The stress at the end of the page

When teaching in MBA programmes, I have often used a small historical example that tends to linger with participants. In the early days of the typewriter, many typists described a peculiar tension. The closer they came to the bottom of the page, the more careful they became. A mistake in the first lines was annoying. A mistake in the final lines could mean starting over.

This was not a matter of temperament or skill. It was a structural feature of the technology as it was used. Typing was linear. Insertion was impossible. Corrections were visible and often unacceptable in formal documents. The value of the page accumulated as it filled. Risk increased with progress.

What is interesting in hindsight is that the stress did not come from the difficulty of typing, but from the irreversibility of the process.

That detail matters more than it seems.

In his book The Shock of the Old, historian David Edgerton makes a deceptively simple argument: we consistently misunderstand technology because we focus on invention rather than use. Technologies become consequential not when they appear, but when they settle into routines, constraints, and institutional habits.

The typewriter did not transform work because it could type. It transformed work because organizations adapted themselves to its limitations. Attention shifted. Pacing changed. Drafting practices emerged. Care concentrated in specific moments. The technology quietly reorganised behavior.

This is a helpful lens for thinking about AI today.

Much of the current conversation still revolves around capability: what models can do, how fast they improve, which benchmarks fall. But as with the typewriter, the real effects are showing up elsewhere. In how work is sequenced. In where responsibility is felt most sharply. In which steps feel reversible and which do not.

In many organizations, AI is not replacing judgement. It is redistributing it. Certain tasks feel safer, quicker, more provisional. Others suddenly feel more exposed. As with the typist nearing the end of the page, caution increases not because the task is harder, but because the consequences of error have shifted.

This also explains a common leadership frustration: the sense that people are simultaneously enthusiastic and hesitant about AI. That combination is not inconsistency. It is adaptation. People are learning, often tacitly, where the new “end of the page” is.

Edgerton reminds us that technological change rarely arrives as clean replacement. It arrives as layering. Old routines persist. New tools are grafted onto them. The stress points are not always where we expect them.

Seen this way, the most important leadership question is not whether AI is impressive. It is whether organizations are paying attention to how its use quietly reorganises risk, care, and accountability.

The future rarely announces itself with a clean break. More often, it shows up as a tightening of attention at the end of the page.

Source: David Edgerton, The Shock of the Old. Technology and Global History since 1900 (Oxford: OUP, 2006).