Innovation & How It Happens
Why Every Technology Has a Ceiling It Can't See Coming
The same mathematical shape that describes how a rumour spreads through a town also explains why the world's best engineers keep getting blindsided by their own success.
The Idea
Most technologies don't improve in a straight line, and they don't explode infinitely upward either. They follow an S-curve: a slow, uncertain beginning where progress feels painfully incremental; a steep middle phase where everything clicks and the gains compound fast; and then a plateau, where the underlying approach quietly runs out of room to grow. The shape is named for the letter it traces on a graph, but the underlying logic is biological — it's the same pattern a population follows when colonising a new environment, expanding until it hits the limits of available resources. What makes the S-curve genuinely useful isn't the description of the past — it's what it implies about the present. When you're in the early flat section, it's almost impossible to tell whether you're watching something that will take off or something that will quietly die. When you're on the steep climb, it feels like it will go on forever. And when you're approaching the plateau, the engineers closest to the technology are often the last to see it, because they're focused on squeezing out the next increment of performance, not asking whether the whole paradigm has a ceiling. The crucial insight is this: the next S-curve rarely starts where the current one ends. It usually begins somewhere that looks worse by every existing measure — slower, cheaper, cruder. Which means the thing that replaces dominant technology almost always looks, at first, like a toy.
In the World
In the early 1990s, hard disk drives were getting better at a remarkable rate. The engineers at the major manufacturers — Seagate, Western Digital, Maxtor — were extraordinarily good at what they did, and they were climbing the steep section of their S-curve with confidence. When a wave of small startups began producing tiny 3.5-inch drives for a new market called personal computers, the established players took note and largely passed. The drives were slower, lower capacity, and worse on every metric their existing customers — minicomputer makers — actually cared about. Clayton Christensen documented what happened next in extraordinary detail. The small drives were good enough for a new class of customer who hadn't existed before. That market grew. The small drives improved. And then the customers who had been buying the larger drives found that the smaller format did the job adequately, and the price was considerably lower. One by one, the incumbents lost the market — not because they failed to innovate, but because they innovated faithfully along a curve that had already found its ceiling. The new S-curve had started in a place that, by every rational measure they had, didn't look threatening. This pattern has replayed itself with film versus digital cameras, landlines versus mobile phones, and now, arguably, internal combustion engines versus electric motors. Each time, the incoming technology looked worse first.
Why It Matters
Understanding S-curves doesn't just explain tech history — it changes how you read what's happening right now. When a new technology is being dismissed as a toy, or as only suitable for a niche market, or as good enough for amateurs but not professionals, that is precisely the description you'd expect of something at the bottom of a new S-curve. The dismissal isn't wrong, exactly — it's just measuring the wrong thing. It's also worth sitting with the uncomfortable implication for expertise. The people most qualified to judge a maturing technology are often the least equipped to see its successor, because their judgment is calibrated to the current curve. Deep expertise can become a kind of perceptual narrowing. For anyone tracking AI, energy, materials science, or biotech right now, the S-curve framework offers a useful discipline: instead of asking 'is this good yet?' ask 'is this good enough for someone, somewhere, who couldn't afford or access the incumbent?' That's where the next steep climb tends to begin — not at the frontier, but at the margin.
A Question to Ponder
What technology in your own life or work do you currently think of as obviously inferior — and is it possible you're measuring it against the wrong curve?
Get a new one of these every morning.
Start learning with Thinkable