The Steadmann Curve

I've been thinking about an idea for quite a long time now and it just came up again today, apropos of nothing it seems, in a conversation I was having with Claude Fable 5... so we worked through it a bit and co-authored this blog post. And while the post is written in my voice, I thought it would be appropriate to share the byline since Claude did most of the research, composition and 99% of the actual typing (I even let it keep the em-dashes... just kidding!)

The Steadmann Curve

Every few months the argument comes around again: is technology on an exponential, or are we hitting a plateau? Is AI about to go vertical, or has it already stalled? Pick your year and the same fight was happening about the web, about smartphones, about the space program. Both sides bring charts. Both sides are usually wrong, because both sides are looking at one curve.

I've been using a different model since my college days, and it has never once let me down... not in describing technology, not in describing culture, and not in describing the shape of my own working life. I call it the Steadmann curve. I'll explain the name later; it's not what you think.

One curve first

Start with a single S-curve, because everything worth building follows one.

When you take up anything new - a job, an instrument, a language, a company - you start flat. Not slow: flat. You put in effort and nothing visible comes back. This is the stretch everyone misreads, in themselves and in technologies. The early flat is not "nothing happening." It's the machinery of compounding assembling itself below the threshold of visibility - vocabulary before sentences, calluses before chords, trust before referrals. A new technology in this phase looks like a toy. A new hire in this phase looks like a mistake.

Then you hit the knee - the bottom bend of the S - and growth turns rapid. Everything you laid down in the flat starts feeding on itself. Skills combine. Each gain makes the next gain cheaper. From the outside it looks like overnight success; from the inside you know it was the flat years, finally paying back everything you invested, with interest.

And then, always, the plateau. The curve that felt unstoppable begins to saturate. The big wins have been won, the expressive space of the skill or the market or the medium has been mostly explored, and more effort starts buying less return. There's a reason the shape is universal, and it's not mystical: an S-curve is simply what happens when something that feeds on itself grows inside a limit. Compound interest meets a wall. Bacteria in a dish, transistors on a die, a career in a field - same math, same fate.

Most tellings stop there, at the tragedy of the plateau. That's where the interesting part starts.

The staircase, and what the plateau is actually for

Here's the heart of the model, and the part I've never seen stated plainly enough in the business books:

The run-out of the old curve funds the run-in of the new one.

A mature curve throws off surplus. Free time, because you're efficient now. Money, because you're paid for mastery. And something subtler - perspective: from the top of a plateau you can see terrain that was invisible from the flat. That surplus is not a retirement package. It is seed capital, and its proper use is to fund the flat, awkward, costly beginning of the next curve... because the next curve always starts below and inside the old one. Below, because a new venture is genuinely worse at first than the mature thing it will eventually replace, and switching carries real costs - you will be bad at the new thing while you're still good at the old thing, and it will sting. Inside, because the new curve incubates within the shelter the old one provides: the day job funds the workshop, the established craft pays for the apprenticeship in the next one.

Then the new curve hits its own knee... and when it does, it climbs past the old plateau. Step complete. Repeat for a lifetime and you get not one S-curve but an interlocking staircase of them, each rooted below and inside its predecessor, each ultimately exceeding it.

If you zoom out from one person to a civilization, the picture doesn't change, because culture is technology. Language, musical forms, architectural methods, systems of law - all of them are tools, and all of them run the curve. A harmonic language starts as an unfashionable experiment sheltered inside a mature style, hits its knee, saturates its expressive space, and its masters' late surplus - craft, audience, patronage - underwrites the awkward beginnings of the next form. The arch, the sonata, the skyscraper, the English language: interlocking staircases, all the way down... and all the way up.

Why the mechanism matters

I'm not the first to notice that new curves replace old ones. Richard Foster built a strategy book on it in the eighties; Cesare Marchetti showed decades of energy transitions decomposing into overlapping logistic curves; Derek de Solla Price proved in 1963 that the centuries-long "exponential" growth of science is actually a stack of saturating S-curves; Ray Kurzweil made the stacked-paradigm version famous. And if I absorbed the idea anywhere, it was probably from Jacob Bronowski's The Ascent of Man, whose entire thesis - that human ascent is cumulative tool-making, where the tools include grammar and perspective drawing - is staircase imagery from the title onward.

