The missing layer in AI training
Most serious conversations about AI are in a language nobody fully speaks. Jargon is everywhere; full understanding is rare. I sit down with leadership teams (in companies, in councils, in the institutions trying to plan past the end of the year) and we work through what's actually changing, in plain English. So the decisions you make this quarter still make sense the next.
Everyone else teaches you to use AI. We teach you to understand it.
The argument
These plans passively assume that the world six months from now will look broadly like today, only a bit further along. For most of human history, that worked; it was near enough. It's the assumption your annual review process is built on, your budgeting cycle, your roadmap, your strategy.
We are not alone any more. There is more than one human-level intelligence now and we're living with it. AI capabilities are compounding, but our response is not. The distance between the two is what Azeem Azhar called the Exponential Gap, and it is the single most important variable in your planning. Almost nobody's measuring it, and most aren't even aware of it.
Perspective
A slide rule was the most sophisticated calculating device on Earth for three and a half centuries. It contained zero transistors. Now, the phone in your pocket contains twenty billion.
Slide rule
0
transistors
HP-35, 1972
15,000
transistors
Today's iPhone
20 billion
transistors
An AI data centre
10 quadrillion
transistors
A single modern data centre contains roughly fifty times more computing capacity than the entire Earth had in the late 1970s. That was the era when many of us first saw a personal computer. And there will soon be hundreds of such centres. Those numbers are virtually impossible to take on board, by any of us. That's the problem we have to solve together.
This time around, for all the astonishing power of AI, the biggest revolution is the rate of change itself.
David Shapton
Who this is for
If your job involves planning, deciding or committing resources on any timescale longer than next week, the Exponential Gap is your problem.
It doesn't really matter if you work in a boardroom, a council chamber, a hospital, a university, or a regulator's office: it's the same dynamic. The same risk. The same work. The good news? I'll show you a reliable and intuitive framework for thinking about it.
If you are
designing a graduate hiring plan for the next five years. The work those graduates will do is the work most exposed to AI capability growth.
If you are
commissioning a mid to long-term technology programme on the assumption that the platforms supporting it will even exist in 2030.
If you are
writing policy or regulation while the systems you are regulating are improving faster than the legislative cycle can respond.
If you are
advising a board, a council or a leadership team that wants a confident answer, and you suspect the confident answers in circulation are wrong.
The approach
I can't predict the future. Nobody can. But I can help you face the right way when it arrives, and that turns out to be most of the battle.
Think of a motorway
Imagine a learner driver joining a motorway from a slip road. Their instinct is to slow down. Those forty-tonne trucks careering past make it seem like the safest thing is to ease off the accelerator. But that instinct is wrong and the answer is the opposite of what you feel. Accelerate. Match the speed of the traffic, and the moment you do, something remarkable happens. The blur goes away. The chaos becomes a system. The trucks are no longer a threat; they are part of an interactive ecosystem of controllable and controlled vehicles, and you are one of them.
That is what learning a common language for AI feels like. From the slip road, AI looks like a terrifying blur. From inside the system, it begins to make sense. You may not understand every vehicle, but you understand the rules of the road, and that is enough to drive.
Think of a sunrise
Now imagine the exponential future as a distant sunrise. A tiny glowing dot on the horizon that is going to light up everything. My job in our work together is not to predict the precise colour or intensity of that sun. My job is to make sure you are looking in the right direction when it rises, rather than standing with your back to it.
Most people, without this kind of training, will have their backs to it, probably admiring what they already know. We will spot the exponential trends together and learn how they behave. The more you know about the causes and nature of those trends, the more accurate your position will be. I can't promise you certainty. I can promise you a much better view.
What I do
I sit down with your board, your executive team, or your senior leadership group (wherever the decisions are actually made), and we work through what's changing, what it means for you specifically, and what to do about it. Plain English. No slides full of jargon.
I speak at conferences, industry events and internal groups. You'll get a talk that leaves your audience with a sharper way of thinking (about exponential change, emergence, embodiment), not a list of tools to try on Monday.
I write long-form journalism and commissioned essays on the fundamentals of AI: the ideas underneath the headlines. Written to be read, not to perform expertise.
What you'll take away
My job is not to make you fluent in AI. It is to make you confident about which questions to ask, which claims to trust, and how the pieces fit into the bigger picture.
01
You'll leave with a set of robust keys to understanding AI (exponential change, emergence, the effect of scale, and how AI "compresses" knowledge into a computable model of the world) that let you read any AI story in the press and know within seconds whether it matters to you. The framework is more useful than any specific fact, because while the facts change, the framework won't.
02
The single biggest cost of unclear language is that decisions get made on the basis of words people are afraid to admit they don't quite understand. After working with me, you and your colleagues will be using the same words to mean the same things. Disagreements get resolved into mutual understanding. Responses come faster, and outcomes become clearer.
03
How worried should I actually be? Is this hype or is it real? Is my job going to exist in five years? Are we behind? Are we ahead? These are the questions senior people carry around privately. I'll answer them, in plain English, with the honesty that comes from not selling you a product.
04
Some AI decisions need to be made now, but most don't. Knowing which is which is worth more than any specific recommendation I could give you, and it is what you'll walk out of the room with.
The stance
Much of AI's media coverage is hysterical or nonsensical; often at the same time. Our response is typically inadequate or off the mark because we're not on the pace with AI. We're lagging behind a phenomenon that's accelerating.
Here's where we should start: Think clearly. Use plain English. Ask honest questions about what's actually changing, at what rate, and what that means for the thing you are responsible for. Above all, be informed and learn to recognise exponential trends. That's what we'll do together.
About
I'm David Shapton. I'm a philosophy graduate, a writer, a former CTO, Director of Communications and Editor in Chief of a technology publisher. When I was born, the state of the art in consumer electronics was a valve radiogram. Today, we can hold PhD-level conversations with AI.
When I started out in digital media, a decent computer cost the price of a small car and stored a thousandth of what your phone now carries in its pocket. Today I get astonishingly detailed answers from frontier AI models, and I have to remind myself that this is the worst they will ever be. This is exponential change, and I have lived through it and written about it for decades. Artificial intelligence is the largest instance of it I have ever seen, by a long way.
I was co-founder and editor-in-chief of RedShark News, and Managing Partner for Content at a US communications firm. I have published more than a thousand articles on technology, and I write and teach about the ideas that actually matter when you're trying to think about AI: exponential change, emergence, embodiment, and what they mean for organisations and individuals trying to keep up. I draw on the work of Anil Seth and David Chalmers when the conversation turns (as it increasingly does) to consciousness, and what it might mean for a machine to have one.
I speak at events like MojoFest, where in 2025 I gave a talk in Dublin to around 600 journalists (from CNN-level newsrooms to independent content creators) about what AI is doing to the way we understand the world.
Get in touch
I take on a small number of engagements each quarter. The best place to start is a short email telling me what you're trying to figure out. I read everything that comes in.
david.shapton@futuretransform.com