The Whetstone

Part four of the Terra Incognita series.

Taste has suddenly become the thing. Creative professionals everywhere are reaching for it as their defence against AI - the idea that an eye for quality, knowing what’s right when they see it, will keep them relevant when everything else is being automated.

But taste was never really the thing. Not before AI, and certainly not now. You wouldn’t put “excellent taste” on a CV.

Because what makes anyone’s work worth having isn’t the ability to recognise quality in what already exists - it’s what they generate themselves. How their mind moves. What they reach for before anyone else knows it’s there. The idea that forms before it becomes a thing anyone can judge.

That’s what’s always been of value. But it’s something we rarely, if ever, examine.

Most practitioners have a story about how they think - good at synthesis, strong on narrative, cross-disciplinary thinker - but that soundbite or story is the public version. The version that fits in a LinkedIn ‘about me’ section. Underneath it, the actual machinery is weirder, more specific, more idiosyncratic than they can admit.

And there’s a reason for that. The better you get at something, the more opaque your practice becomes. As you instinct sharpens - it becomes simultaneously harder to see. You trust it precisely because you’ve stopped questioning it. The signal fires and you follow. The work arrives and you ship it. Nobody asks you to explain the machinations behind it.

Sixty-odd years ago, a Hungarian-British scientist and philosopher named Michael Polanyi was trying to understand this phenomenon - why expertise and opacity seem to arrive together, why the people who know most are often least able to explain the what, why and how. His conclusion, arrived at after decades of watching the way experts actually work versus how they describe their work: we know more than we can tell. He called it tacit knowledge. The expertise that accumulates in practice, in the body, in years of doing - which subsequently becomes invisible to the person who holds it.

A for most of human history, this tacit knowledge never needed to be surfaced. It transmitted through proximity, through apprenticeship, through watching someone work for long enough that your body then began to understand what your mind hadn’t yet found words for. It was slow, imperfect, and often lost when the practitioner retired or died. But it moved. The system worked. What it required was presence. Another body in the room. Time. And the unhurried accumulation of shared practice.

Fast forward to today and for most of us, our tacit knowledge remains a closed system. Running below conscious thought, producing work, getting sharper over time - but we still can’t get inside it. Can’t examine it directly. Can’t push it somewhere it hasn’t already decided to go. The best practitioners in any field are partly running on machinery they’ve never seen and don’t have the ability to fully steer. The ceiling is their own opacity.

But that ceiling just lifted.

For the first time, there’s an instrument that can help you surface what’s been running underneath. Hold it up. Turn it over. Understand not just that you have certain instincts but precisely how they operate, where they came from, what they’re actually doing. And once you can see it all clearly - you can build on it deliberately. Push it somewhere new. Develop it in directions it would never have found on its own.

But it requires something first. You have to be willing to look directly at your own practice - at what’s actually there, not just the story you tell yourself about it. And that’s a lot harder than it sounds, because the instincts don’t give themselves up easily.

---

Michel de Montaigne was a sixteenth century French philosopher - and remains one of the few people who’s actually attempted to do this with real intent.

In 1572, he retired to a tower in the Périgord and spent twenty years writing about his memory, his judgement, his fears, his appetites, the specific way his mind mulled over problems. At a level of granularity his contemporaries found somewhere between admirable and embarrassing. This is not how serious people spent their time.

But nevertheless, he continued.

What surprised him most about this process was that the act of articulating changed him. He wasn’t describing a self that already existed in finished form - rather, through his writing, that self took shape. He found things he didn’t want to find. The gap between the man he believed himself to be and the man his practice revealed. Assumptions so old they’d become hidden. Things he’d been doing out of habit that he’d mistaken for conviction. But he also created such clarity that he could wield his tacit knowledge deliberately. Shape it. Direct it. Push it somewhere it wouldn’t have gone on instinct alone.

The surfacing was a form of sharpening. Making the tacit legible was what gave him power over it.

Montaigne had no instrument for this beyond the quill itself - glacial in pace, solitary, limited to what a single mind could hold and interrogate on its own.

