If you've been reading DOTE for some time now, you no doubt understand that humans prefer Fantasy to Reality in the general case. When Science Daily ran a special issue on the future of computing, which was covered in the New York Times, the optimistic techno-geeks were out in force.
Holographic displays. Robotic restaurants. Computers that replace doctors, translators and drivers. If it’s proximate science fiction you want, you’ll have it, it seems, at the end of the decade.
Looking at 2020 and beyond, readers imagined a future with cures for intractable diseases, direct links between brain and computer, automated everything, contact with alien life forms, sentient machines and no language barriers.
Readers were invited to make predictions and collaboratively edit this timeline, which was divided into three sections: a sampling of past advances in computing, predictions that readers could push forward or pull backward in time with the click of a button (but not, of course, into the past), and a form for making and voting on predictions. Tens of thousands of edits were made.
Look at the timeline to get a taste of what miraculously developments will happen when. Those making the predictions kept changing the future dates—like it matters!
Human love for technology is second only to our love of money. If humans can replace thousands of fellow humans with some machinery which does the same thing those inferior humans were doing, they'll make that "positive" choice in a heartbeat. More money and machines too? You can't beat that!
Let's face it, technology is the only thing Homo sapiens is good at. The historical records makes it entirely clear that humans can't govern themselves, don't understand themselves, and can't change their own behavior. For example, they can't just say No when given the opportunity to apply some technology and destroy the lives of all those fellow (albeit inferior) humans just mentioned. Technology is the solution to all problems, even when a problem does not exist, or the technology itself creates far more problems than it solves. Technology Über Alles.
I thought these statements from the New York Times story summarized the situation very nicely. (My comments on the text are highlighted.)
Optimistic predictions far outpaced negative ones — a wishful view, perhaps, of technology as panacea. The most popular reader-submitted prediction came from Roy in Italy, who wrote that by 2020, “Google will provide everyone with the ability to communicate with everyone else, regardless of the specific language they speak, via their smartphone, with real-time language translation.”
A wishful view, perhaps?
Not all predictions were rosy. In David Gibson’s dystopian view, “humans will become so integrated with electronics that more people will die from computer viruses in a year than from biological viruses.” Readers suggested this would happen about 2170.
We will become so integrated with electronics ... that's the dystopian view?
Many of the negative forecasts were bullish on technological growth, just skeptical about our ability to control it...
Bullish on future technological progress, of course.
Techno-optimism reflects our Great Faith in our future ability to do "wonderful" things we can not do now, and don't have the faintest clue about implementing. I used to work in Artificial Intelligence (AI) in the early 1980s. These were the early days, and various university professors hyped expectations for AI beyond all reason. It was real pie-in-the-sky stuff, just like now. These academics were thus successful at attracting large sums of money for research grants or companies they had started. Needless to say, the whole thing was an Epic Fail by the early 1990s. I'm sure some progress has been made in AI since then. I'm equally sure it doesn't matter.
Take, for example, the grand meeting of mind and machine which is due to happen ... sometime in the far-flung future. I offer you Cohen's Conjectures, #1 and #2.
#1 Wetware (your brain) and Hardware/Software (computers) will not be integrated, not now, not in 10 years, not in 100 years, not in 1000 years, not ever.
#2 Machines will never achieve self-awareness, let alone true consciousness, not as an epiphenomenon or any other way. (And don't assume we will understand the mind/brain problem. We don't and we very likely won't ever understand it.)
My conjectures and hundreds of others I could come up with are based on a simple but profound fallacy about "smart" computing which goes like this: The techno-optimists literally can not tell the difference between algorithmic problem solving and actual creative human intelligence, let alone the potential of human consciousness. They want to believe that a computer beating a human expert at chess, which is just the kind of thing a computer is good at, is somehow akin to Beethoven composing the 9th Symphony, Einstein thinking up special relativity, or Shakespeare writing Hamlet.
As I said, the problem is that the geeks themselves can't perceive the difference between computing complex functions and actual intelligence. They think the two are one and the same. And it doesn't change the situation much if you implement some rudimentary machine learning. (I've done that.) Don't confuse what computers do with actual intelligence.
And let's not forget those cars that will drive themselves. Somebody alert me when that happens because I won't be leaving the house much. You won't be safe on the sidewalks, let alone on the roads. Hell, you're not safe on the roads now, not just because humans are flawed and accidents happen, but also because these same flawed humans are trying to browse the internet or text their friends or talk on the phone while they're driving. I'm sure automated automobiles will make all the difference
Reality is utterly destroyed near the end of the New York Times story.
