# Universal Machines.

Type: Article
Date: 2026-06-15
Tags: Computation, Instruments, Automation, Cognition, AI
Location: Sydney, Australia
Canonical: https://www.aaronroot.net/journal/universal-machines

In 1936, working at King's College Cambridge, Alan Turing published a paper called *On Computable Numbers, with an Application to the Entscheidungsproblem*. He was twenty-three. The paper was answering a question in mathematical logic. Could there be a general procedure for deciding the truth of any mathematical statement? The answer was no.

To prove that no such procedure existed, Turing first had to define what a procedure was and what it would mean for a process to be mechanical, step-by-step, performable without insight.

In 1936 the word *computer* meant a person. For centuries, it had meant someone, usually a woman, whose job was to perform calculations by hand. Astronomical tables, ballistics trajectories, navigation almanacs. The work was procedural and exacting. Teams of human computers worked in observatories, military planning offices, universities, and government departments. Turing's paper was, in its first form, a description of what those people did.

But a person was still too ambiguous. A human computer could remember, infer, improvise, or take shortcuts. To define procedure precisely, Turing had to remove everything personal from the act of calculation. What remained was a sequence of simple acts that could be followed without judgement.

He described a single device with a tape divided into squares, a read-write head moving along the tape, and a set of instructions telling the head what to do at each step. By changing only the instructions, the machine could perform any calculation that could be precisely described. The instructions were the procedure a human computer would have followed. The tape and the head were the paper and pencil. Turing had not invented the activity. He had formalised it.

The capability was general. The purpose would come later, supplied by whoever wrote the instructions. What constrained the machine was rarely the machine itself. It was what people could afford to ask of it, and beneath that, what they could imagine asking at all.

The next decades were spent making the machine real. Colossus. ENIAC. The Manchester Baby. Room-sized systems programmed by hand and used for calculations no person could perform at speed. The narrowness of that work was not a failure of imagination. Other uses had been visible almost from the start. They were simply uneconomical, and the machine went where it could pay for itself, first to calculation, then to the clerical work of payroll, billing, and records. As the cost fell, the range of things worth asking widened with it. Institutions mostly handed computing their existing processes and asked for them faster, cheaper, and more reliable. Existing work made the easy case, because its value was already known. Work that did not yet exist could not be priced at all.

The more important possibility was different. Computation could also be an instrument people thought through rather than merely a machine that processed work in the background. Excel did not matter because it automated arithmetic. Calculators already did that. What mattered was that it turned financial models into structures that could be explored, changed, and recalculated in real time. Modern financial planning grew around that new surface for thought.

Photoshop did something similar for visual work. Retouching existed before Photoshop. What changed was the speed and fluidity of experimentation. Compositing, manipulation, and iteration became accessible enough to shape an entire visual culture.

Real-time collaboration was the same pattern again. The shift from digital documents to shared editing was not primarily about speed. It changed the way groups think together. The software became a place where reasoning happened in the open, as it was being formed.

Automation absorbs existing work. Instruments create the conditions for new kinds of work to emerge. The bookkeepers whose work was absorbed by payroll systems did not automatically become financial modellers. Financial modelling became possible at scale because the spreadsheet made models into things the bookkeeper could shape rather than only record. The instrument shaped the practice, and the practice evolved around the instrument. The compounding begins when the tool makes new kinds of contribution possible.

The same pattern is appearing now, but at another layer. Turing showed that the procedure could be lifted out of the human computer and given to the machine. Agents are beginning to lift the operation of software out of the human user. They call APIs, read records, update systems, and coordinate tasks. They may become the primary users of the instruments through which work is done. This is not a clean choice so much as a drift. The cost of intelligence is falling the way the cost of computation once fell, and the same economics apply. The existing work is absorbed first, because its value is already known. Efficiency pulls software towards machine operation, thinning the interface and moving human involvement upstream into instruction and downstream into approval. The question is whether the result looks like a calculator or a spreadsheet, whether it gives an answer or gives the user a new surface for thought.

The phrase Steve Jobs used was bicycle for the mind. He meant the personal computer not as a machine that did the work for you, but as one that let you go further under your own power than you could on foot. A bicycle amplifies its rider rather than replacing them. The rider is still pedalling.

The best tools do this by removing the work that no longer requires a person while preserving the places where judgement still matters. Trust in the instrument can grow over time. A driver does not need to understand the engine. A finance director does not audit every Excel formula. But that trust is not blind faith. It depends on instruments that remain legible enough for people to direct them. Behaviour you can predict. State you can see. Formulas that do not silently change.

Take the IDE. It was the canonical creative instrument in software, the surface for telling the machine what to become. Some systems now being built on top of it reduce that work to a problem box. Describe the outcome. Approve what comes back. Ask again if it is wrong.

The box is the calculator again, a thin surface over a deep stack. A compiler transformed your decisions within a bounded grammar. The model supplies decisions you never made. Often they are sensible. Often they are useful. But they also tend towards the same median shape for everyone, and the surface gives the person no place to intervene except by asking again.

If programming were only engineering, that would not matter. In the 1990s Richard Hamming argued it was closer to novel writing. Set the Russians and the Americans the problem of putting a man into space and they build much the same rocket, bound by the same physics. Set two novelists on the greatness and misery of man and you get two very different novels. Give two programmers the same complex problem, Hamming claimed, and you get two rather different programs.

The judgement that produced those different programs has not disappeared. It has not simply retreated into deciding what should exist. It runs through the work, and the hooks are where it gets in.

Finance has a name for the difference between what a person adds and what the system would have produced anyway. Alpha is the return above the market. Without alpha, a trader is the market. They are the average.

The same is true of anyone working with these systems. Without hooks, the person produces what the model produces, which is what everyone else using the same model produces. They are the average too. With hooks, they can change the constraints rather than repeat the request, sharpen the tests that define what good means, step in mid-flight rather than judge only at the end. What comes back through them is work the model alone would not have produced.

The universal machine has no purpose of its own. Everything specific comes from what it is told to do, and the deepest instruments are the ones that let a person keep shaping the work as it runs rather than describing the outcome once and approving what comes back. At that depth, entirely new environments become possible.

A real drug discovery environment is not a chat box connected to a frontier model. It coordinates simulations, lab robotics, retrieval pipelines, and experimental histories through one surface, and gives the chemist places to intervene.

Setting the confidence threshold that decides which compounds reach synthesis. Asking why the model ranked one candidate above another. Overriding a prediction the assay history contradicts. Sending a hypothesis to the robots overnight and reading the result against a decade of earlier runs.

It is the IDE's fork at a deeper layer. One path treats the chemist as overhead and collapses judgement into approval. The other builds for exactly this kind of direction. One automates the discipline as it exists today. The other creates the conditions for new kinds of contribution to emerge.

The same choice is appearing in education, in agriculture, in architecture, in finance. None of it emerges automatically from general intelligence. The breakthrough is not the answer machine. It is the construction of instruments that allow humans and systems to think, experiment, and coordinate together in entirely new ways.

The disciplines that emerge from those instruments may not yet have names. The people working inside them may not resemble today's programmers, analysts, or operators. The work itself may look unrecognisable from the outside.

The best instruments automate the floor and expand the ceiling. They remove the work that no longer needs a person, while preserving and deepening the places where judgement can change the outcome. The designers who build them are not just shipping software. They are deciding what humans become capable of next.