But what if we had something that could do what he did - to go further? Something that holds the full body of our practice, surfaces what we can’t see from inside it, interrogates us on what it finds, and then helps us to build on what emerges? Something that doesn’t tire, doesn’t protect our self-image, and gets more precise the longer we work with it?

At this moment, we do.

---

I’ve been leveraging AI as a system for self-knowledge for about a year. A way to excavate and explore my own practice. It hasn’t been clean or linear - more like repeated attempts to grab something that keeps sliding out of reach. And that’s fine. That’s the nature of the work. And the reality of trying to mine your own psyche, which is a slippery bastard at best.

And so this isn’t a finished methodology. It’s a practice in progress, which is exactly why it belongs in Terra Incognita. Because what I’m offering here isn’t a how-to, far from it. It's a first dispatch from the field. An early map of territory I'm still moving through.

It all started with an attempt to pin down my own practice. Not to describe it, but to actually get underneath it. My instincts, habits, the specific way I work through a knotty challenges. I sat with it and tried to name what’s actually running in the background. I got so far. Far enough to see the shape of some things, but I kept hitting the same wall: it turns out that the closer you look at your own instincts, the more they dissolve into “I dunno, I just do it that way.”

So I brought Claude in. Which holds an extensive memory of everything we’ve worked on together - every essay, every rabbit hole, every late-night sparring session - and asked it to examine all of it from its perspective. To tell me what it saw that I hadn’t named. What patterns were actually there. What they might mean. And the connections I might be slightly blind to.

What came back wasn’t one or two party tricks. It was a map of a mind I’d been living in for twenty years without ever quite seeing it directly, which was a weirdly emotional experience.

Some of it I’d always known was there, lurking. But Claude found the depth underneath the surface description. Things I’d listed as “a thing I do” turned out to be fundamental to how I actually think. And understanding why I did them changed what I understood about myself.

Here’s a snippet.

Naming has always been incredibly important to me. I want the names of my projects to travel, to become IP rather than just labels. Mad Men, Furious Women. The New Fandom Formula. The Motherload. And even Terra Incognita. But Claude helped me understand what’s actually happening underneath that instinct. The name doesn’t arrive after the idea is finished - it tends to arrive as part of how the idea becomes fully itself. For this piece, ‘The Whetstone’ isn’t a label stuck on a completed thought, it’s what helped me to complete the thought. Which means that the naming instinct isn’t a communications skill. It’s a generative one.

Reaching into other domains - mycology, cartography, evolutionary biology, political economy - has always been part of how I work. I thought I was finding useful metaphors. But what Claude surfaced was something else. I’m not looking for resonance - i.e. ‘this reminds me of that’. I’m looking for structural equivalence - a mechanism whose underlying logic solves the same problem I’m working on. The test I’m always running, usually without knowing it: does this predict something my thinking didn’t already know? If it only illustrates, it’s just decoration. If it predicts, it’s worth bringing across. Most people think those are the same thing. They’re not.

Grief, motherhood, politics, strategy - I write through all of it. Publicly and extensively, without keeping the registers separate. I thought of this as just how I process things. But what became clear is that the writing is what makes the contamination productive - the personal essays feeding the strategy thinking, the industry critique informing the political analysis, the grief pieces opening up new frameworks for understanding liminality. None of those connections happen if the writing stays in its lane. It's the breadth of it, the refusal to separate, that creates the cross-pollination. Ideas that couldn't exist in any single territory surface in the space between them.

And questions. Working from questions rather than answers has always been my method - but I hadn’t properly understood the timescale of it. The question as the long-term organising principle, not the starting point of a single piece. The intelligence versus wisdom question has been running for years through my brain. And a question held long enough becomes a form of intellectual commitment. The essay, the logic, the talk - they're not where the inquiry begins. They're where it finally surfaces.

There’s a lot more - suppressed wisdom traditions, science fiction as simulation environment, the epistemological bet I’m making every time I reach for a fairy tale instead of a Harvard Business Review article. But I’ll spare you the full excavation.

