When execution becomes commodity, what remains durable?

When the Cost of Making Things Falls to Almost Nothing

Superside was founded in 2015 on a proposition that felt genuinely radical at the time: creative work at high volume, delivered fast, by a distributed global team that could turn around a campaign asset while a client in San Francisco slept. The model was elegant precisely because it treated execution as the product. Speed and scale were not side effects — they were the offer.

Which makes Superside an unusually precise case to sit with right now. Because the thing it built its business around — fast, consistent, high-volume creative output — is the thing that generative AI can now approximate at a fraction of the cost. Not eventually. Now.

McKinsey's 2023 State of AI report found that generative AI was seen as most immediately useful for automating content generation, code, and design tasks. Not strategy. Not problem framing. The execution layer, specifically. If that finding holds — and the evidence since has done little to complicate it — then organisations whose value proposition lived almost entirely in that execution layer are facing something genuinely strange: the market has moved to meet them, but from underneath.

This is the context James Hurst was speaking into at Shift Summit. The talk centred on a specific claim: that AI's value to a creative process depends almost entirely on where in the sequence it is introduced. Use it early, before the thinking has happened, and it accelerates toward the wrong answer faster than any previous tool. Use it after the strategic frame is set, as an execution accelerant, and it behaves very differently. Timing, in James' framing, is the real variable — not the technology itself.

That is a useful distinction. But it also sets up a tension worth sitting with, because James' own published argument from Use Design To Design Change presses on exactly this point from a different direction. The book's structural claim is that features get copied and prices get undercut — that the only durable differentiator is a purpose that shapes culture and behaviour from the inside out. The argument is made about brands, but it applies with uncomfortable precision to the organisations that serve them. If an agency or in-house team has built its identity around the speed and volume of what it makes, rather than the quality of what it thinks, then it has, in the book's own terms, built on the wrong foundation. Not because execution is unimportant — it is — but because execution was always the thing most likely to be commoditised.

Here is where the framework starts to test itself. James is arguing, at Shift Summit, that strategic thinking is the remaining competitive advantage — the thing AI cannot yet replicate. That is almost certainly true. But Use Design To Design Change would push further and ask: is strategic thinking, on its own, a durable differentiator? Or does it, too, need to be embedded in something — a culture, a set of values, a way of working — that is genuinely hard to copy? There is a version of the "strategy survives, execution doesn't" argument that simply moves the commodity problem one level up. If every agency and consultancy pivots to leading with strategy, strategy becomes the new execution.

What the book would suggest, and what James' own career seems to demonstrate across the work at Google, Pinterest, and beyond, is that the differentiator is not a capability at all — it is a position. A point of view so specific that it cannot be abstracted away. The kind of thing that makes a piece of work recognisable before you see the logo.

This connects, perhaps unexpectedly, to the argument running through Weird the Normal / Normal the Weird — James' more recent book, structured as a flip-book with two distinct but interdependent halves. The central discipline it describes is not choosing between the strange and the familiar, but developing the capacity to move between them deliberately. Too much weird burns out. Too much normal stagnates. The skill is the switching. Applied to the current moment in creative services, that framing suggests something interesting: the organisations most likely to find a way through are not the ones that fully automate execution, nor the ones that retreat entirely into "pure strategy," but the ones that develop a genuine practice of knowing which mode the moment calls for — and why.

Superside itself is not standing still. The company has been publicly vocal about integrating AI into its workflow, and there is something worth noticing in the fact that a business built on distributed human execution is now actively reshaping around the same tools that, in theory, threaten it. Whether that becomes a story about adaptation or about something more structurally interesting — a model where the human network and the AI layer develop a genuinely new relationship — is not yet clear.

What is clear is that the question James raised at Shift Summit does not resolve neatly in either direction. If timing is the real variable, then the next question is: who decides when to introduce the tool, and on what basis do they make that call? That decision requires judgment. Judgment requires a point of view. A point of view requires something to believe in.

Which brings it back, almost exactly, to where the book starts. And leaves open the question of whether the creative industry is ready to be honest about which parts of its self-image were always about speed, and which parts were always about something harder to name.