A Thousand Brains

For now, it appears that our best option is to work hard to form enforceable international agreements on what is acceptable and what is not, similar to how we treat chemical weapons.

The point is that anything capable of self-replication, especially viruses and bacteria, is a potential existential threat. Intelligence, on its own, is not.

At the beginning of the twenty-first century, the field of artificial intelligence was viewed as a failure. When we started Numenta, we conducted market research to see what words we might use to talk about our work. We learned that the terms “AI” and “artificial intelligence” were viewed negatively by almost everyone. No company would consider using them to describe their products. The general view was that attempts to build intelligent machines had stalled and might never succeed. Within ten years, people’s impression of AI had flipped completely. It is now one of the hottest fields of research, and companies are applying the AI moniker to practically anything that involves machine learning.

Much of the volume of our brain is wiring, the axons and dendrites that connect neurons to each other. These are costly in terms of energy and space. To conserve energy, the brain is forced to limit the wiring and therefore limit what can be readily learned. When we are born, our neocortex has an overabundance of wiring. This is pared down significantly during the first few years of life. Presumably the brain is learning which connections are useful and which are not based on the early life experiences of the child. The removal of unused wiring has a downside, though; it makes it difficult to learn new types of knowledge later in life. For example, if a child is not exposed to multiple languages early in life, then the ability to become fluent in multiple languages is diminished. Similarly, a child whose eyes do not function early in life will permanently lose the ability to see, even if the eyes are later repaired. This is probably because some of the connections that are needed for being multilingual and for seeing were lost because they weren’t being used.

It is worth emphasizing again that intelligence cannot be measured by how well a machine performs a single task, or even several tasks. Instead, intelligence is determined by how a machine learns and stores knowledge about the world. We are intelligent not because we can do one thing particularly well, but because we can learn to do practically anything.

According to linguists, one of the defining attributes of language is its nested structure. For example, sentences are composed of phrases, phrases are composed of words, and words are composed of letters. Recursion, the ability to repeatedly apply a rule, is another defining attribute. Recursion allows sentences to be constructed with almost unlimited complexity.

Finally, there is the argument of extreme flexibility. Humans can do many things for which there was no evolutionary pressure. For example, our brains did not evolve to program computers or make ice cream—both are recent inventions. The fact that we can do these things tells us that the brain relies on a general-purpose method of learning. To me, this last argument is the most compelling. Being able to learn practically anything requires the brain to work on a universal principle.

Hawkins proposes multiple models of the world, constructed microcosms, informed and adjusted by the rain of nerve impulses pouring in from the senses.

I recently attended a panel discussion titled “Being Human in the Age of Intelligent Machines.” At one point during the evening, a philosophy professor from Yale said that if a machine ever became conscious, then we would probably be morally obligated to not turn it off. The implication was that if something is conscious, even a machine, then it has moral rights, so turning it off is equivalent to murder. Wow! Imagine being sent to prison for unplugging a computer. Should we be concerned about this?

The extreme flexibility of human intelligence requires the attributes I described in this chapter: continuous learning, learning through movement, learning many models, and using general-purpose reference frames for storing knowledge and generating goal-oriented behaviors.