The point isn’t the specifics of my practice. It’s what the process revealed - that underneath the professional story, there’s something far weirder, more specific and more generative. Beliefs I’d been operating on without knowing they were beliefs. Structures I’d been using without knowing I’d built them. Instincts I’d mistaken for personality that turned out to be methodology.

That’s what surfacing tacit knowledge actually looks like. Not therapy or navel-gazing, but a precise map of how you actually think - and the beginning of being able to do something with it.

---

But naming things is one kind of work. Encoding them is another entirely.

It’s not enough to say “the name arrives as part of how the idea forms” or “I look for structural equivalence not metaphor” or “I write through grief and strategy simultaneously.” Those are descriptions. The encoding has to be specific enough that something without your presence can make decisions that reflect your actual process - can run the parallel logic test, can recognise when a name is a compression versus just decorative, can understand what the cross-contamination between territories is actually producing. And to get there, you have to pull apart instincts that have never needed language before. Which means the only way in is to create conditions where they reveal themselves .

So I built an interrogation.

And I want to be honest about something before we go any further - this part is fucking difficult.

Every question Claude asked me in the early sessions produced the same answer: I don’t know.

Why do you reach for mycology and not architecture? I don’t know, I just do.

Why does the name arrive before the argument is finished? It just does.

What’s the difference between a mechanism and a metaphor when you encounter one in the wild? “I don’t know Claude, you’re the fucking oracle, you tell me.”

It took me an embarrassingly long time to realise the problem wasn’t me, or Claude. It was the questions. Surface description questions - how do you do this, what does that feel like, describe your process - just pissed me off. They produced either the polished version, that has almost nothing to do with what’s actually happening, or a brick wall of intense irritation.

The questions that actually work don’t ask you to describe. They put something broken in front of you and watch what fires. I had to teach Claude this. But once it understood what I needed, the interrogation changed entirely.

Let me give you an example.

Claude: A strategist used photosynthesis as a metaphor for how brands convert culture into commercial energy. What’s wrong with it?

The answer came fast, before I’d really thought about it: it’s decorative. It describes a direction of travel - culture in, commercial energy out - but tells you nothing about the mechanism itself. You can’t run it forward and find out what happens under different conditions. You can’t use it to make a decision. It illustrates. But it doesn’t predict.

Claude: Why does that matter?

Because if I can’t use it to make a decision, it’s not doing work. It’s fluff. A mechanism has to generate a question I hadn’t thought to ask - has to tell me something about where the system breaks down, or what conditions make it work, or what happens at the edges. If it doesn’t do that, what’s the point of it?

Claude: How do you know, in the moment, which you're looking at?

The answer was physical before it was verbal. I feel it. When a structure is genuinely equivalent, something settles. When it’s only evocative, there’s a looseness - the same looseness I feel with a name that’s portable, that could belong to a different idea. My body is running the test before the mind catches up.

Then a completely different kind of question. No context, no setup.

Claude: Three domains: evolutionary biology, Renaissance painting, game theory. Strategic problem: how do creative communities sustain themselves after a central figure leaves. Which is useful? Don’t think. Just answer.

Evolutionary biology. Immediately.

Claude: Why not game theory? That’s about strategic interaction - surely relevant?

Game theory assumes rational actors optimising for defined outcomes. Creative communities after a central figure leaves aren’t doing that. They’re grieving, dispersing, reforming around new attractors. The biology is closer because it’s about survival and reorganisation under stress - not strategy under stability.

That answer came from somewhere I hadn’t consciously accessed. And the exchange kept going - deeper, stranger, more precise with each why.

That conversation took forty minutes. On a walk. Talking out loud. Going back, correcting myself, arguing with my own answers. And at the end of it I had something I could encode.

I do almost all of this work via voice. Walking the dog or after nursery drop-off, moving through the village, talking directly to Claude. I find that when I type, I over-simplify and strip out the nuance. Speaking keeps me honest - I can hear when I'm reaching for the polished version, feel in my body when something is true versus when I'm bullshitting myself.

