The wisdom to know which battles deserve your depth
Wednesday Essay | 14 January 2026. The Attention Problem in Partnerships
I’ve been thinking about resistance.
Not the productive kind—the creative friction that sharpens ideas—but the other kind. The resistance that shows up when a partnership threatens to go deeper than either party intended.
Some clients don’t want embedding. They want advice at arm’s length. Deliverables they can review, accept or reject, file away. Clean boundaries between “consultant” and “client,” between “your problem” and “ours.”
At first I read this as preference. Some organisations genuinely work better with external input kept external. Fair enough.
But watching Amelie lead a climb this Christmas—watching her fall, catch herself on gear, want to stop—I started seeing it differently. Sometimes resistance isn’t preference. It’s protection. And sometimes what we’re protecting isn’t worth saving.
The Attention Thesis
The New York Times published something extraordinary last week. D. Graham Burnett, Alyssa Loh, and Peter Schmidt—the Friends of Attention collective—made an argument I’ve been circling for years:
“True attention cannot be measured by a machine. The fullness of our authentic human attention, shared with others, is the power with which we make the world.”
They trace how attention became quantifiable—from James McKeen Cattell’s attention span experiments in the 1880s, through Norman Mackworth’s vigilance studies watching for U-boat blips in World War II, to Donald Broadbent’s model of human cognition as information filtering. Tubes and valves. Flow charts for data.
The paper “Attention Is All You Need”—the foundational text of modern AI—uses the word to mean something that has nothing to do with human attention. It means a mathematical technique for ranking information in data sets. The colonisation is complete.
What strikes me isn’t the history. It’s the implication for partnerships.
When Depth Threatens
Here’s what I’ve learned from clients who resist embedding:
Some genuinely prefer transactional relationships. They know what they need, they pay for it, the exchange is clean. No complaint.
But others resist for a different reason. Embedding means shared accountability. It means being seen—not just your deliverables, but your process. Your hesitations. Your gaps.
Simone Weil, writing in London while Mackworth ran his vigilance experiments, defined attention differently: “The authentic and pure values—truth, beauty and goodness—in the activity of a human being are the result of one and the same act, a certain application of the full attention to the object.”
Attention rooted in love, care, and commitment. An ethics of attention that cannot be sold or stolen.
When a partnership deepens—when it moves from transaction to embedding—it requires this kind of attention. And this kind of attention reveals. It shows the gaps in the organisation. The misalignments. The places where stated values and actual practice diverge.
Some clients resist embedding not because they prefer distance, but because depth exposes what shallowness hides.
Amelie on the Rock
She fell at about 15 metres. Good gear, well-placed. The system worked. She lowered, checked herself—and wanted to keep going.
Here’s what surprised me: she hadn’t lost her confidence. She wanted to climb back up, keep leading, push through that long runout. She knows what I’ve taught her: fear calcifies if you let it settle. The instinct to continue was right.
But I made her stop.
I climbed up because something felt wrong. And indeed—that 5c was more like a 6b, slippery and deceptive. The rock was lying about its difficulty. I made her wait, observe my moves, understand the line properly, then follow.
Not because pushing through proves anything. But because I knew I had to protect her confidence and exuberance for the real victory. That battle—on that deceptive rock, with incomplete information—could have broken something more important than one climb.
The Protection Paradox
This is the part I hadn’t fully understood about resistance to embedding.
Sometimes resistance isn’t fear of being seen. Sometimes it’s wisdom about which battles matter.
Amelie’s instinct was right: don’t let fear settle. But the situation was wrong: deceptive rock, understated grade, a setup for failure that would have felt like personal inadequacy rather than environmental mismatch.
The mentor’s job in that moment wasn’t to encourage forward. It was to say: not here, not now. Protect the capacity to push for situations that deserve it.
I’ve been wrong about some client resistance. I assumed they were protecting themselves from depth. Some were. But others were doing something smarter: recognising that this particular partnership, this particular moment, wasn’t the right place to invest that vulnerability.
