alf » from archive
RT @seriouspony: Past 20 years, when David Gelernter has something to say, I listen. And he has something to say: (h/t @dogriffiths)
Well, my opinion about that essay is that the author: 1. Has a (religious) conservative worldview and is afraid of big shifts towards next levels of complexity in the essence and capabilities of humanity. 2. Does not understand the essence of intelligence deeply enough -- is stuck on the superficial / implementation-specific / low-level properties of computers and minds and is not generalizing much to the higher level properties and processes where the actual interesting stuff happens. ‎- Taivo Lints
The essay is full of things I disagree with -- discussing them all would require an essay on its own -- but here are a few examples: ‎- Taivo Lints
Wrong: "your states of mind (your desire for adventure, your fear of icebergs, the ship you imagine, the girl you recall) exist only subjectively, within your mind, and they can be examined and evaluated by you alone. They do not exist objectively. They are strictly internal to your own mind. And yet they do exist. This is intolerable! How in this modern, scientific world can we be forced to accept the existence of things that can’t be weighed or measured, tracked or photographed—that are strictly private, that can be observed by exactly one person each? Ridiculous! Or at least, damned annoying. And yet your mind is, was, and will always be a room with a view. Your mental states exist inside this room you can never leave and no one else can ever enter." -- as our capability to measure brain states is becoming better and better, we are seeing more and more correlations betwen brain states and states of mind, and we are increasingly capable of inferring the mental state of a person from outside, based on the measurements of the brains state. ‎- Taivo Lints
"software cannot exist without digital computers" -- in my opinion, wrong. Analog computers can be programmed, too, and it is fully feasible to put these programs on carriers read by the analog computer, the same way that the programs for digital computers have to be put on some carriers (disk, memory) in order to be executed. ‎- Taivo Lints
"You can transfer a program easily from one computer to another, but you can’t transfer a mind, ever, from one brain to another." -- depending on how to define the mind, some features of the mind can be transferred. Most of our modern ways of thinking are transferred culturally from generation to generation. Transferring one person's mind to another person's brain is currently not possible mainly just because of our lack of sufficient brain reading and writing ability. Plus each brain has slight biological variations, but then again, as soon as computers have hardware variations, transferring the programs becomes also a lot less easy. ‎- Taivo Lints
"You can run an endless series of different programs on any one computer, but only one “program” runs, or ever can run, on any one human brain." -- on a higher level, brain can also run an endless series of differents programs (including, though very inefficently, even those that the computer can run). On the lowest level, computer is also limited to a fixed set of simple operations. ‎- Taivo Lints
"Software is transparent. I can read off the precise state of the entire program at any time. Minds are opaque—there is no way I can know what you are thinking unless you tell me." -- the precise state of the computer program can be read only when having sufficiently low-level software / hardware access. Similarly, the more we learn to have low-level access to brain, the more we are also able to read out the state of mind. ‎- Taivo Lints
"Computers can be erased; minds cannot." -- O RLY? ‎- Taivo Lints
"Computers can be made to operate precisely as we choose; minds cannot." -- the more sophisticated the computer systems, the less precisely we are able to control them. And the more we understand the brains, the more we are able to control the mind. ‎- Taivo Lints
"Computers are machines, and idle machines are wasted." -- Wut? Not running a computer on 100% load all the time is wasteful? Only in certain specific circumstances, but quite surely not in general. ‎- Taivo Lints
"That is not true of your mind. Your mind might be wholly quiet, doing (“computing”) nothing; yet you might be feeling miserable or exalted, or awestruck by the beauty of the object in front of you, or inspired or resolute—and such moments might be the center of your mental life. Or you might merely be conscious." -- here he's demonstrating a narrow understanding of computation. Feelings can also very well be viewed as (the results of) computations. As can the 'merely being conscious'. ‎- Taivo Lints
"Computationalists cannot account for emotion." -- why? There are plenty of computational models of emotion: ‎- Taivo Lints
"And there is (at least) one more area of special vulnerability in the computationalist worldview. Computationalists believe that the mind is embodied by the brain, and the brain is simply an organic computer. But in fact, the mind is embodied not by the brain but by the brain and the body, intimately interleaved." -- disembodied mind was the "Good Old AI" of the 1960's-'70's. Nowadays, embodied cognition and embodied intelligence are mainstream in AI. ‎- Taivo Lints
"Computationalists often describe the mind as an “information processor.” But feelings are not information! Feelings are states of being." -- as if information in computers would be anything else but states of being. ‎- Taivo Lints
