Доказательство невозможности искусственного интеллекта
В ответ на очередной заход разбирательств (http://ivanov-petrov.livejournal.com/1014927.html) я получил не просто реплику, а развернутую точку зрения. Общий вывод - ИИ невозможен, поскольку интеллект в принципе есть "устройство", компенсирующее некоторые недостатки живого. Если совсем-совсем грубо, смысл таков: жизнь представляет всё в единстве и слитности, интеллект есть компенсация, позволяющая разделять (например, способность разделять представления об объектах). Так что ИИ - противоречие в определении, эта компенсация не нужна неживому.

http://ivanov-petrov.livejournal.com/1014927.html?thread=48285071#t48285071
shkrobius
I finally pulled myself into writing something coherent on the AI. It departs from what has been discussed heretofore by your readers. It also begins with the philosophical framework (Bergson's) rather than unimportant technicalities. The conclusion is that AI is impossible in principle; were it minimally possible, it would be absolutely pointless. http://shkrobius.livejournal.com/136502.html

http://ivanov-petrov.livejournal.com/1014927.html
http://ivanov-petrov.livejournal.com/987875.html
http://ivanov-petrov.livejournal.com/966679.html
http://ivanov-petrov.livejournal.com/959400.html

Can there be such a thing as artificial intellect? To make AI one has to redefine our own intellect in such a way that a facet of it can be modeled. What facet? That depends on what is seen as the essence of intellect. Perhaps the best and also the simplest definition of the intellect belongs to Henri Bergson.

To Bergson, intellect is the power of seeing things as separate from each other. That is all. Intellect is the quintessential property of Life, which is the fundamental notion of Bergson's philosophy, in which Life is opposing matter. The latter is simply how the intellect views the inert building blocks from which Life creates itself. How Life does it and through what adaptations it overcomes the resistance of matter is unimportant; it will find a way. The process itself is creative, meaning that its result is unknown and it is not predicated on the past. The only way to know the goal of Life is to let it create itself. This goal is identified with the Absolute. As Life overcomes the resistance of matter in its upward motion towards the Absolute, intellect is Life looking onward and downward, "adopting the ways of unorganized nature in principle, in order to direct them in fact." It is the contemplation of what passes by, not of what drives Life forward, which is intuition. Intellect is breaking into fragments this continous flux of discarded and deflected garbage. Bergson considers mathematics and logic as activities in which the will is suspended and mind is not active: these are pitiful attempts of the intellect to construct something out of the pieces of rubbish it itself separated from the reality. The pieces have no meaning whatsoever. Russell aptly calls Bergson's intellect a "carver that has peculiarity of imagining that the chicken always was the separate pieces into which the carving knife divides it." Intellect is a flaw in our design; in better designed forms of Life it is not needed, because it is counterproductive and unnecessary. In this respect, my own view is similar to Bergson's, see http://shkrobius.livejournal.com/121946.html

...While the relationship between consciousness and matter instantiated in the instinct of animals is sufficient and well adapted to their survival (from the point of view of the species), humans are not adequately equipped in this respect; hence the necessity of something like intelligence.

...His argument consists of four main steps. First, he shows that there must be an original common impulse which explains the creation of all living species; this is his famous vital impulse (elan vital). Second, the diversity resulting from evolution must be accounted for as well. If the original impulse is common to all life, then there must also be a principle of divergence and differentiation that explains evolution; this is Bergson's tendency theory. Third, the two main diverging tendencies that account for evolution can ultimately be identified as instinct on the one hand and intelligence on the other. Human knowledge results from the form and the structure of intelligence. Intelligence consists precisely in an analytic, external, hence essentially practical and spatialized approach to the world. Unlike instinct, human intelligence is therefore unable to attain to the essence of life in its duration. The paradoxical situation of humanity (the only species that wants to know life is also the only one that cannot do so) must therefore be overcome. So, fourth, the effort of intuition what allows us to place ourselves back within the original creative impulse so as to overcome the numerous obstacles that stand in the way of true knowledge.
http://plato.stanford.edu/entries/bergson

One may or may not agree with Bergson, but suppose he is right. Then AI is hopeless and pointless. Intellect has its purpose only in living things and this purpose is to compensate for the flaws of intuition. It cannot be created "artificially" because to have intellect the object has to be alive and only life brings forth life, so the artificiality is deceptive; at best, one succeeds in creating another deficient life form. Furthermore, it seems completely wasteful to create something which is no more than a handicap. The only worthy task would be enhancing intuition rather than modeling and re-creating a facet of the already facetous and corrupt intellect.

Can there be such a thing as artificial intellect?

см. одно из возражений http://shkrobius.livejournal.com/136502.html?thread=622134#t622134

dennett
Мне кажется, все, что вы говорите - и Бергсон - на каком-то уровне очевидно. Проблема тут в том, чтобы разобраться со словами. То, что Бергсон называет интеллектом ведь можно назвать и по-другому - творческое видение, жизненное мышление, 36-24Х - и так далее. И его определение будет работать. Я не отрицаю, что он усмотрел некий аспект человеческого бытия, важный аспект - и что этого аспекта не может быть - по определению - у неживой материи. (Сам я считаю, что о вопросах, которые задавал Бергсон надо думать по другому - надо задавать другой вопрос - есть ли у живого хоть что-то, чего нет у неживого - чем отличается живое от неживого - важна ли эта граница - эти вопросы фундаментальнее - и мой ответ на них - что живое и неживое в основе своей - одно - нет никакой специальной разницы - из общих принципов единства мира и рассуждений эта разница не нужна)

Что же касается интеллекта и искусственного интеллекта - то самое интересное - не уходить в абстрактные понятия - и оставаться в зоне теста тьюринга - т.е. в зоне практического взаимодействия человека и машины - и думать о том, что случится в ходе этого практического взаимодействия - что помешает человеку увидеть машину как нечто самостоятельно стремящееся к абсолюту. и уже ответ на эти вопросы потом даст почву, для того, чтобы решать телеологический вопрос - вопрос о том, насколько телеологично неживое, насколько оно может быть телеологично - и какие способности - типа интеллекта - ему для этого нужны.

Прошу прощения, что без перевода.