Predictions about the far future — 2100 and beyond — took a broader view of changes that might affect all of humanity. Will we speak telepathically? Maybe by 2484, readers said.
Will we be governed by an all-knowing artificial intelligence? In 2267, perhaps.
Live forever? That could happen as soon as 2100, according to Jay Snipes of Pickerington, Ohio, who predicted, “Medical and computer sciences will learn to map the human brain, preserving the memories, knowledge, and wisdom of selected individuals before they die.”
But finally, in the last paragraph, there was a brief glimpse of Reality among these flights of Fancy.
When, if ever, will these flights of fantasy become fact? Perhaps the most accurate prediction of all belongs to R. Campos of Brazil, who wrote that in the year 2025, “we’ll be laughing at these predictions.”
In so far as I may not be alive in 2025, and for a host of other reasons—the Earth may not be habitable for humans in 2100, can you say Mass Extinction?—I'm laughing at these predictions now.
Bonus Video — An old favorite from my post The Astounding World Of The Future
I took an AI and a machine learning course a few years back. A fair amount of it was based on the failure of AI. What is intelligence and how can you model something you don't understand and are humans really intelligent (are they rational agents). We read of predictions and experiments with often humorous failures. It wasn't until the rise of Complexity Theory that we could see why they failed. Real solutions are so complex, that using computers that exist, they could never solve the problem in the time line of the Universe. Attempts to make algorithms find a solution using heuristics lead to varied results but mostly failures. The course ended up just being an advanced search algorithms course, a brief introduction of various logical systems, and Bayesian learning algorithms.
Since taking those courses, I have very little interest in AI beyond a form of advanced searches and simple learning. I snort at all AI predictions and get annoyed when people try to explain in wide eyed wonder how some DARPA project finished a race or something and how that means HAL 9000 is just around the corner. I get tired of explaining Complexity Theory and that computers that could solve these problems in polynomial time do not exist nor do we know even how they can be built. I just realized to most people computers are magic.
Posted by: Thomas | 12/15/2011 at 10:48 AM
Dave,
excellent post as usual
indeed the mass extinction is what we should worry about instead of
going crazy with techno optimism
i bet there are people who share your sentiment and the sentiment of Thomas which shows that there are instances of homo sapiens that are not completely delusional
the problem is these people are "nobodies" they have no influence on anything simply because they are outnumbered by those who cannot even graduate to understanding evolution and that religion is totally irrelevant not to mention complexity theory and how to create a socio-economic system that will compensate for cognitive biases of specific individuals
thank you again for an outstanding delivery of techno-optimism redicule - anyone who has any sort of ability to think streight and logical would agree with you
Posted by: AlT | 12/15/2011 at 01:53 PM
i wanted to add that any view of the near and distant future that does not include population overshoot, impoverishment of carrying capacity and the continuos die-off on the background of continuos impoverishment of the environment and further decrease of carrying capacity can be quickly crossed out as having no relavancy whatsoever
this is why only a handful of scholars, bloggers and those who comment have somewhat relevant inputs about reality
but even them are all acting within existing mechanism of human condition that have not yet broken down to the degree as to spur those few to start talking to each other and comparing notes
homo sapiens is failing on cooperation under the conditions that are profoundly different from those that contributed to rise of homo sapiens sapiens as the species that is capable of "deliberative capability"
it is logical to think that some individuals within current 7 billion _are_ capable of proper cooperation _and_ the circumstances of their life would be such that would make them aware of this
it is my conjecture that these individuals, if they ever could exist, should be advancced learners and adnaced communicators
i would speculate that provided there is no complete destruction of phenomenological knowledge (science) that homo has accumulated collectively within next 100 - 500 years the requisit subspeciation will occur and the final socio-economic system will be trully sustainable and will be very different from "democracy" as we live now
of course it is quite possible that it will happen only quite deep into destruction of the biodiversity and after the climate change wipes out a big proportion of the population
from our current 7 billion could probably be qualified as extinction (you know some 1 million or 10 millions in total with only few plant and animal species )
exciting life will be in 2475 :)
Posted by: AlT | 12/15/2011 at 02:33 PM
Great work, as always, Dave, and I do read your blog practically every day.
I'm ignorant about complexity theory and am bad at math, can anyone recommend some resources to get started to understand better?
Thanks... Oh, and I wanted add this...
As Ozzy sang with Black Sabbath on the song, Supernaut...
"I've seen the future and I've left it behind."