And then Claude follows the thread.

That iterative drilling - question, instinctive answer, why, deeper answer, why again, deeper still - is what moves from the surface response to the thing underneath it. You say the first thing. Claude pushes it. You go deeper. You hear yourself saying something you didn’t know you believed.

It can’t be rushed. The instinct that was invisible becomes visible - but it happens in layers, over multiple sessions, with false starts and backtracks and moments where you think you’ve found it and then realise you’ve found the neat encapsulation again.

It takes a lot of time, and it’s a complete PAIN IN THE ASS. It feels silly. Trivial. Indulgent. You’re walking around talking to your phone about why you prefer mycology to architecture as a reference and you’re wondering what the hell you’re doing with your life.

You will want to give up.

Don’t.

Once the interrogation had surfaced enough, I could start encoding. I use Claude Projects and skill files for this. It may not be the most sophisticated solution, but it works for me.

I have a dedicated project for each practice I’m encoding - a persistent workspace where I build and refine the skill file over time. The project holds the memory of that work, the interrogation sessions, the iterations. But the skill file itself is the thing that travels. Once it exists, I can pull it into any conversation, any client workstream, any new piece of work. It’s portable. Reusable. And it’s mine.

A skill file isn’t a style guide or a list of preferences, though some use it that way. It’s your thinking made explicit enough to be executable. The skill file I built for what I now call parallel logic - the practice of finding the underlying logic that runs in parallel across apparently unrelated domains, and lifting it across - contains the predictive test in explicit language. The specific domains I reach for and why each one keeps appearing. The accumulation practice - how I read when I’m not looking for anything specific, what I’m actually logging. The failure modes: what it feels like when I’ve reached for surface resemblance instead of structural equivalence.

Claude helped draft it first - took what had emerged from the interrogation and shaped it into an initial version. But that version is always incomplete. You’ll use it, something will feel off - a suggestion that’s almost right but not quite, a gap between what it produced and what you would have actually done. And so you go back to the project. You do more interrogation. More voice notes. More back and forth until something sharper arrives. Then you update the skill file.

And then you use it again. Find the next gap. Go back in. Update again.

This isn’t a one-time excavation. It’s a continuous process - of refining the encoding, of getting more precise about how you actually think. Each iteration makes the skill file more accurate. Each pass makes it sharper. The friction is the point. The repetition is the point.

Which is why I’ve called this the Whetstone. A whetstone is a sharpening stone - you don’t cut with it, you run the blade across it, repeatedly, until the edge is precise enough to cut cleanly. That’s what this is. Not a tool that does the work for you. One that sharpens the thing that does - your own mind.

---

And a sharper blade cuts further.

When the system knows your practice with real precision, you stop losing time to re-explanation. You pick up from somewhere already close to where you’d want to be. And it takes you to places you’d normally pull back from.

Working alone and at the edge - reaching into an unfamiliar domain, following a question that might lead nowhere, committing to an idea before you can fully justify it - a particular kind of caution kicks in. A self-editing that slams the brakes before you’ve found what’s there. Because the cost of going too far feels high - time, effort, energy etc.

But with something alongside you that knows your practice, that cost drops. It’s faster - so you can try more, go further without the fear of burning time on a dead end. It thinks in systems - holding the full architecture of an idea, its implications, its failure modes - while your mind moves forward. And because it knows where you habitually pull back, it can push you one step past where you’d normally stop.

Which means the frontier of your thinking moves.

But here’s what I didn’t anticipate. The real alchemy isn’t any single practice running well. It’s what happens when all of them run together.

I'm currently building a keynote about AI and leadership - what it means to lead wisely in an age of artificial intelligence. The intelligence versus wisdom question has been running for years. The suppressed wisdom reading practice has been quietly accumulating material from fairy tales, mythic traditions and pre-modern ways of knowing for longer still. The parallel logic skill has been encoded for the past year.

