Life and How to Make It

Steve Grand

My Copies


Life is not made of atoms, it is merely built out of them. What life is actually "made of" is cycles of cause and effect, loops of causal flow. These phenomena are just as real as atoms—perhaps even more real. If anything, the entire universe is actually made from events, of which atoms are merely some of the consequences.

  • p. 6 (Harvard University Press, 2001)

Intelligence is first and foremost about common sense. Reasoning (which is only one of the many aspects of intelligence) must be built upon a foundation of common sense.

  • p. 21 (Harvard University Press, 2001)

Each example of life has to compete with others for the right to exist. New clever tricks for persisting are needed urgently every day, and fortunately life has a matchless capacity for discovering them.

  • p. 55 (Harvard University Press, 2001)

Intelligence is perhaps a term that should be reserved for systems that can predict the future.

  • p. 59 (Harvard University Press, 2001)

Logic suffers from a similar one-way effect. Having seen the trick we can work out how it was done because this exercise is a narrowing-down process, aiming towards a known goal. But going the other way is much harder. Starting with knowledge of the parts, it is more difficult to work out the behavior of the whole because we do not know where we are heading.

  • p. 63 (Harvard University Press, 2001)

Cause and effect acts in webs, not in chains.

  • p. 65 (Harvard University Press, 2001)

A simulation of a living thing is not alive, and a simulation of intelligence is not intelligence. On the other hand, intelligent, living things can be made out of simulations.

  • p. 78 (Harvard University Press, 2001)

For fifty years, AI researchers have been seeking to develop serial algorithms that demonstrate the ability to think. To all intents and purposes, they have failed. To make defeat look more like a strategic withdrawal, 'intelligence' is a term that has been debased in computing, and tends to be applied to any program that performs some non-obvious computation. But most of these so-called intelligent programs are really not intelligent at all.

  • p. 93 (Harvard University Press, 2001)

Intelligence is not the ability to follow rules – it is the ability to develop the rules in the first place.

  • p. 93 (Harvard University Press, 2001)

In nature, of course, all intelligent systems are also parallel systems. Even though we feel in our heads like we are one person, thinking one thought at a time, we are really a loose confederacy of a hundred billion tiny, very stupid machines. Our brains are made from vast numbers of neurons, each operating in parallel. There is no central controller and no serial, stepwise process of execution. Intelligence is the result of billions of unintelligent processes operating concurrently.

  • p. 93 (Harvard University Press, 2001)

Adaptation is such a common feature of biological systems that I call it the first law of biology: it can be summed up in the phrase 'nature is lazy.'

  • p. 104 (Harvard University Press, 2001)

A thing is elegant if it maximizes some measure of utility while minimizing its information content (i.e. maximizing its compressibility). Elegance is therefore proportional to utility multiplied by compressibility. Another way of looking at this idea is to distinguish between complexity and complication. Something is complex if it contains a great deal of information that has high utility, while something that contains a lot of useless or meaningless information is simply complicated.

  • p. 118 (Harvard University Press, 2001)

Top-down control leads to complexity explosions, because something somewhere has to be in charge of the whole system, and how much this master controller needs to know increases exponentially with the number of components in the system.

  • p. 119 (Harvard University Press, 2001)

Intelligence is visible, relevant and possible only through the interaction of an organism with its environment, and all such intelligent action is rooted in survival.

  • p. 134 (Harvard University Press, 2001)

There is no such thing as half an organism – life and intelligence are properties of wholes and must be synthesized in a holistic way.

  • p. 134 (Harvard University Press, 2001)

Intelligence without action is pointless. In fact, [...] intelligence without action is not even achievable. Unless an organism receives sensory feedback from its environment and can also alter that environment, it cannot close the loop that makes learning and therefore intelligence possible.

  • pp. 134-135 (Harvard University Press, 2001)

The real world, complex as it is, is entirely self-consistent, and the feedback we receive from it is a natural consequence of that self-consistency.

  • p. 137 (Harvard University Press, 2001)

Life is not the stuff of which it is made – it is an emergent property of the aggregate arrangement of that stuff. Even the stuff itself is no more than an emergent property of a still smaller whirlpool of interactions. Living beings are high-order persistent phenomena, which endure through intelligent interaction with their environment. This intelligence is a product of multiple layers of feedback. An organism is therefore a localized network of feedback loops that ensures its own continuation.

Intelligence cannot be abstracted – we have to build a whole organism. Neither can intelligence exist in a vacuum – it has to be embedded in a self-consistent environment. Life is the sum total of all the feedback within the organism, and between the organism and its environment. The division between organism and environment is not a real boundary, but a convenience dreamt up by our own brains – the universe is really just a single jumble of interactions.

