What Properties Might the Next Machine-Class Have?

The Universal Turing Machine was concocted as a means of understanding how Gödel undecideability would manifest: description numbers (that generate computable numbers as their output) are what we now call computer programs, but not all of them can be relied on to come back with a final output. Turing showed that anything that was computable, could be computed by such a machine. In a similar vein, the steam engine was universal inasmuch that anything that was moveable, could be moved by it (the traction engine, later the tractor, could not only pull its own way along the road, but a payload too, and could alternatively be parked in a field, to drive other machinery). The labels "steam engine" and "computer" have in common the property of not just being the name of machine, but of a whole new class of machines. The abbreviation MC is used in these pages to signify such an enabling-technology machine class.

Themes that are considered here include, the way in which revolutionary machines put people out of work, but that machines create new types of work, and the overall conditions for revolution, and the nature of the human-machine interface. It is noted that there are two acts of genius: inventor and developer, with an ancedotal story behind the inspiration and development of each machine-class, ultimately leading to a changed view of the universe. Themes connected with the implementation include super-machine versus mini-machine, and batch mode versus real-time: off-line versus on-line, and stand-alone versus composite machines.

Revolutionary Machines put People out of Work

Machines during the Industrial Revolution took over the work previously performed by people, thereby putting them out of work; similarly for computers during the Information Revolution. We might, therefore, expect the same to occur during the next revolution. Indeed, we might even say that it would not constitute a revolution unless this did happen. By definition, a revolution involves a major upheaval in society; and, of course, the introduction of a newly invented machine can only be "revolutionary" if the work that it does is already sufficiently important to society. This probably implies that millions of man-hours are already being spent in getting the work done, and the invention of the new machines suddenly releases the bottleneck. The machines start off by emulating the humans, making redundant large numbers of the people who are presently doing it the less effective, manual way.

Thus, the invention of each MC appears at the sudden blossoming of its revolution, not at its budding. Pre-revolution foremen would tell gangs what to lift or pull, but now the engine driver tells the machine instead. A heat-engine could replace large gangs walking on a treadmill. In an early run-up to the Industrial Revolution, we also had windmills and water wheels.

Pre-revolution office bosses would tell their clerks what to compute and think, but now the programmer tells the machine instead (Kurzwell 1992). A computer could replace large gangs working as thoughtless components of a code-cracking algorithm (Hodges 1992). In an early run-up to the Information Revolution, we had television, radio, newsprint, and the Penny Post, the telephone and the printing press. Before the invention of the computer made thinking-work respectable, labourers looked down on pen pushers. Do thinkers now look down on workers in a new domain? New industries will eventually be created, but first, routine, run-of-the-mill semanticists (programmers, authors, actors, designers) might expect most to be put out of work (by AI, for example) in the same way that the computer put run-of-the-mill thinkers, such as office clerks, out of work.

Machines Create New Types of Work

As stated at the start of this page, the labels "steam engine" and "computer" also have in common the property of not just being the name of machine, but of a whole new class of machines. We might infer, therefore, that the next revolution, too, will be accompanied by the invention of the next machine-class (MC). There must be some distinctive invention, or insight, necessary, whose lack, at present, holds us back: a tool for helping us do the new type of work.

The Luddite fear, though proven true, was yet short lived. New applications that had not crossed a mind before, meant that new jobs were generated, too, far more than the number unemployed (Kurzwell 1992) albeit jobs requiring greater education and skill.

By being considered revolutionary, the steam engine and computer each became looked on as the wonder of the age: not only outdoing man's ability in the self same job; and not only finding new applications that had not crossed a mind before; but a fad for use, where none was truly required. Indeed, as well as fearing machines reducing jobs, there were others who feared our new dependency to be a weakness in society, with new machines that clogged where men before had coped.