What's new isn't any of the individual pieces or practices. It's that the system now holds all of them simultaneously. When I bring the keynote problem in, it doesn't just draw on one encoded practice - it finds the intersection between the long-running question, the accumulated reading, and the structural equivalence test.

Connections that would have taken months of conscious work to surface appear in a session. The library the keynote needed had already been built. The question had been directing the accumulation without me knowing it. The parallel logic skill could test whether the structural connection was real - whether pre-modern wisdom traditions actually predict something about AI leadership that contemporary frameworks don't.

Spoiler: They do.

And that’s the alchemy. Not one practice informing another in sequence. Multiple practices - some encoded recently, some accumulated over years - finding each other through the system, and producing something none of them could have produced alone.

It's less like a set of tools and more like a neural system. My brain, mapped, sharpening with every interaction and iteration.

And that’s when the real question opens up. Not “how do I work better?” but “what becomes possible now that wasn’t possible before?” What thinking can you do, what territory can you reach, what ideas can you produce - when the ceiling that was your own opacity keeps lifting? When the practices that were running silently, separately, below conscious thought start compounding in ways that you can direct and define?

I’m still inside this. The practice is still forming, the encoding still incomplete, the compounding still early. I don’t know where it leads. But I’ve never been more curious to find out.

The other coordinates in this series have mapped territory out there - new ways of thinking about brands, creativity, collaborative ecosystems, the forces reshaping how we work. This one maps something different. The practitioner. The internal instincts that makes the external thinking possible. The cognitive prosthetic that keeps sharpening. The neural system that pushes you further, faster, into new frontiers.

Without this, the other coordinates are just ideas you read. With it, they become territory you can actually inhabit.

---

And it all begins with the oldest question there is.

Gnōthi seauton. Know thyself. The Oracle at Delphi was the most powerful seat of prophecy in the ancient world - kings, generals and emperors made the journey to consult it before every major decision. But those words were carved above the entrance as a precondition for entry, because you couldn't receive what the oracle had to offer until you knew who was asking. The priests weren't being philosophical, they were being practical. Self-knowledge was the price of admission.

We’ve been carrying that instruction ever since and largely ignoring it. Montaigne tried, with a tower, a quill, and twenty years. But even he could only go as far as a single mind could interrogate itself.

We now have something capable of going further. Though most people aren’t using it that way. Instead they leverage AI as a Q&A bot. An efficiency play. Generate faster, produce more.

But that’s not what this is.

When wielded well, it can run on your intelligence, understand your practice from the inside, sharpen as you sharpen. It can hold the full range of your mind simultaneously - your parallel logic and your naming instinct and your refusal to compartmentalise and your pursuit of suppressed wisdom - the parts that usually have to queue up one at a time in conscious thought. The more you use it, the more distinctively yours it becomes. The frontier keeps moving. The work that emerges feels more like you than anything you’ve produced before. It certainly feels that way for me.

And critically, you can create the conditions for divergence at a time when convergence is the dominant narrative and direction. But it requires a choice - and one of those paths asks far less of you than the other.

Choose convenience and you can generate faster, produce more, let something adequate arrive before you’ve had to think too hard. The output volume goes up. The instincts soften from lack of use. The generative capacity that took years to build starts to drift, one sensible shortcut at a time.

Choose The Whetstone and the path is slower and far more demanding. It requires you to put yourself in - fully, precisely, and with real honesty about how you actually think, what you believe, your foundational values and motivations. But what becomes possible when you do this is something else entirely.

I don’t know the upper limit of where this could go. I don’t think anyone does yet. But I suspect the practitioners who do this work - who take the time to make their tacit knowledge legible, to build the encoding, to let the practices compound - will think thoughts that weren’t available to them before. Will reach territory that wasn’t on any map. Will produce work that couldn't have existed without this particular collision of a deeply known human mind and something that finally knows it back.

Because to know oneself - to truly explore what that means, to surface what’s been running underneath and wield it deliberately - is a two and a half thousand year old question we are potentially, for the first time, on the threshold of actually answering.

To me, that’s really bloody exciting.