But most of those tellings are observational. They notice that successor curves show up and study why incumbents miss them. What the observation lacks is a mechanism - an answer to why the staircase keeps getting new steps at all. The Steadmann curve's claim is that the plateau manufactures its own successor. Saturation redirects surplus toward exploration almost automatically: bored experts, idle capacity, and resources with nowhere productive to go inside the old curve are the standing army of the next one. The staircase isn't luck. It's the plateau's exhaust.

That mechanism explains the two ways people and companies actually die on this terrain. The first is refusing the dip: riding the old plateau, harvesting comfortably, until the surplus that could have funded the next run-in is gone... and with it the fresh perspective, which decays faster than the money. The second is misreading the flat: abandoning a new curve during the years it's supposed to look like failure, because nobody told you the flat is the work.

It also explains why nobody can time any of this from the inside. Early on, an S-curve and an exponential are mathematically indistinguishable - the data that would tell you whether you're at the knee or the shoulder doesn't exist until after the bend. Which is why every plateau in history has been loudly declared the end of progress, and every knee declared a singularity, and why both declarations keep being wrong on schedule.

The live example

Which brings us back to AI, the argument of the day. Run the model against the record: pretraining scale was a curve, and as it bent, its surplus - cheap compute, accumulated talent, capital, and the models themselves - funded the next curve. Human-feedback tuning started below and inside it (the early versions were worse raw predictors) and then exceeded it. Reasoning training started below and inside that - slower, costlier per answer - and then exceeded it. Every "scaling is dead" plateau announcement has so far been the flat of the next step, misread. The question worth asking about AI is never "exponential or plateau?" It's the Steadmann question: is there visible surplus, and what is incubating below and inside? Right now the surplus is oceanic and the incubators are full. Draw your own conclusion about which part of the S we're standing on... just know that, strictly speaking, you can't. That humility is part of the model too.

About the name

If you ask me casually, I'll tell you I call it the Steadmann curve because I like the name. That's the flip answer, and it isn't even false. The real story is that years ago I decided this model was the core of a business - personal development built around the staircase instead of the hustle: teaching people to read which part of the S they're standing on, to stop mistaking the flat for failure or the plateau for the end, and to treat surplus as seed capital rather than a reward. I believed in it enough to register Steadmann.com. But unfortunately I never got around to doing anything about it. Other curves were paying too well.

As for the name itself... tracing it honestly, it started as a respelling of Stedman, after Stedman Graham, Oprah's partner: the man who stood in the wings of the brightest spotlight in America, supporting without needing to be seen. I didn't notice until years later that this is the model's central character - every next curve starts as a Stedman: below, inside, quietly essential, waiting for its time to enter the light. The extra letters were for the domain name, but they earned their keep: Steadmann is also just "steady man," and steadiness - through the flat, on the plateau, into the dip - is the entire discipline.

Which makes this post something more specific than a retrospective. It's a run-in. The idea has been sitting in its flat for years, sheltered below and inside the work that funded it, exactly as the model prescribes. Consider the knee under construction.

One more thing, for the record. There's an established principle called Stigler's law of eponymy: no scientific discovery is named after its original discoverer. (Stigler, naturally, credited it to someone else.) Price did the math and Foster got the book sales; Foster wrote the book and Kurzweil got the fame. By that law, the surest way to get a curve named after you is to describe it clearly and then wait for history to misattribute it.

The TL;DR

So here it is, described clearly. The Steadmann curve: progress is an interlocking staircase of S-curves, in which the run-out of each mature curve funds the run-in of its successor, which starts below and inside the old one and exceeds it at the knee. It's a description of how technologies advance, but it was derived the honest way - from decades of starting flat, hitting the knee, harvesting the plateau, and paying the dip to start again.

If you're on a plateau right now: congratulations, and be mindful of the resources it's provided you. That time, space and wealth... the seeds of your fruit... are the most valuable thing you own, and they're ready for planting.