Distilling yourself
What principles and values do you actually operate with?
Thanks for being here for the May subscriber edition, and for your ongoing support.
Most experienced leaders I’ve worked with operate on principles they can’t quite articulate. They make consistent decisions, give consistent advice, and have strong instincts about what’s right, but if you asked them to write down the ten things they actually believe, they’d struggle. The principles are implicit, revealed through behaviour rather than stated outright.
This article is about making them explicit, using the same distillation technique from this month’s free article, but pointed at yourself. Because once your principles are visible, you can check decisions against them, share them with your team, and notice when your thinking has drifted without you realising. You might even surface beliefs you didn’t know you held.
The end product is a document: a tiered list of ten to twenty-five personal leadership principles, extracted from your own writing by an LLM, organised by strength of conviction, and mapped to the tactical frameworks you already use.
But it’s not something you file away and revisit occasionally: the real power comes from dropping it into your AI assistant’s context, where it becomes a living part of your workflow. Every decision you draft, every strategy you review, every piece of writing you produce can be checked against what you’ve said you believe, automatically and on demand.
Here’s what we’ll cover:
Why bother? How your principles can drift without you noticing, and why making them explicit changes that.
The raw material. What counts as your body of work, and why writing is uniquely powerful for this kind of extraction.
The extraction process. A step-by-step walkthrough, including the actual prompts I used.
What I learned. The results of running this on my own archive of two years of newsletter articles.
Principles as a daily driver. How to integrate the output into decisions and daily workflows.
Keeping principles alive. Why they go stale, and how to maintain them.
The conversational alternative. A different path for those who don’t have an archive to work with yet.
If you’d like to dig deeper, here are some related articles from the archive:
Gather, decide, execute describes the kind of second-brain system that naturally generates the archive this process could feed from.
Invert, always invert explores how principles help you know what not to do.
Heartbeats: keeping strategies alive covers why frameworks need maintenance cycles, which applies to principles too.
A weekly mind meld shows how sharing your thinking at scale builds alignment, which a principles document enables.
Let’s get into it.
Why bother?
So, what’s the risk of leaving these principles unspoken? Because your team doesn’t just observe your principles, they absorb them. The decisions you make, the things you reward, and the behaviour you tolerate become the template for how everyone around you operates.
Guy Kawasaki, Apple’s former chief evangelist, once described Apple as “Steve Jobs with ten thousand lives” in Leander Kahney’s Inside Steve’s Brain: Jobs had imprinted his principles so deeply that the company became a mirror of its founder. The same dynamic plays out at every level, because every leader shapes the culture around them, whether they’re doing it deliberately or not.
The problem is that consistency drifts, often without you noticing. You don’t wake up one morning having abandoned a principle: it happens gradually, one exception at a time, until what you’re doing no longer matches what you think you believe.
There’s an important distinction here between evolution and drift. Evolution is when you consciously revise a principle because your thinking has genuinely changed: you’ve seen new evidence and arrived at a different conclusion. Drift is when you contradict yourself without realising, because you’ve simply forgotten what you used to believe.
When I ran this process on my own writing, one of the surprises was a principle I’d call scepticism, not cynicism: the idea that new technologies deserve critical evaluation, but blanket dismissal is intellectually lazy. I’d been applying that distinction without ever articulating it, which meant it was invisible to the people around me.
The first step is figuring out where those principles are hiding, which brings us to the raw material.
The raw material
Where do your principles actually manifest? They’re embedded in the things you’ve written, the decisions you’ve made, and the advice you’ve given, scattered across long-form documents, emails, proposals, and messages.
All of it is an externalisation of your thinking process, and that’s what makes it useful: your body of work is the raw material for this extraction, and the richer and more varied it is, the clearer the picture you’ll get.
Not all sources are equal, though. Writing is uniquely powerful for this kind of extraction, because writing is thinking: it forces a clarity that speaking rarely does, and it leaves a permanent record of what you actually believed at the time, not a reconstruction after the fact.
Larry McEnerney’s famous lecture on writing from the University of Chicago captures this perfectly: you don’t think first and then write, the process of writing is the thinking. The clarity comes afterwards, when you revise what you’ve produced into something the reader can follow. That’s why your writing is such a rich source of principles: it caught your thinking in the act.
Writing also tends to represent your best self, which makes it especially useful here. Whether it’s an internal memo or a public article, you probably took time to think through your position, choose your words carefully, and present a considered view. The messy day-to-day pulls you away from that considered version, but the writing preserves it, which makes it useful as a source of raw material.
So, what’s the minimum viable corpus? Ideally, two or more years of documents across different contexts: technical decisions, people management, strategy. But you can get a useful extraction from as little as six months of significant writing, whether that’s internal strategy memos, promotion cases, project retrospectives, decision records, or even detailed interview feedback you’ve written for hiring panels.
One thing to keep in mind is recency weighting: your most recent writing reflects your current thinking, while older material shows where you came from. Both are useful, but they serve different purposes. When I was thinking about doing this, I found that two years was a good cutoff for work that still felt present to me, both in terms of the articles I’d written on here and the capacity I’d been operating in at work.
If you don’t have much of an archive yet, don’t worry: there’s an alternative approach later in this article. But if you do have the material, let’s walk through how to turn it into something useful.
The extraction process
You’ve got a body of work: documents, articles, decision records, emails. It’s sitting in various folders and inboxes, and somewhere inside it are the principles that actually guide your behaviour. The question is how to get them out. What follows is the process I used on my own newsletter archive, but it works on any collection of writing where you’ve been thinking out loud: for example, weekly journal posts, weekly brag docs, or regular writings you’ve shared with your team.
Gather and organise
Since the LLM is going to do the heavy lifting, the format of your documents doesn’t matter much, though plain text works well. The first step is to get everything into one place: export your documents into a folder you can navigate, and organise them by date so you can weight recent material more heavily.
You don’t need every email you’ve ever sent, just the ones where you were genuinely thinking through a problem, making a case, or articulating a position. Strategy memos, decision records, proposals or RFCs you’ve written, project retrospectives, internal blog posts: anything where you were doing real thinking and having your leadership embodied in your writing, not just relaying information.
First-pass extraction
Point the LLM at the folder and give it a prompt that explains what you’re trying to achieve. Here’s the one I used:
