever is not easily produced by the tool.
Your specific experience.
Your intuition.
Your taste.
Your judgment.
Your relationships.
Your humour.
Your scars.
Your obsessions.
Your actual voice.
The stuff that is harder to automate because it comes from having lived a particular life.
This is part of what I am trying to figure out with The Calm Society.
I do not want to use AI to become a faster, smoother, more efficient content machine. I have spent enough of my career around dashboards to know that becoming a machine is not my highest aspiration.
I want to use AI to help me think better.
To notice patterns.
To test ideas.
To organize material.
To sharpen arguments.
To find better questions.
To get to the human conversation faster.
But I do not want it to replace the human conversation.
I do not want it to replace the living.
I do not want it to replace the part where I go for a walk, have a conversation, read something that annoys me, sit in the garden, overthink a phrase for 45 minutes, and then realize what I actually meant while making coffee.
That may not be efficient.
But it is often where the work is.
So I am starting to think about my own AI principles.
They are not fully formed, but they look something like this:
Use AI to organize my thinking, not replace my thinking.
Use it to challenge my ideas, not decide what I believe.
Use it to help with structure, but not flatten my voice.
Use it to remove the friction that drains energy, not the friction that creates meaning.
Use it to get to better human conversations faster, not avoid them.
The important thing is that these rules will probably be different for everyone.
If you are a writer, you may protect more of the writing.
If you are a designer, maybe you protect the taste-making.
If you are a founder, maybe you protect the judgment calls.
If you are a strategist, maybe you protect the wrestling with ambiguity.
If you are a parent, maybe you protect the conversations where nobody knows what the right answer is and you just have to stay in it.
That is not very scalable, I realize.
But neither are most of the things that make life meaningful.
This also matters for people building software.
The default product question is usually:
How do we make this faster, easier, smoother - frictionless?
That is often the right question.
But with AI, I think we need more questions:
What human capacity are we strengthening or weakening?
Are we helping people develop judgment, or replacing it?
Are we helping people express themselves, or making everyone sound the same?
Are we creating more space in people’s lives, or just raising the expectation that they produce more?
Are we helping humans become more human, or just better at behaving like machines?
That last question is the one I keep coming back to.
AI is going to remove a lot of friction.
The challenge is deciding which friction was never serving us, and which friction was quietly helping us become who we are.
That feels like a personal question.
And a creative question.
And, increasingly, a builder question.
So I am curious.
What are you happy to give to AI?
And what part of the process do you want to keep for yourself?
Cameron