The question isn’t just “are you willing to be seen?” It’s “is this the right place to be seen?”
The AI Safety Parallel
This week, Ofcom opened a formal investigation into X over Grok’s deepfake imagery. Prime Minister Starmer called it “disgusting and unlawful.” Liz Kendall announced that creating non-consensual intimate images is now a criminal offence in the UK.
Good. But reactive.
I’ve been tracking AI safety frameworks across jurisdictions. The Chinese approach—articulated through the Cyberspace Administration, the CAICT safety standards, the “core socialist values” requirements—is comprehensive but serves the wrong master. Safety isn’t about preventing harm to persons; it’s about protecting state narratives.
The Western approach moves faster, creates more innovation, but leaves gaps. The Online Safety Act has teeth, but those teeth only bite after damage is done.
What neither approach has built: protocols designed like aviation safety, nuclear safety, structural engineering. Systems where certain categories of failure are excluded by architecture, not addressed through consequences.
The attention economy makes this harder, not easier. When the business model is capturing attention and monetising it, safety becomes a cost centre. You invest the minimum required for compliance, not the maximum possible for protection.
The paper that founded modern AI architecture is titled “Attention Is All You Need.” And it’s true—for the machine. But for humans, attention in the machine sense is exactly what we need to escape.
What Embedded Partnerships Require
My third principle (Adventure Over Comfort) is that relationships form when you’re uncertain together, vulnerable together—when the professional performance drops away, and the actual human shows up.
But Amelie on that rock taught me something I’d missed: there are two kinds of resistance to depth.
The first is protection from exposure. The fear of being seen. The client who keeps you at arm’s length because embedding would reveal gaps, misalignments, the distance between stated values and actual practice. This resistance costs them growth.
The second is protection of capacity. The wisdom to recognise when a situation doesn’t deserve the vulnerability. The client who says “not this partnership, not this moment”—not because they fear depth, but because they’re saving it for where it matters.
I used to read all resistance as the first kind. Now I’m learning to distinguish.
Amelie’s instinct to push through was valuable—I’ve worked hard to build that in her. But on that deceptive 6b masquerading as a 5c, her instinct would have been wasted. Worse: a failure there would have felt like her failure, not the rock’s lie. It would have cost her exactly the confidence she was trying to protect.
Sometimes the mentor’s job is to say: your instinct is right, but not here. Save it.
Attensity
The Friends of Attention call their movement “attensity”—attention activism. Fighting the forces that “frack human beings in order to extract the financial value of their attention.”
I don’t lead a movement. I lead Friday bike rides and occasional mountain climbs with people who become partners because we’ve been uncomfortable together.
But I recognise the same principle: attention that cannot be sold or stolen is attention worth fighting for. Partnerships built on it last. Partnerships built on extracted, quantified, optimised attention don’t.
And I’m learning this: the attention worth protecting includes the capacity for depth itself. Spending it on the wrong partnership, the wrong moment, the wrong deceptively-graded rock—that’s a kind of extraction too.
The clients who resist embedding may be protecting themselves from being seen. Or they may be protecting their capacity for depth for partnerships that deserve it.
My job is to know the difference. And when the situation is right—when the rock is honest about its grade—to be worth the vulnerability.
This is part of an ongoing series on partnership principles. Previous essays covered “Partnerships Over Transactions” and “Embedding Over Advising.”
References:
- D. Graham Burnett, Alyssa Loh, and Peter Schmidt, “The Multi-Trillion-Dollar Battle for Your Attention Is Built on a Lie,” The New York Times, 10 January 2026
- Ofcom investigation into X, 12 January 2026
- UK Secretary of State Liz Kendall statement to House of Commons, 12 January 2026
- Attensity! A Manifesto of the Attention Liberation Movement, Friends of Attention (Crown, 2026)