"All other mental states (not just sensations) are left out, too: beliefs and desires, pleasures and pains, whims, suspicions, longings, vague anxieties; the mental sights, sounds, and emotions that accompany your reading a novel or listening to music or daydreaming." -- many of these are incorporated in various models. Belief-Desire-Intention architecture was popular in the '90-s and in some places probably still is, pleasures and pains are among the concepts used in reinforcement learning algorithms, whims are approximated for example by the random component in behavioral algorithms, suspicion is an assumption waiting for further verification, longing is a type of a goal, etc. ‎- Taivo Lints
"The Iron Rod" -- he is trying to disprove the whole research direction (of trying to find brain correlates to personality traits) by showing that some individual 1848 case was not scientifically reliable, as if the current research would rely on that Gage case. ‎- Taivo Lints
"To behave like a human being (Yiddish: mensch) is to realize our best selves." -- a transhumanist would argue that this realization of our best selves is about becoming more and more advanced over time, as individuals and as a species. Because that's what the humanity has, in large part, been about. But the author is exactly opposing the realization of the progress of humanity in that direction, trying instead to halt it and return it to the previous state most compatible with classical Western religions. ‎- Taivo Lints
By the way, reading this essay after just having read The Transhumanist Wager by Zoltan Istvan makes me increasingly more worried about the possibility of an escalating clash between conservatives and transhumanists. ‎- Taivo Lints
@seriouspony continues, "@Grady_Booch @dogriffiths counting on those more qualified than I am to evaluate this and at least wonder... what's driving him to say it?" ‎- alf
I'm a bit lost here: the author has happened to be a CS professor at Yale. Most of the arguments are laughable, as you noticed, and I would sincerely expect him to understand that very clearly—so, really, what is it behind it? ‎- alf
Talking about "I can read off the precise state of the entire program at any time," in a big enough system one can only read "the precise state" if there was a fair design effort put into making that possible. Again, I would expect a CS professor to know that. ‎- alf
"I hate computers, and I refuse to play with them," he says ‎- Taivo Lints
"What's happened to intellectual life on the right? Conservatives may be succumbing to their default position. Most of the candidates for this year's Republican presidential nomination denied the veracity of evolution; and, according to various polls, Republicans increasingly distrust science. As the world becomes more threatening, many people seek simple answers, and many Americans conclude that an elite—from which they are excluded—must be the source of the ills. They turn on intellectuals, professors, and presumably the specialized knowledge those experts trade in. Instead of resisting that tendency, conservative intellectuals such as Gelernter encourage it. In their flight from elitism, they end up in a populist swamp peopled by autodidacts and fundamentalists. They become cheerleaders for a world without intellectuals, hastening a future in which they themselves will be irrelevant." ‎- Taivo Lints
"Judaism is a strange book. Gelernter's stock-in-trade flourishes are present—captured between the two covers, he writes, is "Judaism at full strength, straight up; no water, no soda, aged in oak for three thousand years"—but the book is also a deeply lyrical, even sensual, accounting of Gelernter's own faith. "He is clearly convinced that he has discovered a truth that is available nowhere else, and he is celebrating it," says David Novak, a professor of the study of religion and philosophy at the University of Toronto. Gelernter, however, casts his deeply personal argument in universal terms as a "common Judaism" (he borrows the term from Israeli scholars) whose "beauties and animating principles can be recognized and (with qualifications) agreed to by all." Not surprising, there is much disagreement. David Biale, a professor of history at the University of California at Davis, takes exception to the very idea that Judaism can be boiled down to an essence. He calls that an antiquated notion with a long pedigree. "These sorts of books were a cottage industry a hundred years ago," he says. Their aim, in part, was tribal boosterism, an attempt to show that Judaism was a modern, even liberal tradition. "But," Biale says, "there was also a genuine intellectual conviction that Judaism could be reduced to a set of beliefs." Such a view, he adds, has largely been abandoned by contemporary scholars, who tend to regard Judaism as a complex, contradictory phenomenon." ‎- Taivo Lints
Gelernter is ... how to say this ... pretty peculiar. he pioneered research into virtual reality and general UI concepts (see "Mirror Worlds" and the Lifestreams prototypes) but he's very Jewish-orthodox and neocon leaning in later years. surviving the Unabomber package was no small feat though, he was much damaged then and barely survived ‎- religion-neutral stethoscope
When strong personal preferences and beliefs go into conflict with reason and observations, then one or the other has to give way, and different people solve such conflicts differently, depending on how rational they are and how strong their preferences and beliefs are. And smart people can be very good at rationalizing away the inconsistencies in their own thinking, and if they are not very good at noticing and avoiding the various common human biases, then they might not even be aware of that rationalizing and selectivity. Or, alternatively, if they do become aware, they may sometimes also knowingly choose to sacrifice some things on the reason's side for the greater goal of having the world fit their preferences and beliefs, though I guess (hope) this happens less often than the not noticing of own biases. ‎- Taivo Lints