  • p. 146 (Harvard University Press, 2001)

Instead of "command and control," we need to "nudge and cajole." [...] Whether you run a school, run a country, manage an ecosystem or write computer software it makes no difference: complex adaptive systems cannot be dictated to – you have to learn how to go with the flow and nudge individual components in order to encourage the system to go in the direction you want it to.

  • p. 149 (Harvard University Press, 2001)

Most lessons in life do not teach you what to do—they teach you what not to do.

  • p. 165 (Harvard University Press, 2001)

This audience had been exposed to the dire predictions of too many doomsayers. Some of these false prophets are actually AI researchers themselves, yet they see nothing wrong with painting a terrible future for humankind in which machines rapidly surpass us in intelligence and we become little more than a source of slaves or domestic pets for the amusement of the master machines. But this is not how it is going to be at all! Machines are not going to take over the world. Artificial intelligence will not create monsters. Almost the opposite, in fact.

  • p. 196 (Harvard University Press, 2001)

I do not find intelligent things frightening in the least—it is the stupid ones that scare me. The more intelligent a person is, the more likely their behavior is to be rational, thoughtful, and considerate. I am sure that this will apply equally well to machines. We are more ruled by machines now than we will ever be in the future, not because machines are smart but because they are stupid. We are slaves to our word processors and databases because they do not understand what we want of them and are not smart enough to work things out for themselves, so we have to adjust our own behavior to suit them.

  • p. 196 (Harvard University Press, 2001)

I accept that many of the tyrants of the past have been intelligent people – you simply do not get to be a dictator unless you are smart, so it is almost inevitable that Hitler, Mussolini, Stalin and Genghis Khan were intellectually capable people. Not intelligent enough, perhaps, to overcome their innate antisocial, megalomaniac, pugnacious character flaws, but these traits are not confined to intelligent people. Most malicious leaders are smart, but it does not follow that most smart people are malicious.

  • p. 196 (Harvard University Press, 2001)

Intelligence is a much richer phenomenon than reasoning alone. Anyone who thinks they can make a machine with the general intelligence of a human being but no emotional or commonsensical qualities at all is wasting their time.

  • p. 197 (Harvard University Press, 2001)

There is some truth to the argument that machines will want different things from humans, and at some point there may be potential for conflict because of this. But this argument is based on a shallow view of how the future will pan out. After all, it is we who shall make these machines; we who shall define for them what their drives are needs are, and what they find enjoyable and desirable in their little metal lives. Why would we design machines whose needs conflict with our own? If we build intelligent cars, they will be designed to take pride in their ability to carry us to our destinations quickly, efficiently and without knocking over a single pedestrian. They will enjoy a good drive as much as we enjoy a good night out. Why on earth would we design them any other way? And just because we give them a degree of autonomy, it does not mean that they will immediately take up a new, conflicting set of desires. They will be just as trapped inside their own drive and reinforcement regime as we are in ours. We humans rebel against our circumstances precisely because we were not designed to enjoy much of what in modern society we spend most of our time doing.

  • pp. 197-198 (Harvard University Press, 2001)

Even though our future is an inevitable consequence of our past, this does not make us mere puppets, acting out a pre-scripted drama. This is because our future, although not optional, is absolutely unknowable.

  • p. 209 (Harvard University Press, 2001)

At some point in animal evolution, something special happened which makes humans and some of our animal cousins different from the others. That something is the capacity to imagine.

  • p. 213 (Harvard University Press, 2001)

You and I do not live in the real world at all: we live in a virtual world inside our heads.

  • p. 213 (Harvard University Press, 2001)

When I trained as a teacher the most important thing I learned from watching children is that their brains are fully competent from an early age – it is only their models of the world that need a bit of tinkering with. The program is all there, but it lacks good data.

  • p. 213 (Harvard University Press, 2001)

The human brain is the most awe-inspiring and powerful simulator in the world – the depths of our imagination have never been plumbed and perhaps never will be.

  • p. 218 (Harvard University Press, 2001)

The problem with insights is that they are like skills: you can't use language to transfer them from person to person.

  • (Harvard University Press, 2001)

Where new knowledge is simply added to one's existing mental store, insights bring understanding, and understanding changes one's whole being.

  • (Harvard University Press, 2001)

Frankenstein's terrible and ultimately fatal mistake was to carry out the act of creation first, and to think about the consequences afterwards.

  • (Harvard University Press, 2001)

Intelligence involves a great deal more than the ability to follow rules. It is also the ability to make up the rules for oneself, when they are needed, or to learn new rules through trial and error.

  • (Harvard University Press, 2001)

The route to the future is often tortuous.

  • (Harvard University Press, 2001)

Understanding the behavior of complex loops of cause and effect is what intelligence is for, and our ability as human beings to perform this trick and predict the future from our knowledge of causal relationships is what enables us to survive.

  • (Harvard University Press, 2001)