Beneath The Surface

In my last post, I argued that AI might serve as our "Gutenberg moment", not by replacing what we do, but by fracturing the monopoly on who gets to create. It leaves a question hanging, one that goes to the heart of what alchemy actually demands of us.

If alchemy requires the idiosyncratic, the half-formed, the things we notice that others don't, then it has a problem with distance, because I do not think that those qualities travel well. They belong to particular people in particular places, dealing with particular circumstances. They are, by their nature, local. They are part of the inconvenient elements that commensuration, the averaging that metrics require to work at scale, must eliminate.

This matters because almost everything we've been taught about value creation runs in the opposite direction. Scale wants universals. It wants what works everywhere to work the same way everywhere. The entire logic of the platform economy, and of the AI goldrush now accelerating it, is to abstract away the local in favour of the general. But alchemy can't work like that.

Scale craves certainty, and wants simplicity to moderate the complexity it encounters. Scale doesn't like handmade. It prefers Lego brick substitutability.

Alchemy requires trust, whether that's in materials, people, or the equipment we use, because alchemy is not science. Going through the same procedures will not necessarily end up with the same result. Alchemy is affected by context and all the small details that exist in materials, place, time, and mood that commensuration must ignore. Alchemy has much in common with Boyd's OODA loop. Boyd understood that the competitive advantage isn't in deciding faster, but in re-orienting more honestly, destroying your current mental model when reality has moved on and rebuilding it from what you actually notice rather than what you expected to see. Alchemy asks the same of us. It requires the constant destruction and re-creation of mental models, not the production of a one-size-fits-all works-every-time process.

Trust is an interesting animal. It takes a long time to grow but is fragile and can be destroyed in an instant. It needs constant nurturing and tending; it erodes over time unless it is frequently refreshed. We can tend it over distance, from writing through to technologies like Zoom, but for trust to do its real work it needs personal encounter. There are so many small signals picked up in the chemistry that exists between us when we're together that technology cannot replicate.

Robin Dunbar reminds us that social connection degrades quickly the more people we connect to, and that our most productive relationships are likely to be with those five to fifteen people we most connect with. Ashby's Law of Requisite Variety tells us that only variety can absorb variety; we need more available to us than is being presented to us. And Sue Heatherington, in her post on Monday, reminds us beautifully that progress comes from paying attention to the small things.

Because the small things are where the value lies. When we sacrifice them in order to measure and compare at scale, we lose texture. We sacrifice mētis, craft, and the power of presence to think in the moment, and end up being captured by the measures we use — until we can only see what we're looking for, not what is there. Serendipity is strangled, as the mediocrity required by speed, efficiency, and productivity blinds us to the genuinely novel.

We lose omoiyari, the quiet, anticipatory empathy that is the interpersonal expression of everything this argument describes. It operates only within Dunbar's inner circles. It depends on sustained attention to the particular person and cannot be scaled, metricated, or algorithmised, because it reads context, mood, history, and relationship in real time. It is perception, not recognition. It is mētis made real.

Consider the difference. Recognition says: "I know your name, your role, your stated preferences." Perception says: "I notice you're holding something back today. I notice the energy has shifted. I'm adjusting before you need to ask." Recognition can be automated. Perception cannot, because it requires being changed by what you notice. It asks something of you. It is, if you like, the OODA loop made intimate — the constant re-orientation that Boyd described, but applied to a person rather than a battlefield.

It is, I think, a small thing that changes everything for those of us looking to find a place outside the brutal economics of scale. It is about turning up as who we are, not who we think others want us to be.

And that raises a major question for us about how we turn up. It's easy to turn up the way others think they want, waiting to be chosen; it's something else altogether to turn up as who they need.

I think it's important, so I'm going to start exploring that in the workshop later this week for paid subscribers. Usual rules - making a path by walking and telling you where I've been….