Comparison of AI computations with thinking and rationality
Computers can pursue an algorithm or formula to well beyond its logical or rational limits. They can compute to the end of time. A computer is limited only by its microchips, and size of its memory. Humans are not good at this kind of activity, a few are, but they are definitely a bit weird. Computers just carry on, like Ravel’s Bolero or a Duracel battery.
Once a computer could show itself better at chess, Jeopardy and any other board game you chose to programme it with, the race was on. Computers do not have any great insights into how to play these games, they compute every possibility that can exist for the game, played by the rules. The outcome for each set of moves is given a score as to how it helps the player. All possible outcomes are computed and scored. The better the algorithms, the better computer player. This is not how Fisher won the World Chess Championships.
This is not how the human mind works. Human minds are not good at repetition, and are nowhere near as a persistent as a computer. Like DNA replication, over time we make errors. For the most part these errors are almost always disastrous, but these errors can, just very occasionally, be heroic.
That computers “Think Rationally” has led to unwarranted superstitions about computers,
That computers will replace humans
That computers are a substitute for humans
That humans can be programmed like a computer
That humans are interchangeable as computer chips
That they can be wired together
AI and philosophy have grown together in some very unfortunate ways.
This has led many scientists down a lot of blind alleys, for example David Gelenter in his book The Tides of Mind, writes “with a little imagination the brain becomes a kind of organic computer” and that this is one of the fundamental assumptions of computational neuroscience, and it is self evidently wrong.
Wrong, the brain is not an organic computer. Very few brains are capable of keeping up with a pocket calculator to more than two decimal places. Computers are, from one perspective, glorified pocket calculators, The vast majority of brains struggle working out who pays what after four of them have eaten in a restaurant together without using an app from their mobile phones. Yet those very brains, even after a bottle of wine can negotiate their ways home across London. Something which even the most advanced AI robot can’t do and it is doubtful it will ever be able to cope with the millions of uncertainties such a journey inevitably brings.
Humans are not good at any of the things that computers are good at, we are distinctly mediocre at those things computers excel at, like infinite 100% accurate computations, remembering pi to a 1000 digits. And that is why we invented computers, as tools to do the things we can’t.
Those humans who can do a little of what computers and have brains that even come close to doing what a computer can, can be distinctly odd. Autistic is often used to describe them, however in the case of Bill Gates, “Just Plain Evil” might be more appropriate.
Technology depends upon material objects that behave in 100% predictably ways, anything else, regardless of whether or not you believe in “Quantum Computing” is Science Fiction (Clarke’s 3rd Law Any sufficiently advanced technology is indistinguishable from magic). The main point about biology, is that it does not behave in a 100% predictable fashion. It is constantly on the move, and constantly adapting to its environment in order to express itself, in some form or other. Could anything but biology behave in a biological manner? It would be biology. Carbon based life-forms seem to have a flexibility, dependent on the Carbon atom, not available elsewhere in the periodic table, but it is theoretically possible.
The Computational view of Human Nature, depends upon neurones acting like binary switches. Neurones do not behave like binary switches, they have a subtly and modulation that an “On/Off” switch, cannot have. Yes both are like fireworks, once fired they have done their thing and the information leading to their firing has now been lost, possibly, maybe – but that is an assumption as far as the neurone at least is concerned. However the whole nature of information is questionable
No more are neurones binary switches – they are communication wires, extremely sophisticated wires, but they do far more than switch on and off. They one way switches, but they can be at varying states of readiness, requiring only a light touch to make them fire, or a massive stimulus, depending on the charge on their membrane, increasing and decreasing their sensitivity to any given stimulus that depends upon their surrounding inputs. In contrast to this, a computer switch is either on or off -either a 0 or a 1
Nagel, Searle and Gelenter dissent from the computational view of the mind, and so they should. The computational view of mind, is fixed in an understanding of the brain that places the neurone centre stage. The neurone is a wire, albeit a highly sophisticated set of wires, but it is fundamentally a wiring network.
The twenty first century question that challenges neuroscience and neuroscientists is – where is information generated that leads to consciousness, leads to creativity to real intelligence (as opposed to the computational ability associated with computers – anything but an intelligence, unless you think working through an algorithm is “Thinking” as opposed to “Working through an algorithm” or several algorithms, or algorithms containing feedback loops, all clever stuff but it doesn’t take a genius to do it.
Gelernter suggests that we may not have the brains that make us capable of understanding consciousness any more than a parrot can understand how to play chess. My personal belief is that in order to understand consciousness, or at least before we stop trying to understand it, we need a more accurate model of the brain. And that is “Neuroscience behind MoodMapping”, an alternative view of the brain, that is consistent with what we know and equally provides us with an alternative framework based on the facts as were familiar with, understood within a different framework.