Conditions for Revolution

To build the next MC would be an Everest-climbing challenge, an academic study for the inquisitive interest of it. Only if the new MC offers enormous economic advantage, though, will it be taken up by an entrepreneur, in big corporations, affluent countries, military powers, and other empire builders to invest in the idea being developed, for it to receive the massive funding it requires for it to be exploited, and snow-ball on a grander scale (Garside 1996). To achieve man's projects, in the past, rather than needing more force to be brought to bear, leadership and management enabled less to be squandered on divergent or uncoordinated activity.

To whom would power and wealth accrue? To those who control semantic information and expertise. Money, presently available for human operatives, could be diverted if a machine were promised to support a less skilled practitioner (Simons 1983). The more pressing the necessity, the more significant and revolutionary the invention that, at last, it mothers. The elderly, the bodily and mentally sick, the unemployed; the famine-stricken, war-torn, terrorist-attacked, organised-crime regions of the world are society's present pressing needs (Simons 1983).

However, the next MC might not be a break-through in any one of these, but in what is currently considered a more minor one. In World War Two, the hope would have been for a revolutionary weapon, not for an artificial brain. But the Atomic Age, and not the Atomic Revolution, was all that came of that. Over recent decades, there have been 'revolutions' in all sorts of fields: business revolutions, medical revolutions, genome maps. In coming decades, there will be 'revolutions' is all sorts of others: nanotechnology, quantum computing, DNA computing, 3D printing. What distinguishes these from the three that were selected for this book? The chosen three are all industrial revolutions, each spanning a whole range of industrial production (in agriculture, materials and information).

We might not presently even be fully aware that there is a problem in the area that the next MC will work. After it is invented, though, we will find it difficult to imagine how we had ever underestimated the problem, or how we had failed, for so long, even to conceive that a solution to it was possible.

It has to be an area that already consumes a large amount of effort to do it manually. Machines to help with care in the community (looking after the agèd, and baby-sitting), or in policy-making (such as machines to volunteer solutions to the problems in Africa and the Middle East), are all high-level goals. But the next MC also needs a more immediate, pragmatic, wealth-generating application. 'Operation Restore Hope' will have to come much later; the first role will be far more modest, and practical than this, more at the level of an automatic home-help.

In his lusts for wealth and power (Garside 1996), man has always had to tune his plans and strategies to his strengths and weaknesses, and so has a vested interest in analysing and understanding his own makeup. The creation of an artificial version of himself will partly be a by-product, and proof of principle, and highlight the areas that needed more analysis. In the shorter-term, it will also give him an object that can be sent off to do the undesirable jobs that need doing for the fulfilment of his plans.

Human-machine Interface

A finger on a valve to lift a two tonne weight; the heat-engine is an amplifier of man's strength (Fei1983). A finger on a keyboard to design a giant ship; the computer is an amplifier of man's mind. A finger or artistic work to convey emotion by tactile contact or language; an emotions machine might be an amplifier of man's heart (whatever that means).

We can imagine having a fairground attraction to do Indian Arm-wrestling, to determine whether a steam-engine cab gave human-comparable strength. Likewise, a chatbot could be the subject of a Turing test, to determine if a machine can think. So, we could have something similar to determine if a machine can display motivation, intention, taking the initiative and having projects.

The steam-engine arose initially from studies in physics (such as Boyle's and Charles' laws). The computer arose initially from studies in mathematics (such as Hilbert's Entscheidungsproblem). Which studies would give rise to the next MC?

Two Acts of Genius: Inventor and Developer

Who was the developer who separated fulcrum and load, to turn the prizing tool into a lever? And who the developer who separated rim and axle to change the roller to a wheel?

Savery, the inventor, unified the pumpate with the work surface. Newcomen, the developer, separated them with a metal piston face. And Watt, the developer, separated the condensor and the cylinder.

Leibenitz, the inventor, unified mechanics and number representation. Babbage, the developer, separated instruction and data streams. von Neumann's team, the inventors, then re-unified them both again; with Fortran and Cobol teams, the developers, separating them yet again.

Maxwell, the inventor, unified energy and syntactic information. Shannon and Chomsky, the developers, separated syntax from semantics.