Posted by: Shawntoh | 12/15/2011 at 04:01 PM
A great line:
"Will we be governed by an all-knowing artificial intelligence? In 2267, perhaps."
Did they mention the month, day and time, I wonder?
I wish I could remember the fantastic things that were just around the corner, in the 80s and 90s. The flying car, of course. And I recall reading an article about the computer of the near future - this was round about 1990, targeting 2000. I don't remember all the features but it certainly had a screen you could fold and a reactive keyboard membrane (instead of real keys), plus a bunch of other stuff that, apparently, was all technology that was already in the labs. Still waiting for that dream computer, 20 years later.
And, on the subject of AI, I recall, when working at IBM in the early 90s, one smart Fellow was given carte blanche to dabble in whatever he wanted. Neural networks was his choice - I wonder where he got with that.
I would have lapped that sort of stuff up then; I was totally inside the box. Now I despair at such beliefs. They keep us from trying to adapt to this mess.
Posted by: Mike Roberts | 12/15/2011 at 04:52 PM
Conjecture 1 is able to be proven wrong and I'll tell you how. We assume computers will have to become more than they are now - ie empathy, self aware, compassion, etc. Overlooked is the fact that humans are loosing these qualities and our role models are now autistic eggheads. You see we have looked at the AI problem backwards.
Remember the Turing Test - if a blinded questioner is unable to tell if the answerer in human or a computer then you have achieved AI. Please watch as many interviews as you can on You Tube of Sarah Palin, esp the one in which a Fox interviewer asks her who her favorite founding father is and the only answer she could give is all of them. She - I shit you not - would fail the Turing Test - thus we have achieved AI in reverse!
Posted by: JC | 12/15/2011 at 09:18 PM
The less realism, the more optimism, of course!
Here is another one, seems like new solar breakthrough came to the rescue!
http://www.physorg.com/news/2011-12-scientists-solar-cell-electrons-photocurrent.html
"Scientists report first solar cell producing more electrons in photocurrent than solar photons entering cell"
cheers, Alex
Posted by: Alexander Ač | 12/16/2011 at 05:39 AM
We don't know what is going to happen five years from now, let a lone 100 years into the future.
Posted by: Honesty | 12/16/2011 at 07:17 AM
Ah Transhumanism, the Yang to Neoprimitivism's Yin. Gotta love it. :-/
Posted by: Mister Roboto | 12/16/2011 at 09:41 AM
I was using Computational Complexity theory. The Wikipedia page doesn't make it very clear. Basically the idea is that an algorithm can be analyzed to see how complex it is and thus how long it will take to complete. Complexity is rated as an equation that determines how many operations are needed and that determines how long. These growths in time are based on growth of input. They can be simple doubling, squaring, cubing, and so on until you reach n to the nth power. At that point the time to complete gets so big, such that after inputting 20 items the time to complete is so big it might as well be forever. The problem is that almost every problem in AI exists as one of these n to the nth problems. Computer Scientists are clever and they have managed to reduce some of this to n to a really high degree, but that means that these problems are still very complex and slow.
Posted by: Thomas | 12/16/2011 at 11:28 AM
Great posts- really remarkable, but if I don't have enough problems, now there's a insoluble "mind/brain" problem?
Seriously, folks, as Alex Rosenberg states over and over in his super-brilliant "The Atheist's guide to Reality," "the physical facts fix all the facts." What is it about the neural circuits firing away that bedevil you?
Please keep up the brilliant work - but how can you stand so much determined surfing, when there is just so much ignorance and bullshit rocking the 'net?
Posted by: Martin | 12/17/2011 at 06:08 AM
Ah the march of human progress and civilization. Except that's not what it is at all. It's always been pushed by specific sections of the human population that through varying circumstances were able to wield undue amounts power and influence over all the rest, forcing those outside of their societies to bow to their demands through force. They did this namely by constantly expanding, appropriating, and then hyper-exploiting the surrounding land bases, which practically always lead to the eventual collapse of the societies themselves.
Most indigenous peoples were happy enough living "primitively" carrying on traditions that had allowed them to survive perfectly well for centuries upon centuries, with just enough flexibility to change with the environment. But then you get the parable of the tribes situation, causing the forced civilization, expansion in complexity and arms race that has lead us down this long wondrous road to pointless crap like ipads and nerds that want to have robot bodies(transhumanists) so they can beat up the guys that picked on them in grade school. And it is all wondrously unsustainable.
Hooray for the myth of progress, it only leads us to our imminent doom.
Posted by: Wanooski | 12/17/2011 at 02:49 PM