Rise of the Machines

Rise of the Machines. It conjures the portent of a dystopian future, Terminator-like or Matrix-like, pitting man versus machine in a struggle to survive and thrive.

 

A clear distinction is made between man and machine. But can we clearly nail down the differences? Humans are alive. But the machines seem just as alive. Humans are intelligent. Well, so are the machines, although we may not understand their black-box intelligence with the newest machine-learning algorithms. Humans are carbon-based. While you might think of machines as being silicon-based now, that distinction is being slowly eroded as we approach a cyborgian future. A sufficiently clever Turing machine might even fool you into thinking it was human. Sci-fi has no trouble imagining such hybrids, from Star Trek’s Data to Westworld’s Dolores.

 

In marveling at biochemistry and molecular biology, we imagine being made up of a host of molecular machines. Microscopic engines whir and hum along allowing us to accomplish a variety of tasks. (Life’s Engines is an excellent book!) Perhaps the origin of life, and the arrival of LUCA, the last common universal ancestor, was a landmark event in the history of Planet Earth. Maybe the story of biological evolution is a story of the rise of the machines.

 

We’re good at identifying man-made tools and machines. The industrial revolution led to the development of man-made engines; in the early days, these were machine-like behemoths that burned fuel and turned it into energy we could harness. You can think of an engine as an energy transducer. One form goes in. Another seemingly useful form comes out. These days our machines are much smaller, much more compact, and run by tiny engines. They seem intelligent too. We call some of them smartphones. But there’s a part of us that wants to distinguish man-made machines from seemingly natural machines.

 

Forty years ago, the physicist-metallurgist Alan Cottrell wrote an intriguing article titled “The Philosophy of Engines” (Contemporary Physics, 1979, vol. 20, pp. 1-10). The paper begins with the second law of thermodynamics, and considers how open systems, that experience a constant influx of energy from the surroundings, organize a subset of themselves into seeming life-like engines to transduce energy and dissipate it. (Most arguments about why living systems do not break the second law run along similar lines.) There are some excellent analogies in Cottrell’s paper, and although it has some equations, I found it relatively easy to follow as a non-physicist and I recommend reading the article if you find this to be a topic of interest. Here are a handful of quotes.

 

From the standpoint of statistical physics an engine is a strange object, a highly improbable configuration of atoms and quanta, which displays an enormously exaggerated motion in one of its modes. The description of such objects is a familiar task in [biology]… In fact, the total phase space of a large material system covers all possible configurations of the system, among which may be those which we recognise as engines, or artifices generally… But, according to classical statistical thermodynamics, these are unique configurations and hence extraordinarily rare…

 

So long as physics lacks a historical dimension it cannot deal with the special properties of living systems. Similarly, it cannot account for the existence of purposively made things, such as transistor radios, internal-combustion engines, ant-hills and birds’ nests, all of which are statistically improbable yet exist abundantly on earth today. Although we usually recognize such organized systems as these through their special structures, it is nevertheless not through structure that they are to be properly defined. In fact, they do not differ significantly in purely structural aspects from non-functional systems such as misconnected electrical circuits, abiotic polynucleotides and machine-like sculptures…

 

Cottrell uses an example of a self-propagating crack in a crystal to illustrate a simple kind of autocatalysis that is also replicative in nature. Comparing the process of deformation (subjecting a material system to an outside energy source) in something like plastic versus a brittle material illustrates how energy is stored in various dislocations providing an interesting interplay between stability and instability, with different degrees of meta-stability which Cottrell equates to homeostasis. Some definitions of life suggest that it exists at the “edge of chaos”, between seeming order and randomness, in a strange liminal space with a fuzzy boundary.

 

In another short readable article (with just a handful of equations), “Dissipative adapation in driven self-assembly” (Nature Nanotechnology, 2015, vol. 10, pp. 919-923), the physicist Jeremy England sketches out a scenario for how systems can be driven towards becoming seemingly more ordered structures that then function as microscopic machines to maintain themselves, possibly even becoming more complicated over time as energy flow through the system. The trick comes from how seemingly random motion coupled with absorption of “work” energy provides the route to self-sustaining structures. Here’s an example of England’s clearly-written prose.

 

The absorption of energy from a drive allows the system configuration to traverse activation barriers too high to jump rapidly by thermal fluctuation alone, and if this energy is dissipated after the jump, then it is not available to help the system go back the way it came. Thus, while any given change in shape for the system is mostly random, the most durable and irreversible of these shifts in configuration occur when the system happens to be momentarily better at absorbing and dissipating work. With the passage of time, the ‘memory’ of these less erasable changes accumulates preferentially, and the system increasingly adopts shapes that resemble those in its history where dissipation occurred… the structure will appear to us like it has self-organized into a state that is ‘well’ adapted to the environmental conditions.

 

Cottrell says structure is not enough, and function has to be considered beyond mere structure. England, by providing a mechanistic explanation, inadvertently tries to collapse the distinction in the sense that once you have a mechanism, you have a machine. The two words ‘machine’ and ‘mechanism’ have similar roots. Does this mean that man-made machines and the ‘natural’ molecular machines of biology are fundamentally no different from each other? Is it the case that the difference is only a matter of degree? Are living things merely complicated rather than complex? Robert Rosen would disagree. By his definition, if life is complex, then a mechanism cannot be comprehensively conceived and no algorithm can be devised to compute it. Mathematical equations will always be lacking and cannot represent the complex system, although it can provide interesting insights into a ‘reduced’ model of such a system.

 

Can something that started out as a man-made machine cross the threshold of complicated to complex, if such a threshold exists? Perhaps we can only wait and see if today’s smart devices will be tomorrow’s Rise of the Machines.