Inspiration and Development of each Machine-Class

The inventive insight of a boiling kettle, whose lid was lifted by the steam (a lever like mechanism), was developed with the contemporary technology of bored-out canons of expanding gas that pushed up on a metal mass.

The inventive insight of a mathematical game, played on a typewriter mechanism (Hodges 1992) (a mechanical engineering mechanism), was developed with the contemporary technology of wireless, radar and telephony.

Entropy for the heat-engine; GIT for the computer (via THP). The need for reversibility in computation.

View of the Universe

The current understanding of the working of the human brain follows the fashionable invention of the age (West 1996): a system of cogs and wheels, a telephone switchboard, a computer. Similarly, the current understanding of the working of the universe follows the fashionable invention of the age (Barrow 1991): a giant system of wheels, a heat-engine, an information engine (Zurek 1990).

Super-machine versus Mini-machine

The first steam engines and computers were big, expensive and stand-alone; corporately owned super-machines (the factory engine and the central mainframe) occurred before the personal mini-machines (in household-tools, and in workstations and laptops).

Beam engines to lift and pump, and rotary engines for belt-driven factory machines: peripherals distributed around the workshop. Only later were structured servo-motor assemblages designed.

The batch-mode processing of formulae and payroll files, and real-time servicing of dumb terminals: peripherals distributed around the building. Only later were structured assemblages designed (SimD and MimD).

The first wheels were used in assemblages from the start, with many needed for a cart. But still, corporately owned super-wheels (to transport monoliths) occurred before the personal mini-wheels (on carts).

The first levers, though, were small, cheap and stand-alone; only later did big, expensive pieces of stand-alone equipment appear: corporately owned super-levers (to cleave and lift out monolithic slabs) occurred after the personal mini-levers (to open nuts and shellfish). Only later were structured assemblages designed (like cantilever bridges).

Perhaps the lever and the wheel, then, are only weaker members of the sequence.

Batch Mode versus Real-time: Off-line versus On-line

Real-time tools: used 'on-line' within the application, to deal with situations as they arise. Steam locomotive passengers wait while the work is being done. Microprocessors respond directly to the washing machine or fly-by-wire environment.

Batch-mode tools: used 'off-line' within the workshop, to make new tools for later use: steam hammers to build new steam hammers; and workstations to design new computers.

Energy and time: the concerns of Watt, of Heisenberg, and of TD2. Workshop tools can churn out tools of lesser power, or fashion the next, more powerful generation of tool. And, workshop tools can take long months to make a part that must serve for decades in a suspension bridge, or minutes in a rocket stage.

Stand-alone versus Composite Machines

The new MC will not replace our current manual ways, but will be added to our box of tools. We still use saws and hammers, and other human-powered tools, despite our engine-powered ones. We still use lathes and drills, and other human-controlled tools, despite our NMC.

And each new MC does not replace its predecessors, but becomes a tool within a brand new field. We still use heat-engines, and they fill our roads. We still use steam-engines, to power the Analytical Engine and its descendents.

All of the previous MCs can yield us interesting machines when linked up with their predecessors (Table 1): A locomotive is a heat-engine up on wheels (00110). A crane is a heat-engine with a lever (00101), a computer with a motor to control the fluids and conveyer-belts within industrial plants (01100), a computer with a motor and wheels forms a mobile robot (01110), a computer with a motor and a lever gives a robot arm (01101), a computer with a motor, wheels and levers gives a mobile robot with lifting arms (01111). The new MC might be added to computers to give them motivation (11000) and would be an especially useful addition to a robot (11111). It would also be a useful tool in its own right (10000) (just as the computer is a useful tool without being connected to a physical robot body (01000)).

Table 1. Stand-Alone and Composite Machines (where 1 means strong, and 0 weak, with fractional values possible in between)

The prospect of building successive classes of machine on suitably structured assemblages of their predecessors is explored in the next chapter. Alternatively, the reader is referred back to the table of contents.